2015 Fantasy Football Projections using OpenCPU
412We are releasing our 2015 fantasy football projections in an OpenCPU app. The app allows you to calculate custom rankings/projections for your league based on your league settings. The projections incorporate more sources of projections than any other site, and have been the most accurate projections over the last 4 years. You can access the Projections app here:
http://apps.fantasyfootballanalytics.net/
For instructions how to use the app, see here. We will be updating the projections as the season approaches with more sources of projections. Feel free to add suggestions in the comments!
Share
Wow love it !!!!! I’ve been doing a average of the top in my opinion 4 sites and using their numbers . Our you getting the fantasy guru and pro football focus projections ? Like I told you in a tweet I used those two and 4 for 4 and football guys to get my average which is a lot of work. Also how do you print a cheat sheet ? This so cool I’m a kid in a candy store !!!! Hope my friends don’t find about this site !!!! Thanks great work pal will be touch
Hey John,
Our tools are free, so we’d need some funds to include subscription-based projections like Fantasy Guru and Pro Football Focus. Let us know if you’re willing to share them with us!
Cheers!
-Isaac
Just added the ability to download the data to .csv. Cheers!
-Isaac
Hey Isaac,
Great stuff on this site, actually inspired to me learn R and I’m just finishing up a dynamic mock draft program in R that I can continuously update in-draft.
I was wondering where the option to download the projections in the .csv format is? When I click download on the projections page I only am able to download as a .txt or .html file, which doesn’t seem to do too much for me.
Thanks!
Hi Ricky,
The “Download” button exports a .csv file. It’s in the top right of the page here:
http://apps.fantasyfootballanalytics.net/projections
Glad to hear you’re learning R. Would be interested in seeing your mock draft program if you’re interested in joining the team.
Cheers!
-Isaac
How do i find the file to download
Go to the app (http://apps.fantasyfootballanalytics.net/projections) and click “Download” in the top right corner
What about adding two points to qb tosses for TDs above 50yds?
Hi Mike,
We’ll put this on our to-do list.
Thanks,
Isaac
Appreciate all the work put into this site! One hiccup I found in the projection app is that once you remove players or keepers from the list online, when you export the new rankings it will still contain those players thought to be removed. It’s not a huge deal, I can go ahead and remove them but VOR and other stats are likely off.
Thanks, Mike. I’dd add this to our to-do list.
-Isaac
This is fantastic, and will save countless hours that I normally put into aggregating the projections on my own each year.
One thing I am wondering (and it’s purely cosmetic), is it possible to “freeze” the player name column so when I scroll to the right I can still see the player name when I’m viewing the stats in the most rightward columns (like AAV and Risk)?
Hey Sean,
Great suggestion, I’ll put it on our to-do list.
Thanks!
-Isaac
Hey Sean,
The player ID columns are now frozen.
Thanks!
-Isaac
Super app as doing this myself is too time consuming and I trust your accuracy much more than mine. I have the same question about printing- are you able to make an option to print all or portion of this? Also how about adding into custom scoring the option of bonus of 5 points for 300 yards passing or 150 yards rushing? Again, this is superb. Thanks
Bonuses are challenging because I’m not aware of any sites that forecast how many 300+ yard passing games or 150+ yard rushing games each player will have. Let us know if you find any projections for these estimates!
-Isaac
Just added projections for bonus points for 300+ yards passing, 150+ yards rushing, etc.
Cheers!
-Isaac
Perfect- and incredibly fast on your part! I am curious- were there very many sites that listed 300 yard and 100/150 yard projections?
Not very many, so we’re always looking for more to add. Let us know if you find any!
Just added the ability to download the data to .csv. Cheers!
-Isaac
Could you add an option to export the results to excel or a csv file? That would really be handy.
The app is terrific, I used your rankings, especially the ‘risk’ category to draft last year, and won the league.
Exporting to excel or cvs file would also allow printing whatever portion of ranks you wanted.
Yes, this is definitely on our list, thanks!
-Isaac
Just added the ability to download the data to .csv. Cheers!
-Isaac
I really like that you can exclude players to see who is remaining. It is also really slick the way that clicking on a player in the “exclude” list drops that player back into the pool of players. Would it be possible to click on players in the main list and have them automatically go to the “exclude” list? This would be really handy during drafts… just click on players that have been drafted to exclude them rather than having to type in their names…
Great suggestion, Aaron. I’ll put this on our to-do list. Thanks!
-Isaac
Hey Aaron,
You can now exclude players by clicking their name.
Cheers!
-Isaac
Hi Isaac, love the site.
I am using google chrome and the data from the app doesn’t all fit in the window, meaning its quite hard to view the players risk and points on the fly.
Is there a better browser to be using, or anyway the layout can be changed to fit it all in without the need to scroll?
Hey Pete,
We’re modifying the tool to be wider and fit the width of your screen so you won’t have to scroll horizontally. Hope to post the updated version shortly.
Thanks!
-Isaac
Just modified the app to fit the width of the screen.
Cheers!
-Isaac
Is there a way to see the projections behind the app? Like rushing yards, passing yards, etc…?
Hey Scott, good suggestion. This is on our to-do list, as well.
Thanks and keep them coming!
-Isaac
Issac, would also love to see the actual projections (yds, TDs, etc.) . When do you think they’ll be available?
Hi Alex,
We’re currently working on the auction tool to get it ready for draft season. It’s possible we’d be able to work on exporting the stat projections after that’s finished.
Thanks,
Isaac
I am on-baord with Scott and Alex. I would like to download the projected stats and not the calculated points.
Hi Jody,
Yes, this is on our to-do list after we finish the Auction tool.
Thanks,
Isaac
I was wondering if any progress has been made on making the actual projections accessible? Also, is there any way we can have access to actual stats that you use to calculate the accuracy of the projections? Thanks!
They are available!
https://fantasyfootballanalytics.net/about-the-site/faq#actualPoints
Cheers,
Isaac
Just wondering how updated are the Analysts projections?
Hi Bobby,
Right now we’re updating about once per week. We will update more frequently as the draft season approaches.
-Isaac
This is a huge step up from the Shiny Apps you have released previously. I also wanted to mention that I have had a premium membership at Fantasy Football Nerd for a couple of years now. While they claim to do the same sort of thing you are doing with compiling different projections and weighting them based on accuracy, your methods are much more transparent reassuring me that what you claim is going on under the hood is really happening.
I also had a couple of observations/suggestions about the new app:
1) Some of the data you are using seems to be very outdated. There are several players with mismatched teams. Couple of examples: Reggie Bush is listed as being on Detroit when he is actually on San Francisco. Knowshon Moreno is listed as being on Miami when he is actually a free agent. Also, Knowshon Moreno by your stats is projected to score about 41 points and has an ADP of around 242 which is kinda weird considering he is currently team-less.
2) Would it be possible to separate the player names and their affiliated team into separate columns? It would be nice to be able to sort and filter players by team.
Hey Ryan,
Thanks for the encouraging comments. We’re developing export tools so you can see the underlying data so you can truly see what’s going on “under the hood.”
1) Good catches. We updated the players table so the players now have the team designation from NFL.com. Moreno still shows as being with MIA on that website, so we are considering using MyFantasyLeagues player list to update Free Agents. Knowshon Moreno’s ADP is currently ~ 267 on MyFantasyLeague, which is down from the 241 that is in the database from the last scrape. Let us know if you see other issues!
2) This is possible–I’ll put it on the to-do list. Thanks!
-Isaac
The projections app looks great! Would it be possible to add a bye week column? Also, it would be really useful to be to be able to exclude players based on their bye week.
Hi Rick,
I can add this to our to-do list. My position is that it’s generally best to ignore bye weeks. It makes sense to draft the best possible team regardless of bye weeks because you don’t know how the bye weeks will help or hurt your chances. For instance, it might actually be beneficial for all of your starters to have the same bye week. Yes, you might lose that particular week, but you’ll be more likely to win all of the other games. I’m open to adding the column, though, as long as the visual presentation doesn’t become get too complex.
Thanks!
-Isaac
I would honestly prefer if you don’t add bye weeks, because the more simplistic you can keep the visual output the better.
Thanks for your input. One option we’re considering is to allow users to pick which columns to display.
Great app. This automates basically what I was trying to do manually, and with even more sources.
A few things I was curious about:
Do you run into problems where a given stat category is only projected by a subset of the sites, and do you then just average that category among the sites that project it?
What source do you currently use for ADP? In the snake draft app you could choose the ADP source. Any plans to include that here as well?
Is it feasible to include tackle for loss as a category for IDP, or is it too hard to find projections for it?
Again love the application. Thank you.
Also I found a curious phenomenon. When looking at Adrian Peterson with Yahoo scoring, I get a lower projected points total from averaging Yahoo and ESPN than from Yahoo and ESPN individually. Any idea what is going on here? I would expect the average projection to be the average of the individual projections.
Nevermind: I did some research, and the answer to the phenomenon was in the answer to my first question. ESPN does not project fumbles, so when Yahoo’s fumble projection was used alone, and added to the mean of the other categories, the result was lower than both individually.
Hi Eirik,
Yes, it is common for a site to have projections for some categories but not others. We do average across sites for those categories that are missing for a given source. We are currently working on letting the user pick the source of ADP. Hope to have it ready soon. Haven’t seen projections for tackles for loss (apart from sacks), but let us know if you find any sources! It very well could be that the average of the two is lower than the individual projections because ESPN doesn’t project fumbles.
-Isaac
Thanks for the reply.
Yahoo has projections for TFL, but not sure how common it is among other sites. I would imagine just finding IDP projections is hard enough. I know I struggled last year finding anything outside of Yahoo, when I did this manually. I can always pseudo include a TFL estimate manually in the app anyway, by finding a decent estimate as a linear combination of sacks and tackles, for DE, LB and DB separately, and tweaking the point values for those categories.
Hey Eirik,
Thanks, we’ll put TFL on our to-do list!
-Isaac
Hey Eirik,
You can now select the source of ADP.
Cheers!
-Isaac
There are numerous ways I found around the fantasy football web that discuss how to pick baselines for each position when determining VoRP. I notice currently in the app the baselines are set at 11-30-37-7 for QB-RB-WR-TE. Could you tell me what basis you are using to derive these?
Hi Sean,
Yes, we set the default baseline VOR value as the number of players drafted at each position within the top 100 picks, based on ADP on MyFantasyLeague.com, as described here:
https://fantasyfootballanalytics.net/2014/06/custom-rankings-and-projections-for-your-league.html
If you want to use different baseline VOR values, the app lets you change them!
-Isaac
Thanks for the response. I think that’s a really good way to set the baseline, would it be possible to add an option on the app to select how many teams are in the league. So it uses top 100 for a 10-team and top 120 for a 12-team? Or do you feel that shouldn’t really make a difference?
I don’t know if it would make a difference or not. I hope to test which baseline values are the best by comparing which ones predict the players who are most likely to end up on championship teams. Do you know where we can get that info?
I think the optimal baseline would depend on more than just number of teams, for example the positional requirements in your league. I would also argue that it’s difficult to find any baseline that is objectively correct. I’d argue that the specific baseline is not that important, since you shouldn’t just rely on VOR anyway. The Dropoff column combined with ADP is just as important, since VOR alone doesn’t tell you anything about how many points you’re likely giving up by passing on a player.
Eirik brings up a good point. Isaac, how much does positional requirements affect VOR drafting/your VOR numbers? Do you structure your app with a particular positional line-up in mind or particular # of teams drafting?
There are many different ways to determine VOR baselines (https://fantasyfootballanalytics.net/2013/04/win-your-snake-draft-calculating-value.html) and, to my knowledge, there’s no strong evidence for which is best. Ultimately, we plan to test which approaches are “best” by comparing their results to the percent of championship teams a given player is on (let me know if you know where we can find these data). Right now, we’re selecting the average number of players drafted at each position in the top 100 picks (i.e., 10 rounds for a 10-team league). Although the VOR baseline values would be somewhat different for a 12-team league (i.e., 120 picks over 10 rounds), I suspect that players’ ranks would not change much, but that is a testable question. I like the idea to have different league configurations for VOR baselines, and will put this on our to-do list. In the meantime, I’d suggest playing around with the VOR baseline values given your league settings to see how much they actually affect your player ranks.
-Isaac
Is the higher number in the risk category associated with less risk?
Hey Daniel,
Higher risk values reflect greater risk, as described here:
https://fantasyfootballanalytics.net/2014/06/custom-rankings-and-projections-for-your-league.html
-Isaac
Hey, do you think you can add a slot for rushes and receptions of over 40 yards?
In order to add it, we’d have to find a site with projections for how many rushes/receptions over 40 yards each player will have. Do you know any sites that have projections for these categories?
Isaac I love this site! I been doing this for years (compiling) and glad I found an easier way now. Yahoo! projects completions over 40 yards, rushes over 40 yards, and receptions over 40 yards. I’m not sure how accurate they are though. I don’t know of any other sites who makes these projection, so the output would only be Yahoo! projection. I’m pretty sure they’re using a correlation to some other stat to make that prediction.
Will put this on our to-do list. Thanks!
-Isaac
Hey Darren,
Just added projections for passes/rushes/receptions over 40 yards.
Cheers!
-Isaac
Thank you Isaac! I’m not sure if this was asked previously, but is there any way we can see the projected stats such as yards, TDs, receptions, and etc? Also, with the defensive scoring settings, anyway you can add points against and return yards? I know Yahoo projections include return yards, but I’m unsure of the other sites. One of my many leagues uses a scoring system that subtracts 0.5 point for every point the defense gives up. An example of this is if Seattle gives up 16 points, the defense is subtracted -8 points. If its no way to calculate points against, it’s fine. I’ll manually calculate it if you’re able to provide converged projected stats.
THANK YOU!!!!!!!!
Please set up a donation site!!!!!!!!!!!
Hey Darren,
We can put DST points allowed on our to-do list. Do you know of any sources of projections for DST return yards? We have exporting stat projections on the to-do list, but we’re hard at work getting the auction tool ready for draft season!
Thanks,
Isaac
Hey Darren,
We just added points against and return yards for DST.
Cheers!
-Isaac
Hey Isaac,
I’m not able to reply to the message below, so here’s the reply. Yahoo projects return yards. I’m looking at the additional sites and I’m not able see anyone else who projects the return yards.
CBS also projects return yards.
I see return yards for Yahoo but not CBS. Where do you see CBS projections for return yards?
I have a league using CBS platform and I’m able to see projected return yards by changing the categories from standard view to advance view. I tried to duplicate the report by using the general projections available to anyone not in a league and I’m not able to view the return yards. I could possibly provide you the link and see if you can access it.
Can you access it when you are not logged into your league? If not, we wouldn’t be able to scrape it unfortunately.
I cannot access it without being logged in. I’m still searching for an alternative. If I find a solution, I’ll be sure to let you know.
Thanks Darren. Just added a donation box on the lower right sidebar.
Thanks!
-Isaac
Thanks for updating return yards and points allowed. One small adjustment, I’m not sure how the points allowed or return yards are calculated. I think the same way yards are calculated for individual positions, it should be calculated for the DST for return yards and points allowed. So, one point per 10 would be better than the current setup. Great updates though!
I’m venturing of to join an auction league so I can use the new app the team is creating! Can’t wait!
THANK YOU for FFA!!!!!!!!!!!
Just updated DST return yards as requested. Will work on presenting the points allowed/threshold fields in a more intuitive way.
Thanks!
-Isaac
Since D/ST point allowed was mentioned, what about adding DST data settings for yardage allowed?
Love this app.
Hi Ben,
Do you know any analysts that provide projections for DST yards allowed? We’d need projections in order to include them.
Thanks,
Isaac
Hi Isaac,
Thanks for the prompt response.
I only know of a few (4 to be exact.. I know.. not a lot). I can probably hunt around some more later. Here are links to what I’ve got:
http://games.espn.go.com/ffl/tools/projections?display=alt&slotCategoryId=16
http://www.fantasysharks.com/apps/Projections/SeasonProjections.php?pos=D
http://subscribers.footballguys.com/myfbg/myviewprojections.php?projforwhat=td&projector=2&profile=0
http://subscribers.footballguys.com/myfbg/myviewprojections.php?projforwhat=td&projector=53&profile=0
Hope that helps!
Thanks Ben. We’ll add this to our to-do list.
-Isaac
Thanks Ben. We’ll add this to our to-do list.
-Isaac
Hey Mark,
Just added projections for passes/rushes/receptions over 40 yards.
Cheers!
-Isaac
Also, can you compare the NYTimes Sablich brothers accuracy to your accuracy? Its free access, they ranked 2nd to John Paulson by .1 in multi year accuracy rankings on fantasypros.
Hey Mark,
We have plans to compare the accuracy of rankings vs. projections in predicting performance. We’d probably focus on ECR rankings but might consider other sources of rankings. I typically suggest to use projections instead of rankings for many reasons, though (https://fantasyfootballanalytics.net/2014/08/use-projections-not-rankings.html).
-Isaac
Isaac,
Thank you for this awesome work! Kudos!!
Is there a way I can download the data i get based on my league settings from here: http://apps.fantasyfootballanalytics.net/projections ?? or atleast have a Print friendly view ?
Quick feedback: A save button for the league settings would be great too, else you have to re enter every setting again as there is no way to save the data 🙂
Thanks once again
Hey Bala,
These are both on our to-do list. Hope to have the data ready to download soon!
Thanks,
Isaac
Just added the ability to download the data to .csv. Cheers!
-Isaac
Fantastic, works great. Thanks again!
Thanks a ton Isaac!! Works like a charm!
Cheers
B
Hi Isaac, How would I change it to a 2QB league with 2RB 2WR 1TE and a flex
what I mean by comment above is would VOR need to be changed?
Hey Mark,
As it is currently calculated, VOR does not depend on how many players you need at each position. We have plans to include roster settings and a Dynamic VOR down the road.
-Isaac
Just wanted to say thanks! This site is very handy for us FF nerds.. You do an excellent job!
Regarding the drop off number, what two numbers is this the difference of? It is not clear from the app.
The dropoff is the difference between a player’s projected points and the average of the projected points of the next 2 players at the same position, see here:
https://fantasyfootballanalytics.net/2014/06/custom-rankings-and-projections-for-your-league.html
so based on your article written about whose projections are most accurate, I should choose average in the open CPU app and not weighted or robust?
Hi Mark,
The weights we use now reflect the historical accuracy of the projections. We did not examine the accuracy of a weighted average last season, so we don’t know how it would have compared. Applying the current accuracy weights would likely have been more accurate than an average over the last 3 years, because the weights were chosen based on analysts’ accuracy over the last 3 years (which is in effect cheating when examining their accuracy in relation to past years because that’s how the weights were chosen). It’s not clear whether those analysts who were most accurate in the past will continue to be the most accurate in the future. Our evidence suggests that a simple average is fairly accurate. If you think some analysts may continue to outperform others, you might use the weighted average. If you think it’s a wash, a simple average is probably your best bet. If you want to scrub out the crazy projections, you can go with a robust average. Next season, we’ll compare the accuracy of all of them!
Hope that helps,
Isaac
Hi Everybody!
Isaac. Great tool. Used it last year. Hope the other teams in my league don’t find out about it 🙂
Anybody else using Mac Safari 8.0.6 having trouble downloading the data? I’m not getting the file to go through to the download folder.
Thanks!!!!
Figured it out. Little bit different than last year. Thanks!!!!!!!
How did you figure it out? I am having same issue
Hi Bill,
You might try the suggestions outlined by me and others here:
https://fantasyfootballanalytics.net/2015/05/2015-fantasy-football-projections.html#comment-31033
Hope that helps,
Isaac
How frequently does the program update it’s data from the various sources? I love everything about this, but out of date information (even a few days old) is a disadvantage come draft day…
Hey John,
We update the data fairly regularly. We will be updating the data even more frequently as draft season approaches. You can always let us know when your draft is, and we’ll make the data are up to date!
-Isaac
Hi Isaac, My draft is August 15th. Any chance you could be sure to update the date on the 13th or 14th?
Really loving this site. You’ve done an incredible job with the data.
Yes, we’re updating them regularly. We will make sure to update then.
Thanks,
Isaac
Will there be updates in July and August
Yes, we will update the projections regularly through the draft season.
Hi Isaac,
There seems to be some duplication in the data – when filtering via the “Compare Player” option I get two versions of each player; one with rank, VOR, points, ECR Rank and Pos Rank and the other without all of those – all other categories the information is the same.
Thanks Pete, will fix as soon as possible!
-Isaac
Hi Pete,
We think we removed all the duplicates. Let us know if you find any more!
Thanks,
Isaac
Thanks for sorting so quickly. One Doug Martin is bad enough….
There are more: DeAnthony Thomas, Demarcus Ware, Dri Archer, Ifeanyi Momah, Terrelle Pryor (x4!).
I don’t think those are duplicates—I think that has to do with the same player being listed (and projected) at multiple positions. That’s why they have multiple rows. We will work on modifying the positions of those players to reflect their different positions, once we finish the auction tool.
Thanks!
-Isaac
…I don’t think so. Looking at it right now, Thomas is listed as WR in both cases, Archer at RB, Pryor is listed at QB four times, etc.
That’s what I mean—the underlying data lists those players at multiple positions, but the data grid isn’t currently showing the different positions. Will update it when we can—they will still have multiple rows (1 for each position they are listed at).
When do you think the auction draft optimizer will be updated with the upcoming seasons projections and values?
Hey Sean,
We’re hoping to release it later this month.
Thanks,
Isaac
I’m trying to figure out what I should use for my baseline – I have a 10 team, 2QB league with a Kicker but no D/ST or IDP. It’s a keeper league where we all keep 8 players every year. Does anyone have good ideas on how to calculate this? I’ve looked at a lot of things online that all make sense, but they all change the numbers I should use for my baseline, which then significantly alters the dataset. Would love to know if anyone has worked through this particular issue yet or has some thoughts on how I should calculate. Things I’ve tried:
– Make an assumption based on how many of each position are kept and who has been drafted in the top 20 each year. Something like: 28 QB, 28 RB, 37 WR, 7 TE.
– Only use the number of starters at each position x number of teams in the league. So 20QB, 20RB, 30 WR, 10 TE, 10K.
– Try to find data on who was drafted in what order in past drafts – not sure I’ll be able to do this. We do our drafts in person using a big board rather than online.
– Find specific examples online – I haven’t been able to find this yet. I’ve found plenty of explanations for how I might be able to generate this number, or sites saying ‘just let our software calculate this for you’, but nothing concrete.
Really appreciate any help in advance!
If you’re unable to find the data from last years draft I would suggest you use the # of starters (your option 2). Essentially the formula works in a way that the more of a position you draft the more valuable players at that position are (relative to their total score). It’s preferable to know exactly how your league drafts to further customize your draft board to fit your league. But if that’s not possible go with the # of starters at each position (My Opinion).
Have a question on the page setup.. I can’t scroll over to input the Wr’s scoring settings.. Is there a layout setting I need to change? Thanks
Can you tell me which browser you are using? Also, are you accessing the tool from a laptop or desktop computer, or perhaps trying from a phone or tablet?
I haven’t seen any problems like this myself, but am happy to try to get to the bottom of it!
I’m using Chrome.. I just can’t scroll over to the right to input all the WR scoring settings. On my Iphone it works fine
Hm, very strange… I’ve been using chrome with no issues. Would you mind sending a screenshot of what you’re seeing to isaac@fantasyfootballanalytics.net? Thanks so much for helping us track this down!
Do you feel that VOR underestimates players who are missing time? Especially when missing time due to suspensions? For example Tom Brady’s projections from source websites will only be projections based on 14 games or so (pending appeal), but that doesn’t factor in the fact that I can start my backup quarterback in those missing weeks and score an extra 20-30 points total (probably). So should I look at the average game value of a mediocre backup and add those points to Brady’s VOR (sense I have to draft a backup at some point anyway)? Or is that idea flawed in some way?
Hi John,
Good point. VOR (and projected points for that matter) are calculated for the whole season, so if a player is missing games due to suspension or injury, you could impute the value of a replacement, if you’d like. I will look into calculating a separate variable for 1) VOR per game available and 2) points per game available.
Thanks,
Isaac
Hi Isaac,
I use Safari on a Mac. When I press “download” on the upper right hand side of the projection app, it shows up in a new page as a huge bunch of text. Is that a csv file? My goal is to have the projections in an excel spreadsheet. It’s possible that I’m an idiot and don’t know how to translate the text into the proper format, but could you give me some advice? Thanks!
Andy
Hey Andy,
Are the values separated by commas? If so, it’s a Comma Separated Values (CSV) file that you can open in Excel.
-Isaac
Yes, they are separated by commas. How do I open it in Excel? When I click download it just opens up a webpage with all the text and values in one big block of text. Sorry if this is a stupid question!
I’m not too familiar with Mac. Can you click File, Save As?
OK Nevermind! It worked when I used Firefox instead of Safari.
Yes, they are separated by commas. How do I open it in Excel? When I click download it just opens up a webpage with all the text and values in one big block of text. Sorry if this is a stupid question!
Hi Andy, like Isaac, I’m not a Mac guy. However I think I may be able to help you.
1) Download the file and SAVE AS (not open) it to a place you can easily find it on your hard drive (desktop isn’t a bad place)
1a) If the document opens automatically in your web browser or word processor simply close the document.
2) Open Excel
3) Go to File>Open>Double Click the File you saved to your hard drive
(Note to Windows users: steps 1-3 can be done on a PC by simply right clicking the downloaded file and selecting “Open With” > Excel)
4) A strange window will appear in Excel – possibly looking like a hot mess of random text. Look for the option to separate by commas and make sure it is the only option selected for separating the text. Once you’ve selected the proper options that window should look like an organized excel spreadsheet.
The common mistake is to simply double click the file that you downloaded. This will generally open the data in either html format (Explorer/Safari/Firefox) or word processor format. You want to open it FROM Excel so it knows to import that information into a spreadsheet.
Good luck!!! Let me know if that worked for you.
Thanks John for your helpful comment! Everything is working great now.
Hi guys,
Would it be possible to add a scoring category for 100+ and 150+ rec yds for RBs as well? Thanks.
Hey Gage,
Do you know any analysts that provide projections of 100+ and 150+ receiving yards for RBs? We’d need to find projections in order to include them!
Thanks,
Isaac
Love the App. When I plug in 10 pt bonuses for +300 yards passing, +100 yards receiving or rushing it changes ranking drastically. To the extent of having Demarius Thomas the #1 pick and say Jamal Charles the #13 pick. Does this sound right for my league bonuses and how closely would you follow these rankings while drafting. Thanks.
Hi Chris,
What scoring settings are you using? If I just add 10 pt bonus for 300 yd passing, 100 yds rushing and 100 yds receiving as well as 1 pt per reception then the top 5 are Antonio Brown, Eddie Lacy, LeVeon Bell, Jamaal Charles and Arian Foster – which seems reasonable to me. If I remove the pt per reception then the top 5 are Marshawn Lynch, Eddie Lacy, Arian Foster, Adrian Peterson and DeMarco Murray.
-Isaac
No ppr just bonuses for 300+ passing and 100+ rushing and recieving. Non IDP. -2 for fumbles. 1 or for every 20 yards passing. Only changes I made. Does it every time. 1. Demarius Thomas 2. Antonio Brown 3.odell beckham 4. Julio Jones 5.Calvin Johnson 6. Marshawn lynch.
I must be doing something wrong. Any idea?
I can reproduce your results with those settings. Those scoring settings heavily favor WRs. On a week-by-week basis there are more receivers with 100+ yards receiving than running backs with 100+ yards rushing, so it is more likely that a receiver will have a 100+ yard receiving game than it is a running back will have a 100+ yard rushing game.
Last season there were 5 WRs with 100+ receiving yards per game on average, only one RB (DeMarco Murray) had 100+ rushing yards per game average.
Hope that helps!
-Isaac
Also, I’d like to donate to the cause if possible. Maybe so that you guys could add in PFF projections? Let me know how please.
I’m with Gage. This site needs a “Go Fund Me” link prominently featured. I’m not in a financial position to give right now, but will definitely support in the future. It’s obvious that you’re not doing this for the money, and there may be legal barriers to taking other peoples free projections, averaging them, and then selling the product. But if it’s not illegal you should get on that, because this site is totally worth supporting.
Also no advertisements??? I mean don’t go overboard and clutter your page, but surely you can find some room. If you’re putting in this much work and sharing it then you deserve a little profit.
Thanks guys, we’ll look into ways to allow donations. Also, we already include PFF projections (they are the same as Yahoo).
Much appreciated.
-Isaac
Just added a donation box on the lower right sidebar.
Thanks!
-Isaac
Possibly you could have a spot on the app that displays when the projections were last updated, so people don’t have to constantly wonder how up to date the projections are and always ask you.
Hi Sean,
Good suggestion, this is on our to-do list. We’re hard at work on the auction tool so it can be ready for draft season!
Thanks,
Isaac
Hey, Isaac, love the site! I am a middle school teacher and I am putting together a fantasy football elective that would emphasize learning and using statistics and math skills. I am going to divide the class into teams and have them compete week to week. In the midst of their competition, I am hoping they learn a little and develop an appreciation for math and statistics! This is my first run at this, so I’d love to hear any suggestions you have about how I might structure things. Thanks!
Hey Clinton,
Glad to hear you’ll be using fantasy football to teach stats. Would love to see your teaching materials when you develop them. I think understanding the differences between the mean and median, and when to use each, can be very important for understanding how to interpret statistics. You might read this article on teaching stats with fantasy football:
https://scholar.google.com/scholar?oi=bibs&cluster=9564064504560829125&btnI=1&hl=en
Best of luck!
-Isaac
Isaac,
I would love to see projected first downs for rushing and receiving added. We find it’s a better way of rewarding offensive production than PPR.
Hi Bob,
Do you know of any sources that provide projections for those stat categories? We’d need those projections to include 1st downs as stat categories in the tools.
-Isaac
Yahoo offers first down projections. Regardless, you can also get a pretty good estimate from yards/20, so just add p/20, where p is the points value for a rushing or receiving first down, to the values for rushing and receiving yards in the settings.
For example if we apply this model to the top 50 RBs in 2014 and look at rushing FD, we get an R-squared value of 0.94, which means this simple model explain 94% of the variations in the data.
Doing the same for WR and receiving FD gives a slighly worse result with R-squared equal to 0.79, but this is still quite good. If you want a slightly more accurate model, you can try first downs = yards/40+receptions/3 which gives an R-squared of 0.9. Again just adjust the points for yards and receptions in the program to include receiving first downs in this way.
Hi Isaac,
Just found your site through the fantasy librarian. Love your work. I’m definitely going to incorporate it into my drafting software for auction drafts. I noticed a comment from “Ricky” above who said he’s incorporating R into his own software to re-adjust the data on the fly as the draft goes on. What am I missing? How can I get the code for your software? I would love to take it a step further like he apparently has. I really only care about auction drafts and the various sites auction drafting webpages don’t have nearly enough functionality for a good auction draft. So, I write my own using Oracle APEX.
The Auction shiny app in R is still calculating based on last years data. When will you be changing your source data to 2015 projections? I’m seeing some projection data for 2015, but the apps don’t seem to be using it. Maybe I’m just using an old link?
Also, how do you calculate the projected auction cost per player? Are you just pulling it from the scraped data? When I add up the total projected cost for the full draft it’s well over a $2400 12-owner league with $200 budget per owner. That also doesn’t include defenses or kickers, which should be a lower % of the total budget anyway, but would push is a little higher, so I’m finding it’s about 50% too high for the overall money spent. I can always prorate the dollars down to the correct budget, but I’d like to know the methodology behind it. I care about it so I can give the league good average auction values for our keepers. I find most of the sites out there use Fantasy Toolbox numbers based on mock auction drafts and those numbers aren’t that good. If anyone else has some ideas I’d love to hear them as well.
Thanks, Tim
Hi Tim,
The code for the tools isn’t public, but all the functions it uses (calculating risk, projections, etc.) are public on our GitHub repo (https://github.com/isaactpetersen/FantasyFootballAnalyticsR). It includes all the R scripts.
The Projections/Snake Draft tool has been updated for 2015 (http://apps.fantasyfootballanalytics.net/projections). We are currently working on updating the Auction tool so it’s ready for draft season. Stay tuned! The calculation of the auction values is described here: https://fantasyfootballanalytics.net/2013/06/win-your-fantasy-football-auction-draft.html. We will be modifying their calculation this year.
Thanks,
Isaac
Hi Isaac,
I love this app. The best thing I have seen in years. I love being able to incorporate different scoring systems to get different rankings. One glitch I found is Charles Clay is listed as a running back instead of tight end. Thank you for all the hard work this had to take. This is my “go to” app for 2015.
Thanks
Tim
Thanks, Tim. We’ve isolated the issue and it’s related to the duplicates that others have observed. The fix may have to wait until we finish the Auction tool, but we’ll fix it.
Thanks!
-Isaac
Isaac,
I love the numbers you are capable of showing on your website. Is it possible for you guys to incorporate dynasty projections into the equation?
Zach
Hi Zach,
Do you know of any sources for dynasty projections? We’d need to find projections to include them. What would dynasty projections involve? Projections over the next 4 seasons?
-Isaac
Isaac,
That’s a great question. The thing with dynasty is no site does an actual scoring projection past the current year. Majority do a ranking system most of which revolves around age and previous achievement.
With that being said I think you guys could do something no one else does and unfortunately I don’t have the time to program. I would advise a system that takes into account a numerical calculation that dictates a prime age for each position. An example would be WR and their “prime” say is 27. You could take the relevant average from previous season(s) along with the current season to project the optimal point total a player can accomplish upon his peak and then show the potential decrease in potential for subsequent years out to say 4 years.
While basing a player’s value off of a scenario like the one above and rankings them accordingly. Something else to include might be a ppg total for year. Especially if a player is hurt. But the ultimate goal would be to show that, assuming he doesn’t retire early, Aaron Rodgers can play at a similar to Peyton Manning has until he is 38-40 and there for is more valuable for the next 3,5,8 years than Manning will be.
Feel free to ask more questions I’d love to help as much as possible.
Zach
Hi Isaac,
I love this website you have created. I have created similar projections and draft board within the limits of Google Sheets, but what you have done is really impressive. I specifically really like your Projections app that crowd-sources projections. I have one question. When I try to download the Projections from the app, the entire Points column contains the value “null”. Can you tell me if I’m perhaps doing something wrong? Thanks!
Hi Kevin,
The “actualPoints” column will be null unless you are examining historical data. The “points” columns represents projected points, and shouldn’t be null.
Hope that helps,
Isaac
I’d like to be able to see a list of the actual averaged projected stats so I can use them to do some of my own calculations. Can you make that available?
Hi Larry,
Yes, this is on our to-do list. We’ll look at this after we finish the Auction tool so it’s ready for draft season.
Thanks,
Isaac
Hey Isaac
Getting an error using Chrome on Robust and Average when I try and update the table saying “column not found weight In call ‘[”
Do you think its a browser thing?
Thanks Pete, just fixed it!
-Isaac
Hello Isaac,
I was wondering it is possible to have a column that displays a players projected points per dollar that is spent based on the AAV?
exp. LeVeon Bell 309.97pts AAV $53 = 5.85 points for every dollar
Thanks,
Nick
Hey Nick,
Will put this on our to-do list. In the meantime, you can download the data in Excel and create a new column where ROI = points/AAV.
Thanks!
-Isaac
Is there a way that you can include the date that the projections were last updated or published for the various sites, like fantasypros does for their projections? Thanks
Thanks, Chris. This is on our to-do list.
-Isaac
Great site, really like the tool. Regarding analysts weighting in the projections: can you also display when the raw projection values were actually updated by each analyst, and if they projected receptions (some sites don’t)?
I imagine over time, some projections will be badly out of date. Having more info about how/when the projections are sourced will help users with customizing the weights. Thanks.
Yes, this is on our to-do list.
Thanks,
Isaac
When trying to get the projections for my league settings, I am getting an error in the app that reads “object VORrank not found”
Thanks Sean, just fixed it. Let us know if you still run into problems.
-Isaac
Hi Isaac,
I absolutely love all of these analytics, thanks for putting this together. Having just found this site, I’m wondering if traffic on popular draft days (i.e. 8/29) becomes high enough to strain the server causing difficulties utilizing the auction optimizer. Is there a software/ excel version? Thanks!
Hi Scott,
We’ve upgraded our servers this year, so we anticipate being able to handle heavy traffic. We’re working on an offline version, but not sure if it will be ready for this season.
-Isaac
Can “ESPN Pre-Draft Auction Values” be a setting for AAV? I notice the “ESPN” AAV you have are what players are actually being drafted for on average in the espn leagues, which are obviously very helpful. But i know many people in my league might simply be looking at the AAV ESPN automatically assigns to each player. For instance, if I do an espn mock draft, I see many players with AAV values set over $60, but for your app the highest is around $55.
Hey Sean,
Do you know how to get access to pre-draft auction values without joining a draft or mock draft? We’d need a fixed and accessible link to scrape them.
Thanks,
Isaac
Quick question about ADP. Are you generally just taking overall ADP stats from sites over the entire off-season, or is there the ability to take them from REAL drafts as opposed to the larger population of mock drafts? Also, are ADPs aggregated over the entire offseason, or are they taken over a recent period of time (ie the past two weeks)?
I ask because I find the best use of watching these ADP values is to anticipate who is getting over-drafted or under-drafted. But given that the vast majority of drafts up until now have been mocks (or even dynasty), these values could very mis-representative of the actual ADP trends. Some people experiment wildly in mock drafts, and even with training camp injuries it could take a while for reality to overcome previous guesses (ie Foster).
Just wondering. I think that MFL for example does allow to filter ADP by real as opposed to mock drafts.
Regards,
Elliot
Hey Elliot,
Many sources of ADP don’t say how they were calculated (mock vs. real drafts) and over what time frame. We are updating MFL ADP to include only real drafts over the last 2 weeks so that we have enough drafts for a representative ADP estimate, but so it also captures recent trends.
Thanks,
Isaac
The values for ceiling/floor seem awfully tight to the points averages to the point of highly variable and steady guys being very close together in those values. Have you considered loosening up those values to try and show a little more spread for the guys with a wide range of projections vs those who are quite clustered?
Hi Ethan,
I agree. Right now, ceiling and floor are being calculated with a 95% confidence interval around the point estimate (point estimate +/- 1.96 standard deviations divided by the square root of the number of sources). We’ll look into calculating ceiling and floor with different intervals (e.g., based on +/- 1 or 2 standard deviations without dividing by the square root of the number of sources). This would widen the interval estimate.
Thanks,
Isaac
Hi Ethan,
We just modified the calculation of floor and ceiling to be the 10th and 90th percentiles of the projected points across analysts, which makes the interval between ceiling and floor much wider for most players.
Thanks!
-Isaac
Awesome, I think it really helps show the difference between guys like Roddy White and Martavis Bryant now.
I wrote my own program that does pretty much all the same stuff but you have more sources that you are pulling in and I don’t have to fiddle with the data parsing year to year when sites change their layout so this is way easier 🙂
Great Site.
I have a question – I wanted to get your insight on another avenue for fantasy analytics.
Shoot me an email if you get a chance.
lazerlayell@yahoo
Great App! My department is having a fantasy league for the coming season…I know little about NFL but I know a little bit about data analysis and Shiny apps…Will definitely rely on the app for my draft and strategies!
Just some quick questions..is the projection app (http://apps.fantasyfootballanalytics.net/projections) already set up for the coming season? How to set up the AAV and ADP value in the data setting (I am going to play in yahoo fantasy)? What does the weight mean in the data setting? they are not summing up to 1…
Thanks!!
Vincent
Hi Vincent,
Yes, the Projections app is setup for the coming season. You can select Yahoo ADP and AAV. For more info on the tool, including the inputs and how to set weights, see here:
https://fantasyfootballanalytics.net/2013/09/win-your-fantasy-football-snake-draft.html
Hope that helps,
Isaac
Hi Isaac,
Thank you so much for providing this data! I’ve been writing my own site-scrapers for a few years now. I’m so happy to see someone else has done some of the heavy lifting for me.
FYI: I’m finding duplicates of certain players in the exported csv data:
2 2506410,Alex Smith
2 2507596,Marcus Thigpen
4 2531332,Terrelle Pryor
2 2541096,Ifeanyi Momah
2 2543596,Dri Archer
2 2543638,DeAnthony Thomas
2 2553446,Malcolm Johnson
2 2553649,Devin Gardner
I double-checked all of these and confirmed that in each case it’s the same player name, same player id, same team, same position, different projections.
Hi Alan and Jacob,
We are aware of this issue. It’s because those players are projected at multiple positions (even though the tool doesn’t show the different positions). We will fix this after we finish the Auction tool.
Thanks,
Isaac
Are the stat projections from the various sources in your app current for 2015? It projects J Bell significantly ahead of Abdullah and most things I’ve read seem to favor Abdullah. If the projections aren’t current, will they be updated by the end of August or do I have to download the software to get up to date info?
I should have read the comments above. Sorry for wasting your time!
Could you add to the footer maybe, the date of the last stats scrape/pull?
Hi Todd,
Yes, this is on our to-do list after we finish the Auction tool.
Thanks,
Isaac
Can you include custom roster inputs? I use ESPN but but a custom roster. Our flex spot is WR/RB/TE
Hi Chris,
We have custom roster inputs planned for the Auction tool (to be released soon). You can account for custom rosters in the Projections app by changing the VOR Baseline values. For more info, see here:
https://fantasyfootballanalytics.net/2013/09/win-your-fantasy-football-snake-draft.html
Hope that helps,
Isaac
Wow! You are doing this the right way, can see a major future forming here. Furthermore Isaac, your personal attention and accountability are First Class! Had a couple thoughts for functionality improvement:
1. Bye week column inclusion on CSV (simple but helpful piece of data).
2. One could assume or not assume that each expert has TRULY considered strength of schedule for each team/player. I have a fairly strong history of success which influences my excel files post ECR plotting based on this theory. *for example look at first 5 of Ind Colts games (Gore will thrive), or Defenses which AFC N -vs- NFC W will have to share commonly. This will negatively influence on average 8 of the 16 individual games for skill players within these two divisions. Rank each defense on multi measure, designating numerical value of (-.xxx) versus passing, running stats and you may be amazed at year end accuracy. There are always variable of performance there, but no doubt Sea, St.L, provide a greater statistical (-xxx) versus N.O/Atl opposition. This is fairly predictable, but often overlooked.
3. This is HUGE for users with so many different starting compositions out there. Influence consideration of 1/2 QB: 1/2/3 RB: 1/2/3 WR: 1/2 TE: 0/1/2/3 Flex and so on. If you are in a league which incorporates 1 versus 2 QB’s or more commonly a 2 WR league versus a 3 WR league then the weight of each position drastically changes relative to other positions. Would be a great (and potentially difficult) feature to customize to # of starters per position. This can be accomplished, your team seems to have the intelligence to perform this the right way.
I wrote quite a bit here, only out of respect to what you are doing. Your insight is the “correct” way of predicting statistically driven performance. I think that the 3 items mentioned here could help your already amazing direction gain better unique user function along with improving the accuracy of your sites output with schedule influence. This could set you much farther ahead and apart from competitive sites. Thank you for all you are doing, I appreciate it greatly! Dax
Hi Dax,
1) Bye weeks are on our to-do list. Note that it’s my position that you should not draft players based on their bye weeks because you never know how bye weeks will help or hurt your chances. If you have many players with the same bye week, you might be more likely to lost that particular game, but you would be more likely to win all the other games. I would draft the best player available (points/ceiling/floor, risk, dropoff, ADP difference, etc.) given your positional needs.
2) As you point out, projections may already take into account strength of schedule (SOS). It’s also unclear to me how accurate projections of SOS would be. I’ll put this on our to-do list.
3) I like the idea of having configurations for VOR baselines depending on number of teams and lineup requirements. This is on our to-do list.
Thanks for your kind thoughts,
Isaac
I downloaded something before that had a detailed list of projected stat categories, but I can’t find that anymore using the projections app at http://apps.fantasyfootballanalytics.net/projections. It also had all of the projections from the different sources (projectedPts_espn, projectedPts_ffn, projectedPts_nfl, etc.) Am I just looking in the wrong place? Is there another link that provides this?
Thanks!
Tyler
Hi Tyler,
This is on our to-do list.
Thanks,
Isaac
Hi Isaac,
I swear I pulled this info before though. Did you used to have this option and have since changed? I pulled a csv file from somewhere on your site that had the projections broken out and the different stat categories. Am I crazy or did you used to have this as an export?
-Tyler
Hi Tyler,
Last year’s Shiny apps provided projections for stat categories. We’ve migrated away from Shiny to OpenCPU to handle more simultaneous users. Projections for stat categories in the new apps is on our to-do list.
Thanks,
Isaac
Hi Issac,
Thanks for the replies. Appreciate the work you are doing. Are projections something you expect to have this season or is this something that isn’t necessarily high on your to-do list? Also, in the mean time is there anyway to just download the raw data you guys are using to a csv file?
Maybe I’m weird, but when drafting I like to have a sortable list of projections to key in on specific categories when players are close in value. Would be nice to have a robust data set of multiple projection sources.
Thanks,
Tyler
Hi Tyler,
We’re hoping to get to it within the next couple weeks.
-Isaac
Issac,
I love the site. I have been following your work since the first post. I am messing around with the Projections app and I keep getting the below error. Any thoughts?
Thanks!
“An error has occurred:
error in evaluating the argument ‘conn’ in selecting a method for function ‘dbGetQuery’: Error: object ‘siteDB’ not found
In call:
dbGetQuery(siteDB, paste(“SELECT subscription from analysts a join sites s on a”
Am getting the same error here.
Ryan and Red Scott.
This should be fixed now.
How often is the app updated to reflect the change in consensus projections and rankings?
Hi Marc,
We update the data regularly. A “Date Updated” infobox is on our to-do list.
Thanks,
Isaac
that answers my question also. Just discovered the site. Amazing.
Isaac,
Great site, love the transparency of everything you guys produce. Are there any plans to add more league settings into the projections? I’m in a 12 team league that has a offensive player slot which tends to drive prices up for QBs. Would love to be able to adjust for league size and starting position types.
Thanks for all the work,
Tim
Hi Tim,
Our upcoming auction tool will account for league size. For more info, see here:
https://fantasyfootballanalytics.net/2015/08/how-do-auction-values-differ-by-the-number-of-teams-in-your-league.html
We have future plans to allow further customizable AAV but that will likely have to wait until next season.
Thanks,
Isaac
Love the site, and the work put into it. I have a question about the floor/ceiling determinations. I understand these are calculated based on confidence intervals, but the actual projection value varies within the floor/ceiling range. For example, the following are the top two players, per FFA projections, after I enter my custom league data into the app:
1) A. Brown – 317.56 (floor), 355.81 (ceiling), 339.73 (projected) = 58% (percentage above the floor that the projection number falls)
2) L. Bell – 271.21/341.57/307.12 = 51%
Why would one player’s projections fall at a different point within the floor/ceiling range than another? If the answer involves a complex discussion of standard deviations, statistical models, and similar mathematical analyses, please feel free to answer “Stats, man.”
Hi Jeff
Isaac essentially answered this question in a previous comment:
“We just modified the calculation of floor and ceiling to be the 10th and 90th percentiles of the projected points across analysts, which makes the interval between ceiling and floor much wider for most players.”
If you for example have the 10 player projections:
105 106 114 127 132 141 145 146 148 149
The average is 131.3, the floor is 106 and the ceiling is 148. Thus the average here lies 60.24% above the floor, and of course this number will vary from player to player.
I play in a Yahoo league where we reward return yards as 1 pt/20 yards. I noticed that when I input that and a few other changes (7 pts/rushing td) that Jarvis Landry becomes a WR1 (#11 WR), ahead of other WR’s such as Mike Evans, T.Y. Hilton, Emmanuel Sanders, and Deandre Hopkins. This is a boost of 20 spots when compared to the normal settings, which seems a little bit extreme. No one is drafting Landry as a WR1, so I’m not sure what to make of this.
Hi Joe,
Looking at the projections Yahoo has Landry projected at 1225 return yards, Fantasysharks at 705 return yards and Rotowire’s is 1000+ yards, so awarding points for return yards will definitely tilt the value in Landry’s favor.
Dennis
Love this site. I have a draft 8/15/15 this coming Saturday. Any chance I can get a verification that the projections are updated? I would love to use this tool in all my drafts this year. I like the idea of a box showing the last date update, too.
Thanks
Hi Quinn,
Yes, they are up-to-date and we will update them before your draft. A “Date Updated” infobox is on our to-do list.
Thanks,
Isaac
Hi Quinn,
Projections have been updated tonight with the latest updates from the sources.
Dennis
Hey guys, I love the site and I’m trying to use the fantasy app for my draft but noticed that NFL scoring is not available. The closest thing is ESPN, which changing some numbers is not a problem but I also know their flex includes TE’s and the NFL’s doesn’t. Is it possible for you guys to add a NFL scoring option?
We added the scoring rules for the NFL managed leagues
Thanks man just saw it.
Does 4×4 submit PPR projections or only standard?
Hi TAS,
We use their stat projections which are independent of whether it is PPR or standard. You can determine in the scoring settings whether you want to calculate points and ranks based on PPR or standard scoring.
Dennis
I came came across this site when searching “most accurate fantasy football projections”. From the look of things a lot of time and effort has gone into the data. thanks for all your efforts !
My question is….is it as simple as entering my league data at the top of the page and hitting the download? Am I missing a secret access where I need to pay for your services?
Yep, it’s that simple—they’re free!
Cheers!
-Isaac
Using the Rankings/Projections app, how do you get the AAV? I’m assuming it is based on an average of real auctions of 10 team leagues… If so, I can apply some formulas in Excel to change values for my 12 team leagues, but I wanted to make sure it was in fact an average from 10 team leagues… Also is there a setting I can change somewhere that will update it for my 12 team league that I missed?
Thanks again!
Hi Brian,
Yes, the AAV in the Projections App is currently determined by the AAV of real drafts from 10-team leagues. If you want to see the AAV for leagues with different numbers of teams, use the Auction Draft Optimizer:
http://apps.fantasyfootballanalytics.net/lineupoptimizer
It’s not a linear transformation, so it’s not a simple formula you can apply in Excel. For more info how AAV is calculated, see here:
https://fantasyfootballanalytics.net/2013/06/win-your-fantasy-football-auction-draft.html
For more info how AAV is modified according to league size, see here:
https://fantasyfootballanalytics.net/2015/08/how-do-auction-values-differ-by-the-number-of-teams-in-your-league.html
Hope that helps!
-Isaac
Wow I found this site today and the work you guys are doing is amazing. I have two questions for you guys:
1. Sorry if this is asked all the time, but will you guys be releasing a 2015 version of the draft optimizer?
2. Have you looked into as a separate study the use of regression to find the parameters other than projections to find the overall VOR of a player? Explanatory variables such as age, years in the league, team, average defensive opponent, number of players at same position on same team, etc.?
Thanks so much and keep up the good work!
^^^
My bad… For my comment above I meant a snake draft optimizer rather than the auction optimizer that you have posted here.
Thanks
Hi John,
You can use the Snake tool here:
https://fantasyfootballanalytics.net/2013/09/win-your-fantasy-football-snake-draft.html
Not sure what you mean by looking at things beyond projections for examining VOR because VOR is determined by projections (and the projections of the typical replacement). I’d imagine that many of the analysts include the factors you mentioned in their estimation of projections (e.g., age, experience, strength of schedule, etc.), but they are proprietary black boxes, so we don’t know for sure.
Hope that helps,
Isaac
Hi Isaac –
This is great work! Do you plan on uploading updated rstudio project to github so that we can run the optimizer offline?
Hi Jacob,
We do have plans to release an offline tool and to update the repo but I’m not sure we’ll be able to do it before this season starts.
Thanks,
Isaac
Hey guys, your site is revolutionary! I stumbled upon after searching for fantasy draft projections and at first I was a little intimated. You guys simplify some terribly complicated stats (for me at least) that is amazing. So thank you. I was wondering if:
1. there was a way to save settings so you don’t have to input each time you run the app. And 2. I’m in a 2qb, 2rb, 3wr, 1rb/wr, 1te, 1k, 1dst, and full point PPR league. How or should I be adjusting the VOR considering these positional needs?
Thanks again, Eric
Hi Eric,
We have the ability to save settings on our to-do list. In terms of changing the VOR, you can change the VOR based on how many teams are in your league. If there are 10 teams and you want to use the “Worst Starter” as your baseline then set QB baseline at 20; TE, K, and DST baseline at 10. For RB and WR it depends on how you see the Flex position used. If you expect half the teams to use WR in flex and the other half RB then set the RB baseline at 25 and WR baseline at 35. These numbers need to be adjusted if you have a different number of teams in the league.
For PPR, just change the scoring rules to have 1 point per reception for RB, WR and TE.
Hope that helps
– Dennis
Did you lose access to a bunch of sites for their projections? For 2015 the list of sites used is much shorter now than it was earlier this month.
Hi Ethan,
We decided not to include subscription projections.
-Isaac
So theoretically if I draft from my projected rankings down the team I end up with should do well? For example when do I switch to pick a starting TE over say a bench WR or RB
Hi Michael,
See the strategy section here:
https://fantasyfootballanalytics.net/2013/09/win-your-fantasy-football-snake-draft.html
We generally suggest drafting starters before bench players.
Hope that helps,
Isaac
I have a VOR question. How would you modify the VOR based on say leagues with 2QB, or leagues that flex is QB/RB/TE/WR
Hi Nathan,
That depend on what approach you want to take. If you want to use the “Worst Starter” approach in a 12-man 2QB league then the QB baseline would be 24. For the flex positions you can look at draft results from similar leagues and see how many of each positions were taken in the first 100 picks or 10 rounds and set the baselines from that. Alternatively you can try and guesstimate how many teams are using what positions in the flex slot, and add that to your baseline for the respective positions.
Hope that helps
– Dennis
Can the model be updated with projections sometime today? I have a draft coming up tonight and want to use the latest data.
Hi John,
Data were just updated recently.
Thanks,
Isaac
Thanks, U DA MAN!
love the site, isaac.
i’ve searched all over, but i can’t seem to find an answer to this: what do the metrics in the ‘sleeper’ column in the downloaded .csv from the projection app indicate? for example, after making adjustments for my league, i see that ryan fitzpatrick has ‘1’ in the sleeper column and geno smith has ‘2’, and so it goes down to andy dalton at ’20.’ does this mean fitzpatrick has the most or least sleeper potential? or does it mean something else entirely?
thanks!
Hey T Bone,
See here:
https://fantasyfootballanalytics.net/2014/07/identify-fantasy-football-sleepers-app.html#comment-31968
Hope that helps,
Isaac
Hi Isaac,
Can you comment on the risk distribution shift between 2014 and 2015 top 10 players? More specifically, was there a calculation change between the two years or an introduction of younger players with less history?
Here is the data on the top 50 using FF analytics scoring, looks like the risk distribution shifted.
risk 2015 100.0% maximum 9.3886
risk 2015 99.5% 9.3886
risk 2015 97.5% 9.2813775
risk 2015 90.0% 7.73688
risk 2015 75.0% quartile 6.470625
risk 2015 50.0% median 5.64375
risk 2015 25.0% quartile 4.588575
risk 2015 10.0% 3.94787
risk 2015 2.5% 2.9451
risk 2015 0.5% 2.9429
risk 2015 0.0% minimum 2.9429
risk 2015 Mean 5.675304
risk 2015 Std Dev 1.47839205357216
risk 2015 Std Err Mean 0.209076209266636
risk 2015 Upper 95% Mean 6.09545837281508
risk 2015 Lower 95% Mean 5.25514962718492
risk 2015 N 50
risk 2014 100.0% maximum 7.8856
risk 2014 99.5% 7.8856
risk 2014 97.5% 7.869375
risk 2014 90.0% 6.37343
risk 2014 75.0% quartile 4.931775
risk 2014 50.0% median 4.1699
risk 2014 25.0% quartile 3.4504
risk 2014 10.0% 2.90507
risk 2014 2.5% 2.7358825
risk 2014 0.5% 2.7281
risk 2014 0.0% minimum 2.7281
risk 2014 Mean 4.41809
risk 2014 Std Dev 1.31362426433784
risk 2014 Std Err Mean 0.185774525048896
risk 2014 Upper 95% Mean 4.79141788522771
risk 2014 Lower 95% Mean 4.04476211477229
risk 2014 N 50
Hey Joel,
Good catch! Risk in 2014 was calculated using only uncertainty risk. This season, we are combining uncertainty risk with injury risk from Sports Injury Predictor.
Hope that clarifies!
-Isaac
Absolutely, thank you for the clarification.
I have a league where we draft two IDPs (from any position); is there any way to have an option to consolidate the three defensive positions into one? Otherwise the VOR gets screwed up. Not certain how common this league setting is.
Thanks, great tool.
Hi John,
We have plans to allow flex positions to handle this. In the meantime, you can modify the VOR baseline values so that the VOR baseline values across IDP positions correspond to about the same number of points at each position. That way, they have a similar “Typical Replacement”.
Hope that helps,
Isaac
Hey guys, I get this error when I use NFL scoring with the VOR baseline 14,39,48,11,2,6:
An error has occurred:
cannot popen ‘/usr/bin/which ‘uname’ 2>/dev/null’, probable reason ‘Resource temporarily unavailable’
In call:
system(paste(which, shQuote(names[i])), intern = TRUE, ignore
Hi Craig,
The server is having a hard time keeping up with Reddit load at the moment. I suspect it will be working again later today.
-Isaac
I’m looking to try and add footballguys and fantasypro projections back into the calculations, can you give me some guidance on how to determine their associated weights ? I was going to include David D, Bobby H, and Fantasy Pro’s- could you give me the associated weights or advise on how I can figure them out myself based on adding them to the list ? Is there an easier way to do this with excel ?
Hi M,
We’ve calculated weights based on historical accuracy (see here: https://fantasyfootballanalytics.net/2015/02/best-fantasy-football-projections-2015.html). For example, you could use R-squared or 1-MASE.
Hope that helps,
Isaac
Apologies if this is a double post, but it looks like it didn’t go through the first time.
Hi Isaac,
I’ve been looking for something like this for forever; this is great! One question: why is Kelvin Benjamin still projected to be a top 20ish WR? Is that due to the source sites not updating their own projected stats?
Thanks!
Hi John,
The issue with K. Benjamin is that not all sources are updated. CBS and FOX still have him projected at 1000+ yards so he will remain up there in the rankings until those are updated.
Hope that helps,
Isaac
Hey Isaac,
I love the tool (I’m a big R-user and Stats graduate student). Have you considered using some unit-less measure that makes it easy to compare upside potential for the bench players? For instance, the coefficient of variation. Or perhaps, using some measure of skewness to see how much upside potential exists. Just a thought!
Best,
J.T.
Hi J.T.,
We have a couple ideas in mind. The CV would be one possibility. Another would be Ceiling minus Floor. In the meantime, looking at ceiling would give you a good idea of bench players’ upside.
Thanks!
-Isaac
Great website.
Is there a way to tell when the last time the projected data has been updated?
Thanks.
Following up on last question. Do you have set days for updates? Middle to end of week seems to make sense in order to give sites time for their updates after weekend games and injuries. I would think that the heaviest draft times are next two weekends. Love your work!!
Hi H,
We update the data very regularly. We’re updating the data tonight! Will make sure everything is up-to-date for weekend drafts.
Cheers!
-Isaac
Hi Reggae Boyz,
We have plans to include a “Last Updated” infobox soon. We update the data regularly and are updating them again tonight.
Thanks,
Isaac
Thanks Isaac. Keep up the good work.
One question about what values that the tool uses to determine VOR.
I understand the concept of setting the base value based on the number of players drafted at each position (e.g,. 11 QB, 35 RB, etc). But what confuses me is that in the resulting table that is produced by your snake draft there are no players that have a “0” for VOR. Shouldn’t, by definition, the 11th best QB have a “0” value? Same for the 35th best RB? But that doesn’t happen. So either you are not using the 11th QB’s projection as your baseline for a QB’s VOR, or I am not understanding VOR correctly at all (or possibly both).
Hi
If you go to the “About” column in the app, and scroll down to VOR, there is a link to an article on how this is computed:
https://fantasyfootballanalytics.net/2013/04/win-your-snake-draft-calculating-value.html
Here is the relevant paragraph for you:
“To get a more robust estimate of the typical replacement player, I then averaged the number of points from this player at each position with the player of the same position who was ranked 1 above and 1 below the player.”
So VOR is not e.g. the 11th best QB, but the average of QB 10, 11 and 12.
However the app seems to have an error for RBs. With default settings Crowell is currently RB30 and is listed with VOR of 0 and 122.26 points. However he should have a VOR of -1, since the average of RB 29, 30 and 31 is 123.36. Furthermore every other player around him has a VOR value consistent with the replacement value 123.36.
*the average of RB 29, 30 and 31 is 123.26
*consistent with the replacement value 123.26.
The issue you noticed with RBs was fixed with the latest update.
Thanks,
Isaac
Just noticed ECR is using the Fantasypros’ PPR ranks – any chance this could be changed to standard scoring, or an option to select between the three FP options (standard, PPR, half PPR)?
Dynasty ECR would be cool down the road as another option.
Hi John,
We use the ECR that corresponds to your league settings. We use Standard ECR rankings when all reception scoring settings = 0, half PPR when one reception scoring is between 0 and 1, and PPR when one reception scoring >= 1.
I’ll add Dynasty ECR to our to-do list.
Thanks,
Isaac
Ah, I had thought it might be related to the reception setting but when I tested I forgot to change that to 0 for all three RB WR TE positions.
The problem is this league has very weird scoring settings (2 points for every 5 rushing yards, 12 points for TDs, etc) that even though it has 1 point/reception the other settings overwhelm it so much that player value actually resembles standard more than PPR. So the ECR for standard probably matters more to me.
Great site, just sent a donation.
Thanks for your support, John!
-Isaac
Any plans on an update later today (Tuesday)? I have a draft at 630 and am anxious to see how the projections move around with the loss of Nelson (specifically as it relates to Rodgers, Cobb, Adams, et al.). Thanks!
Awesome site & tool! Thanks!!
Having trouble deciding what VOR baselines to use for a 2QB 10 team PPR league. I keep getting DST & K positions near the top. Any suggestions on what to use?
Thanks,
David
Hi David
This is very hard to answer without any info on the scoring settings you’re using. There is no reason K and DST should be near the top, unless your league’s settings have very abnormal scoring for those positions. Just using standard settings and a QB VOR of 20, I get over 10 QBs in the top 50, and no K or DST before 141.
Thanks Isaac! I’ll try it again.
We do a PPR league where 1 pt is awarded for 10 yrd rushing & receiving. TD are worth 6 points. No special DST settings. We have 2QBs, 2RB, 2 WR, 1 TE, 1 RB/WR/TE flex, 1 K, 1 DST. QBs get 1 or for 25 yd passing. Fumbles & Int lose 2 points.
Oops. I meant Eirik! Sorry!
Ryan Matthews seems to have disappeared from the snake draft tool.
Weird… there he is now.
Nope… spoke too soon. Gone again (at least when I change the settings for ESPN standard scoring and non-IDP players)
Hi Michael,
He shows up for me with those settings. Could you double check? Also check to make sure you haven’t excluded him or set a risk tolerance. If he’s still not showing up for you, please give me a detailed list of the steps you take.
Thanks,
Isaac
Could you add a category for first downs? My league awards 1/2 point per first down, passing, rushing or receiving
LOVE THIS SITE!!!
Thanks
Matt
Hi Tesno,
We’ll add this to our to-do list.
Thanks,
Isaac
All,
As we await a “last updated” feature (hopefully for each site scraped) for the benefit of anal-retentive OCD types I checked each set of projections individually for the “Jordy/Davante” test.
I was very pleasantly surprised that ALL of the sites incorporated into the consensus had Jordy Nelson deleted, and Davante Adams elevated many spots, in points and rankings with the exception of Fox and Walter Football (and examining their websites, it’s because they have not updated their projections, not because FFA didn’t scrape them recently. Fox still has Kelvin Benjamin in their rankings, yeesh).
So use the projections with confidence, but you may want to deselect or weight zero Fox and Walter.
FFNerd stands out as being a little odd for only ranking a small number of top players, but it has deleted Nelson.
I noticed the default calculation type is weighted. Should this be Average is that is your most accurate setting?
Hi Chris,
We discuss the accuracy of the different calculation types in the “Interesting Observations” section here:
https://fantasyfootballanalytics.net/2015/07/accuracy-of-fantasy-football-projections-interactive-scatterplot-in-r.html
You’re welcome to use whichever one you like!
Hope that helps,
Isaac
what date’s of expert’s rankings are in your projector? cause some experts I see gave their opinions over a week ago, thanks
Hi Evan,
We update the data very regularly. A “Last Updated” for each source is on our to-do list.
Thanks,
Isaac
Great site. Are there any plans for a weekly lineup optimizer for sites like Fan Duel & Draft Kings? Thanks
Hi Steve,
Yes, we have plans to release weekly projections. Stay tuned!
-Isaac
Any progress on the weekly projections? Great site Isaac!
Have loved using the site and crunching the numbers before my draft this weekend. Is there a way to get the actual average projections along with the projected point totals?
Hi Brian,
We’re working on that, hope to release it soon.
Thanks,
Isaac
“Soon” like this week? My league draft is this weekend, and I want to see what the numbers are that arrived at the point totals. For at least some players, I’m going to likely want to adjust the projection numbers based on how I think a player will actually do. I will also want to try and figure out the reasons why when the rankings for my league settings differ from the rankings from more “standard” scoring (for example, for my league, the tool ranks Jamal Charles as the #1 RB rather than Leveon Bell or Adrian Peterson, albeit by only a point or two).
Love your site and data. Tremendous tool! I have a question and a request.
Am I right in thinking that the sources for your projections have changed from say the spring? I seem to recall seeing Football Guys and others in your list and yesterday when I went to re-run my leagues projections I see that they are not there and others have been added. Since the Guys made your accuracy list I valued their projections being a source of yours??
My request can you add a 60 and 80 yard bonus for rushing yards for running backs?
Thanks,
Brian
Hi Brian,
We decided not to include subscription projections. Do you know any sources that have projections for 60/80 yard bonuses? We’d need to find projections to include them.
Thanks,
Isaac
That’s a shame, especially since FootballGuys are so much on the same page as you when it comes to VBD/VOR.
But of course I get not using subscription projections (would totally understand if there were legal issues).
As long as we can get one more projections update (fingers crossed, this weekend?)
A 60 and 80 rushing yard bonus is highly unusual, so I doubt any site has projections for them.
Hi Isaac,
I’m absolutely amazed by Your apps and all of the site content. They’re very useful. I’m a bit confused regarding ‘league scoring’. Can You please briefly explain the difference between FFAnalytics and NFL league scoring? I’m aware of scoring settings (points for TDs, return yards, etc) but what is league scoring about? I’m preparing my draftboard with Your app for a custom fantasy league on the NFL website. What league scoring do you recommend?
Thanks,
Geza
Hi Geza,
The League Scoring just alters the points for scoring settings (points for TDs, return yards, etc.). You can easily further customize them by typing in new values into the textboxes.
Hope that helps,
Isaac
Thanks for the fast answer. Your user-friendly approach is another advantage of your site 🙂
Real late to bring this up, but I’m in several leagues with DST points for yards allowed, instead of points allowed. Almost all projection sources list yards allowed, so I wondered if this was something you could add to the DST options.
Hi Elliot,
This is on our to-do list.
Thanks,
Isaac
Isaac,
All of my drafts are this weekend so wanted to be using the latest snake optimizer app. Can you let me when your updates will be done? Thanks for all you are doing!!
Hi Henry,
We just recently updated the data.
Thanks,
Isaac
Thanks. Will that be the last time this week?
I assume all of the website projections have projected for Brady based on him missing 4 games. With this morning’s news that he will be playing all 16 games, do you have any estimation of (a) if/when the sites will update their projections on Brady, and (b) when your projections will be updated to reflect that change?
I have a draft on Sunday. While it won’t make or break my draft if I don’t have updated Brady data, I’d like to know the chances of that happening.
Hi Michael,
We just updated the data tonight.
Thanks,
Isaac
Isaac,
Congrats on an excellent site and tool — I am thankful for your hard work!
I’m in a Yahoo league with more or less standard settings, except there are two QBs and no individual defensive players. Do you have any suggestions on setting the VOR baselines? I’m guessing that the default VOR settings would pretty significantly undervalue QBs.
Best,
Patrick
Hi Patrick,
You can raise the VOR for QBs. For instance, you could look at the number of QBs drafted on MyFantasyLeague.com with your settings (or even better, your prior league drafts) within the first Top 100 picks.
Hope that helps,
Isaac
Hi,
Do you plan on posting weekly projections or will this site only update the season long projections?
Hi Andrew
Isaac answered this question in a previous comment on this page.
“Yes, we have plans to release weekly projections. Stay tuned!
-Isaac”
Looking over the auction value app, one function I love about it is the ability to generate “best” teams based on floor, ceiling, and points scored. I’m wondering if there’s a similar functionality available for snake drafts.
Based on your draft slot, the ADP of players, and the number of teams in your league, one might slice the draft into layers of players. For instance, if you draft 5th in a 10 player league, one might roughly predict that the players with the 5th – 15th ADP will be drafted before your pick in round 2. Then the 16th – 24th will go before your 3rd round pick, and so on. Each set of 10 players is a “slice”. As a rough way to determine targets, one could try to assemble starting lineups based on picking one player from each slice.
This also has me wondering if we won’t give a better sense of scarcity and the effects that runs on a position can have. The Dropoff metric is useful, however it only covers three players. Most leagues have 10 or more teams. ADP and league size may give a more useful insight into which players you should select and when.
I should add that the number of players in a slice varies based on your draft position… so in my example slices would either be 9 or 11 players
Hi Ben
I’ve actually done some work on a related idea, essentially simulating managers picking according to ADP to derive the best players to take. I think it works quite well, but some more testing must be done.
I personally use a drop-off based on ADP, i.e. the drop-off from the current best player at a position to the best player at a position with an ADP higher than my next pick. This essentially measures the expected loss in points by waiting a round to pick that position.
-Eirik
Hi Eirik,
That’s exciting! Looking forward to seeing the results. I’ve noticed that the best pick may not be the player with the highest VOR. Pretty interesting stuff.
-Ben
Hi there. I am getting an error when I try to load custom values. Maybe some kind of reset is needed? Hopefully, you can fix it soon as I have a draft in a few hours.
Thanks!
I am also having this problem. Thank you for all the hard work!!!
Hey guys,
See here:
https://fantasyfootballanalytics.net/2015/08/2015-auction-draft-optimizer.html#comment-35121
-Isaac
The app appears to be not working ATM although based on other posts it looks like its a server load issue. Just wanted to give you a heads up! Keep up the good work!
I can’t seem to load any data in any of the apps. The app was fantastic for an auction draft last night. Great work guys!
Works for me, try clearing your cache.
-Isaac
Works for me, try clearing your cache.
-Isaac
Awesome work! =)
Isaac-
Great website, I love the approach!
Like others, I think it will be great if your tool could save a custom scoring setup.
I play a salary cap league on Fireleague.com, and can customize the scoring to get about 90% of how Fireleague scores:
Stat Points
Offense
Passing Yard 1
Rushing Yards 2
Receiving Yards 2
Kick-off Return Yards 1
Punt Return Yards 2
Touchdown 60
Reception 2
Incomplete Pass -2
Fumble Lost -45
Interception Thrown -45
Two Point Conversion 20
Kick-off Return Attempts -10
Kicking
Field Goal Made 1-39 yards 40
Missed Field Goal 1-39 yards -30
Field Goal Made 40-49 yards 60
Field Goal Made 50+ yards 80
Extra Point Made 10
Extra Point Missed -20
Defense
Game Played 200
Point Allowed -10
Defensive Touchdown 60
Fumble Recovered 45
Interception 45
Sack 20
Safety 45
The only things I didn’t see in the options were:
Fireleague differentiates between punt return and kick return yards. Also kick return attempts are tracked.
Kicker:
Missed point after attempts are scored differently from missed field goals.
Missed field goals only count if from 39 yards or less.
DST:
Fireleague defenses start @ 200 points for a game played, and then points allowed against are -10 each.
I just used the <= on 7 point intervals with averages, but I saw the box labeled 'Points Allowed' at the bottom. Is that for just tracking whether the opposition is shut out?
If there's a better way to model Fireleague, please let me know. Thanks for all your hard work!
-Brian
Hi Brian,
The Points Allowed box at the bottom is for assigning (typically negative) points for every point allowed. If there are stat categories you’d like us to add, please find analysts that provide projections for those categories.
Thanks,
Isaac
Thanks Isaac! Setting <=99 equal to 200 and putting -10 on the Points Allowed box gets predictions in line with what I was expecting!
Can't wait until custom scoring schemes are added (or ability to save).
Thanks,
Brian
Saving is on our to-do list. Not sure what you mean by custom scoring schemes. We do have different default configurations/schemes that you can alter.
Are there any sites that do “rest of season” projections as the season progresses besides NumberFire? If so, would it be possible to scrape those and add as an option?
Hi John,
I’m not aware of any. Let us know if you find any.
Thanks,
Isaac
Would be great if this was made for other sports as well. Hockey for me 🙂
Thanks for all of the great information on this site. What day will you typically update the projections app for current week predictions? Thanks again!
They’re up!
Not that any sane person would own him, but your database has never added Jacksonville’s Kicker, Jason Myers
First of all, thank you for putting all of this together. I love the approach, and making all of this free to the public is really cool of you.
I’ve noticed that D’Qwell Jackson (LB for IND) hasn’t been in any of the weekly projections, but he shows up in the overall season projections (and is quite highly ranked for my league). Any way you could look into this? Maybe the apostrophe in his name is effing something up.
Thanks for everything!
Should be working now, thanks!
Hi Isaac,
My results for the Lineup Optimizer are empty/blank. I Cleared cached and browsing data. Are you having any issues on your end? Any suggestions?
Thank you,
Shane
We’re working on it, thanks.
Hi Isaac,
Love your website and the analytics tool. Is there any chance you could produce a similar tool for fantasy hockey? If you don’t have any interest in developing a full tool (which would be fair enough), could you perhaps look at creating some of the R scripts to scrape the data from the various sites which offer projections?
Cheers,
Rhys
I have a full-time job on top of this, so don’t have time to do that unless we get others to help, but you could certainly adapt the scripts for other sports. The FFA community would really appreciate if you shared them with the community. For more info how to do that, see here:
https://fantasyfootballanalytics.net/2014/10/download-run-r-scripts.html
Ok, I’ll see what I can come up with. Just hoping to be able to scrape all the data I can before the season starts so there is a good amount to compare accuracy at the end of the season.
I would also love to get this going for hockey!!
Hey did you get much work done on fantasy hockey tool?
WR Willie Snead of New Orleans never appears in the weekly projections; He’s been getting stats for New Orleans and I know he has projections out there- I think he’s not included in your stat scraper.
Hi there – are there any plans to add actual results to the app for historical lookups? Or does someone know of an easy place to pull it from?
Also would it be possible to add an export to csv for the data table on the app?
Thanks!!
You can download data in .csv here (click Download in top right):
http://apps.fantasyfootballanalytics.net/projections
Actual weekly data are on our to-do list.
Thanks Isaac – great stuff as always. Is there a good or easy source for historical actuals you would recommend if I wanted to pull them. I am trying to pull the points with Fanduel scoring and all I really need is the player and the points by week.
Points will depend on scoring settings. You can pull historical actual stats from Yahoo or ProFootballReference.
Is there a way to save my scoring system so I don’t have to enter it every time? My league uses Myfantasyleague.com (there’s an API here: http://football.myfantasyleague.com/2013/export).
Can you maybe set a cookie so I don’t have to keep doing it?
P.S. I made a donation to your Paypal, great work!
Oops, the API link is http://football.myfantasyleague.com/2015/export – just change 2013 to 2015 🙂
Hi Clayton,
Yes, this is on our-to do list. Thank you very much for your support!
-Isaac
Additional players not being scraped: RBs Jarryd Hayne and Terron Ward, WRs Jordan Norwood and Seth Roberts. Also, K Kyle Brindza’s team should be TB, not DET
How about since we are now one quarter through the NFL season doing an article on how your pre-season projections are doing versus actual results so far? And then do them again at the halfway mark?
Not sure I’ll have the chance to post mid-season, but I do examine the historical accuracy annually.
https://fantasyfootballanalytics.net/2015/02/best-fantasy-football-projections-2015.html
Hey guys – I notice that the Victiv (now Starsdraft) lineup only has 8 positions when they actually use 9. Maybe they used to use 8 but now they have QB, 2 RB, 2 WR, TE, DST, 2 Flex.
Thanks, we’re working on it. If you change the starter requirements in the dialog box then the app seems to be returning the right number of starters. The app does not provide the right number of starters at each position by default.
Also is there a lockdown time for projections. It seems that I am seeing some variations in the projections from a couple of days ago.
We’re updating the projections several times a week. First scrape usually happens Tuesday night, then we try and keep the projections updated through Saturday to help out with lineup decisions. When possible we will update the projections daily.
Are you running an update today? Last update was 12/10 and it is into the evening now on 12/12 in a week with a lot of moving projections due to football news.
Yes, we updated the data today.
Hey Guys,
How do you calculate your floor and ceiling on your projections app?
I figure you may have been using a method like this article you posted to determine it? https://fantasyfootballanalytics.net/2014/07/weekly-variability-simulation.html
However, i’ve noticed that the floor/ceiling for each player are not normally distributed from the mean value you’ve given. Is it the weekly variability for each given player based on variance for statistics (TDs,Yards, Receptions, etc) or the variance between the experts’ projected means?
Floor and ceiling are calculated by the 10th and 90th percentiles of projected points across analysts. For more info, see here:
https://fantasyfootballanalytics.net/2014/06/custom-rankings-and-projections-for-your-league.html
Scrape is still missing Chris Boswell, K PIT
This should be fixed now. Thanks!
Hi guys – is it possible to get salaries as a column on the projections tab? You can see salaries on the lineup optimizer but not the projections link.
Thanks.
We can add this to our to-do list. In the meantime, as you point out, you should be able to get this info from the lineup optimizer.
I second the notion for a salary tab but I just came across this site and I have to say this is the best out there!
You can get salaries here: http://apps.fantasyfootballanalytics.net/lineupoptimizer
Hi Isaac – thanks for making all of this open to the public! So in the Projections App, there are options for previous games in the seasons. Are these projections are the exact same as they were before those games took place? IE: those estimates don’t contain any look-ahead bias, right?
Right, seasonal projections are not modified after the season starts, and weekly projections are not modified after the week’s games start.
Awesome thanks! Also do you aggregate actual realized fantasy points? If not could you recommend a data source for that?
You can download historical actual seasonal fantasy points from our tool. Historical actual weekly fantasy points is on our to-do list. In the meantime, try Yahoo or ProFootballReference.
Hi,
I noticed that in some of the projections, the average point lies outside of the interval given by the floor and ceilings. For instance, the Week 7 DraftKings setting gives the Jets DST as scoring 5.7 points, with a floor of 4.2 and a ceiling of 4.73, which seems to me to be impossible. Also, how are the floors and ceilings worked out? I am trying to best fit a normal curve based on the 3 given points, but being unsymmetrical I don’t really know how to.
oh no! i went to the weekly lineup optimizer and got an error page with a bunch of code saying that i needed to contact the system administrator.
Should be up now!
Thanks for all the help this year. Used your tool exclusively while prepping and this year was one of my safest, most solid drafts. Some suggestions/comments:
1) Comparison on accuracy of robust vs weighted – I’ve used weighted but I’d like to see which was better
2) snake optimizer tool for next year – the one that shows the BPA for each round by position – some players in the pre-draft ranks were super high for their ADP and the snake tool helps with figuring out who you can wait to draft later.
3) incorporating keeper and ROS rankings if possible – only a few sites do these regularly however so I understand that it would be difficult
4) please continue the risk values, those were really helpful with some picks – helped me decide on AP 1st overall. I’d also be interested in seeing how accurate these values were and if there were any trends in regards to age, position, adp etc
Thanks again for all your hard work and for bringing me to the ship!
oh no – having trouble on the projections app site page – which i really have loved using !
Error
An unexpected error has occurred. Please contact the system administrator.
Exception: Packets larger than max_allowed_packet are not allowed.
Controller: projections
Action: Index
Really nice tool here.
I was wondering if and when the projections for the wild card and subsequent playoffs would be on the site.
Yes, we have plans to continue calculating projections through the playoffs. Cheers!
Thanks!
When will they be available?
They are available now!
Another thing I thought of: a lineup optimizer for yahoo/DK/FD would be awesome.
We already have one!
http://apps.fantasyfootballanalytics.net/?app=lineupoptimizer
hey isaac,
rough estimate for when the 2016 projections app will launch?
Thanks!
Hopefully this week!
Great job on this projections tool. I started to use it yesterday, but after the 7/14 update I now get the following error:
An error has occurred:
object ‘projData’ not found
In call:
merge(projData, playerRisk, by = “playerId”, all
To reproduce it go to Change Data Settings and hit Load. It doesn’t matter if you make any settings changes or not.
Should be fixed, thanks!
Isaac – This is an awesome tool but I have a question, why can’t I get the players deleted from the list when I click on their names? It worked for me earlier this year but not now. Thanks,
Shawn
Hmm, it works for me! You might try deleting your cache, disabling ad blockers, and/or trying a different browser.
Is something off with it? I’ve been messing with the settings a million times, and I cant seem to get eddie lacy to show up higher than 30th, but no where on any projection is he so far down. What gives? It also has Lamar Miller extremely low, like 7th, and I cant seem to get that to correct either. Is there an error?
Our default rank for Lacy is 23, which is highly similar to his consensus rank on FantasyPros (24). Our default rank for Miller is higher than his consensus rank on FantasyPros (11).
https://www.fantasypros.com/nfl/rankings/consensus-cheatsheets.php
I absolutely love this app. Being able to include return yardage proj. is giving me a clearer picture of the value for this type of players. It is also giving me a clearer picture on when the value is right to snaq a QB. I’ve always been a value based drafter and this nifty little tool is proving to be invaluable. I’ve been using these results as my rankings except for RC’s, DEF and K and I’m thrilled with the results. An example; I usually don’t grab a QB until rounds 8 or 9, but I’ve been able to grab Rodger, Luck or Brees in the 5th round and get replacement value guys like Lockett and Abdullah. We’ll see if their production is at that 5th round value, but I’m feelling pretty confident.
Hi Isaac,
Love the site!
Here is the issue I noticed:
I downloaded projections for week 1 draftkings fantasy points last Friday (9/9/2016). I then went and downloaded those projections again today, and I noticed that some of the projections were different. I was just curious where the differences come from. Perhaps it’s that the experts are updating their predictions as late as Saturday and you are recomputing projections based on this? If so, it would be interesting to know when in the week your projections are typically final.
Also, a minor thing is that in the past you have included “actual points” in the downloadable projections spreadsheet for past weeks. I noticed that this column contains all missing values when I went back and downloaded week 1 projections.
Thanks!
Hi Erik,
We use different projections from other sites. We calculate projections that are customized to your league settings using the “Wisdom of the Crowd”.
https://fantasyfootballanalytics.net/2014/06/custom-rankings-and-projections-for-your-league.html
We have “actual points” uploaded for prior seasons. We haven’t done that for the current season, yet.
Hope that clarifies!
-Isaac
I can’t find where “actual points” is, could you please help me locate this?
Hi Bryan. We are working adding this data right now, thanks for letting us know!
Isaac,
Just curious…..
How often do players cover their spread when receiving the Q designation for the given week?
James
When I load the Projections App no players show up until I edit the points weighting. Then when I type in a player to compare it doesn’t automatically filter anymore. This is new behavior this week.
Also, my account shows ‘Free account’ even though I’ve paid and subscribed. Tried it numerous times.
The same is happening for me as well.
Your account shows up as active on my end. Have you tried clearing your cache and logging in again? We have more trouble shooting tips here: https://fantasyfootballanalytics.net/about-the-site/faq#appNotLoading.
Ok. I’ll try. What about the other questions?
When I load the projections app, its just blank. Then I have to go and change the point values and it populates. Then when I try to put players into the compare box, nothing happens until I click on, say, accuracy and then back again. Something in the dynamic changing of the players shown has changed.
We will look into this issue, thanks for providing detail.
It’s also definitely still saying “FREE ACCOUNT” next to my email when I login
Getting this error today: ‘connection with pending rows, close resultSet before continuing’
Still having same problem with results not dynamically changing when I add a player to the compare box or first load the screen.
Where can I get actual support with this? Account also still showing Free.
This app doesn’t seem to be working. The numbers are very skewed compared to any of the sources and the gold mining that comes out reflecting it. The player positions are all out of wack.. please help. If you were to try and pull just ESPN source for example, the numbers would not match up with the ESPN sources for that week.
Example as taken from espn’s league projections, week 15 has projections as follows for these players;
David Johnson 20.9
LeSean McCoy 20.1
Le’Veon Bell 19.8
Devonta Freeman 16.8
That’s under standard scoring, week 15 projections. Now if you open the projections App here, and apply ESPN scoring, and make ESPN the only source with a weighted value of 1, that data wont match.
LeSean McCoy (BUF) RB 21.19
— LeVeon Bell (PIT) RB 20.81
— David Johnson (ARI) RB 20.47
— Devonta Freeman (ATL) RB 19.36
Please advise, as I feel something is off with the app. When you release your gold mining, the data looks much closer to what should be expected. Not sure what is skewing this.. but the weekly projections always look off compared to where they should be and the gold mining results, which are on point. I’m essentially trying to extrapolate the gold mining data before you guys release it, using the projections app but it never seems to be right.
Look forward to your reply!
Hi Jesse,
Our data is a snapshot of projections at a given time. It looks like the ESPN projections were updated after our latest data scrape (Sun 11AM CT) so there are some slight differences. For various reasons, we can’t be expected to have an exact copy of individual sources. Plus, the range is more important than the point values themselves, so using our site with only one source defeats the purpose of aggregating for gaining the increased accuracy of the Wisdom of the Crowd.
Hope that helps,
Isaac
Hello Isaac, I appreciate your site very much. Do you have a timetable as to when you will update the adp projections app for 2017 post draft?
Thank you very much.
Jeremy