2016 Fantasy Football Projections
264We are releasing our 2016 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 tool here:
http://apps.fantasyfootballanalytics.net
For instructions how to use the app, see here. We also have a Draft Optimizer tool (see here). See our articles on how to win your snake draft and auction draft. We will be updating the projections as the season approaches with more sources of projections. Feel free to add suggestions in the comments!
Share
Issac,
Thanks for the updated projections!
Any chance we’ll get the ability to add in custom anticipated costs in the auction optimizer this year? I know there is the rank cost multiplier, but I was wondering if this could be done on a more granular level. This would helpful in customizing the optimizer to the unique market of each auction.
Thanks for all that you do!
Hey Andy,
I hope we can add that feature–it will mainly depend on our ability to find a reliable front-end web developer:
https://fantasyfootballanalytics.net/2015/10/call-for-help-front-end-javascript-net-web-app-developer.html
Let us know if you or someone else might be able to help.
Thanks!
-Isaac
FYI – The list of players that populate in the “Compare” filter tool do not match the players in the stats database. For example, Todd Gurley (LA) vs. Todd Gurley (STL).
Hi Joe,
Thanks, we’re aware of this issue. It’s on our to-do list.
Thanks!
-Isaac
Thanks, Joe. This issue should be fixed now!
Is there a way to change ADP based on a 2 QB league? Would it be to increase the QB VOR? Thanks, I’m really excited about this tool.
Hey Joey,
Yes, to change player ranks for a 2-QB league in a Snake Draft, you can change the VOR baseline values. For more info, see here:
https://fantasyfootballanalytics.net/2013/04/win-your-snake-draft-calculating-value.html
If you’re doing an auction draft, you can specify multiple QBs in the league settings:
http://apps.fantasyfootballanalytics.net/?app=lineupoptimizer
Note that these won’t affect ADP.
Hope that helps,
Isaac
A few questions:
1. Why is Peyton Manning in the 2016 projections?
2. Is there any way I can export the raw stats projections from the app, and not just the composite score?
3. Can you include targets in your categories? It’s really crucial for my league’s scoring.
4. Have you considered using Adam Harstad’s EVoB in place of VBD for measuring player value?
5. Let me know if you are interested in some formulas for auction pricing.
Hey Mike,
1. I don’t think he is in the 2016 projections. The app doesn’t show the 2016 projections by default currently–you have to manually change the data settings to 2016. We’re working on a fix to show the most recent data by default.
2. Yes, click “Download” in the top right, and then “Raw projections”.
3. What sites have publicly available projections for targets? Please provide the URLs.
4. We do have plans to include VOR per game, which appears to be very similar to EVoB.
5. Yes, we are interested in formulas for auction pricing, especially given the (changing) available pool of players.
Hope that helps,
Isaac
Hey Isaac,
When I run the app for 2016 I keep getting an error: “Type of RHS (‘integer’) must match LHS (‘double’)”. I take it you know this? Also:
1. Do you plan on having VOR per game numbers affect the baselines? In other words, in a 12-team league that starts 2 RB and 2 WR…if the projection is that the top 24 RBs start 312 games and the top 24 WRs start 336 games, will your baselines adjust for what “starting” means (i.e. putting the starting RB baseline lower than starting WR)?
2. It sounds like you plan to include subscription projections in your app this year. Will you publish the names or will it just be “Subscription 1” etc.?
3. How do you adjust for the double-counting that occurs when you include composites (like Fantasy Football Nerd or FantasyPros) in your rankings?
4. ESPN used to project targets but I’m not seeing that right now.
5. Do you have any plans to adjust your weights by position? It seems like that could end up being more accurate than going broad-brush on it.
I will send you my auction values work.
The tools are working for me. How did you generate the error? Please provide all steps you took/changes you made so we can reproduce it.
Thanks!
-Isaac
I agree with Mike. The adjustable baseline is a must which apply to both his point in #1 and #5.
I also have a formulas to adjust auction values based on the players available and the funds available to teams. I am more than happy to share.
Hi Stephen,
Thanks, we’re definitely looking into those issues. Feel free to send your formulas to adjust auction values based on the players available and the funds available to teams.
Thanks,
Isaac
This article addresses how values fluctuate during the draft. Is there a way to email you a spreadsheet I have created that uses your projections to calculate the values as picks are made and recalculates the BUTA (bid up to amount) as players are nominated.
http://www.rotowire.com/basketball/showArticle.htm?id=21393
We have a script to calculate the Bid-Up-To value:
https://fantasyfootballanalytics.net/2013/08/calculating-bid-up-to-values.html
I just checked the box that says “Impute Replacement-Level Points for Missed Games”
Thanks, looking into this. In the meantime, you can generate projections without imputing replacement-level points for missed games.
This should be fixed now, thanks!
It’s Christmas in May… thank you!
I’d love a formula for auction values. My league has a $680 budget and starts 2 QBs… so its difficult to find a values that will work for that. I’m trying to convert SGP into NFL metrics. Which is working but still has some iffy points for positions.
So if you have a formula that works for you… i’d love to hear it.
Other issues with the app:
1. When running 2015 projections for the IDP positions and choosing only NumberFire, the results show that NF projected players up to 40 points. But that makes no sense that they would project the top IDP’s to score 40 points on the season. What kind of error is this?
2. When running 2015 projections for the IDP positions and choosing Walter Football, RTSports, NFL, FoxSports, FantasyFootballNerd, ESPN, EDSFootball, or CBS, I get: “argument ‘env’ is missing with no default”. Does this mean that these sites don’t project the IDP positions? If so, can you have their checkboxes disappear whenever there are no available projections from that source for that position? That will eliminate lots of headaches.
3. It looks like FantasySharks is the one source you have reporting for IDP for 2016. There is an error with the Michael Bennett stats. It looks like you’re giving identical raw projections to both Michael Bennett’s, and they are much too high. Such that both Bennett’s are your top IDP scorer. Looks even to be too high to have been adding them together.
General questions/features:
1. Do you adjust weights by position? Have you found any projectors to be better at some positions than others?
2. Could you possibly include “estimated categories” like Targets, TFL, and QB Hits? These categories are in many scoring systems (like mine), but are often not included in projections. The app could do what I do, which is use a fixed formula for that category based on a historical regression per-position. For example, maybe WR target projection = 1.2*Rec + .012*Yds – 0.5*TD or something, but RB target projection = 1.1*Rec + .006*Yds + 1*TD. I could do some of that math if you like…to come up w/ formulas for these estimated categories. That would make this app the must customizable on the market.
Thanks, Mike. Looking into these.
The Michael Bennett issue is fixed now, thanks!
Isaac, still looks like Michael Bennett is causing troubles. When I run the composite projections, only his Yahoo numbers are being considered. Why is he not registering from Fantasy Sharks and FFToday?
Should be fixed now.
An issue I came across. When exporting 2016 rankings a lot of players have incorrect ceilings. (ex. Reggie Bush had a ceiling of 414.085 points) I can replicate the league settings if needed.
Thanks, will look into it!
The ceiling issue should be fixed now, thanks!
Isaac,
Love the site and the intellectual approach! Couple of quick comments/ suggestions:
1) Would it be possible to show the latest date/ availability of the various analysts projections just so you know who is in/ out of the projections Will additional analysts be available closer to the season like Football Guys (referenced as very accurate in earlier works…)
2) I like the flexibility you give for DST points, Will there be something similar available for yardage? I also wanted to suggest additional stats that at least my league scores like x+ yard TD pass/rush/receiving bonuses, fumbles lost in addition to overall fumbles and fumbles forced. I assume some of this is just time/ focus and some may be inconsistencies of what is provided across the analysts sites.
3) Any thought to putting together a webinar or even a course on how you use R? I am trying to catch up with the different links provided but I think I would benefit from more focused/ direct instruction and would pay for the service.
Thanks again for all you do! I am sure it is not easy to do “sausage making” in such an open and public way and I appreciate your openness to critiques.
Hey Bral,
1) Good suggestion. Many sites don’t list when the projections were last updated. It’s not trivial to check if the data has changed (to know if it’s “updated”), but we can certainly add this to our to-do list.
2) Do you know which publicly available sources include projections for these categories? Could you provide URLs?
3) I love the idea of creating a webinar! We are thinking about creating videos to show people how to (1) use our scripts, (2) learn introductory stats using fantasy football and/or, (3) learn how to do better in fantasy football using advanced stats. Thanks for the suggestion! In the meantime, we have something exciting to release soon that should make the scripts more accessible to more people. Stay tuned!
Thanks!
-Isaac
Any word on whether you will be covertly using any subscription sites in this year’s rankings? The couple extra R^2 points from your analysis got me excited. It also cuts down big time on my manual work.
Also, how are you correcting for the double-counting coming from using projections like FantasyPros, who are just putting together composites of the other web sites you’re using?
Yes, we hope to include subscription sources in the projections. We are working on de-selecting FantasyPros by default, but users can easily de-select them manually.
Thanks,
Isaac
1) Got it, then with it being early I should assume if it is visible for selection/ deselection then there is data available for that analyst? Will additional options become available to select/ deselect as we get closer to the season?
2) An example of available stats is on Fantasy Sharks, which you also use:
# PLAYER TM YDS ALLOWED 0-99 100-199 200-299 300-349 350-399 400-449 450-499 500-549 550+ PTS AGN 0 1-6 7-13 14-17 18-20 21-27 28-34 35-45 46+ SCKS INT FUM DEFTD SAFTS PTS
These are the column headers, hard to see but yardage is a reported stat for DSTs. I will look for other examples.
3) Sign me up!!!!
Great work as always!
http://www.fantasysharks.com/apps/bert/forecasts/projections.php?League=40651&Position=6&scoring=53326&Segment=554&uid=4
Link to see format directly (hopefully!) @ Fantasy Sharks
1) Yes, if it is visible for selection then there is data available for that analyst. We will be adding more analysts as the season approaches.
2 & 3) Thanks!
Could you break out DL and DB by specific positions (DT, DE, CB, S)? All the leagues I am in do this. Also, where do you get your position designations? Do you use NFL, Rotoworld/MFL, or something else?
We get position designations from NFL.com and MyFantasyLeague. We’ll add specific IDP positions to our to-do list.
Is there an option to print out the projections after I have set my league settings?
https://fantasyfootballanalytics.net/about-the-site/faq#print
Michael Bennett issue still persists. You’ll also want to look into the curious cases of Neiron Ball and Jasper Brinkley at LB having strangely high numbers. May be a scraping error at Fantasy Sharks.
Thanks, we’re hoping our next update fixes these issues!
Awesome! When can we be expecting that?
Thanks, we’re looking into these issues. We’re working on some fixes that we hope to release soon!
These issues should be fixed now, thanks!
Why does FantasyFootballNerd projections only include 15 QB’s and TE’s, and doesn’t work for RB and WR?
Re: the top 15:
https://fantasyfootballanalytics.net/about-the-site
You should be able to get projections for RBs and WRs now. You’ll have to change the VOR baselines to be less than 15 to get VORs for them though (because they only publish 15 players publicly available).
Also, for the “Fumbles” column for, say, QB’s…this is actually referring to Fumbles Lost, correct? I think the sites you are scraping from are counting fumbles lost, but that’s an important distinction. Maybe change the language? My leagues count them differently.
Will there be any attempt to distinguish fumbles from fumbles lost?
Yes, this is on our to-do list. Thanks!
I was just looking in depth at your kicker projections. Love how you extrapolate kick distance proportions using Fantasy Sharks in your composite. However, I noticed you don’t include the XPA and (by association) XP Miss data from Fantasy Sharks. Could you scrape that data also, and include it in your scoring options? My league counts XP misses (which are suddenly relevant).
Thanks, on our to-do list.
Also, not sure why it’s happening, but composites by distance aren’t being calculated for Matt Bryant or Nick Folk.
They’re being calculated, just not showing in the downloaded file currently because they’re based on imputed data rather than actual projected stats. But they are calculated for the projected points. Updating the downloaded file to show imputed projections is on our to-do list, thanks.
You asked where you can find publicly-available targets projections. FantasyData.com has top 50 at every position publicly available now. Please add these. https://fantasydata.com/nfl-stats/fantasy-football-weekly-projections.aspx?fs=0&stype=-1&sn=0&scope=0&w=-1&ew=-1&s=&t=0&p=1&st=FantasyPoints&d=1&ls=&live=false&pid=false&minsnaps=4
If you scrape team-by-team at FantasyData.com, you can get all projections for any position, not just top 50.
And finally 4for4 has published their projections free! http://www.4for4.com/fantasy-football/rankings/WR/STANDARD
Any ETA on 4for4 and FantasyData’s projections being included?
We have FantasyData up there (top 50 at each position because it would’ve required too many scrapes to get all players). We’re working on adding subscription projections.
FFToday made their projections live on 5/30. Both offense and IDP.
We added FFtoday projections. Cheers!
Hi Isaac
Not sure if I have missed this, but have you thought about adding a clustering algorithm to the data? That would lead to tiers of players and help in the drop off of targeting players.
I know there is a ‘dropoff’ metric present, but it gives raw numbers and looks at drop-off to ‘next 2’.
Any thoughts?
Yes, that’s on our to-do list, thanks!
The update looks awesome, Isaac! Congrats. I noticed you use MFL ID…is there any chance that you could involve a checkbox feature that is something like “Use position designations from…” and then have different sources? If your app could automatically import and use MFL’s position designations you would be my HEROES! You’d save me hours of having to re-assign positions from different sites’ defaults.
Thanks, will add to our to-do list.
Found a bug: for your Fantasy Sharks IDP data, you are reversing fumbles forced and fumbles recovered. Understandable, since their nomenclature is ambiguous (they use “Fumfrc” to mean forced fumble and just “fum” to mean recovered fumble).
Should be fixed now.
One question and one suggestion.
Question: Why, when I impute for missed games, does Vontaze Burfict’s score not change at all? Yahoo has him missing 4 games. Related to this…can you include “games played” as a column in the raw data export for excel?
Suggestion: I believe Yahoo posts Targets projections. I know most sites don’t do this, so it would be misleading to just impute Yahoo’s targets projections into a composite. What if you used Yahoo’s “Catch Rate” (Receptions divided by targets), then applied this ratio to the composite Receptions number in order to come to a “Consensus Targets”? That’s what I do in my personal methodology and I find it works very well for my scoring system, which counts targets.
Bug: Calais Campbell’s stat line from Yahoo is showing all zeroes.
This is not a bug, this is Yahoo’s projection for Campbell.
Bug: “Art Jones” from Fantasy Sharks is being not identified as FFtoday/Yahoo’s Arthur Jones.
Should be fixed now.
Bug: You have a number of IDP’s missing entirely…likely because of differing position designations on different sites. Ones I noticed:
Brandon Graham
Connor Barwin
Hauoli Kikaha (you have Yahoo only)
Trent Murphy (Yahoo only)
Dante Fowler
Should be fixed now.
First of all, thanks for all your hard work. This site and apps are awesome.
When using the projections app, is there any way to split it out by expert and not just site? Like if I wanted to give more weight to Scott Pianowski of YAHOO! SPORTS and less to Brandon Funston of YAHOO! SPORTS? Also, if I have a subscription to a pay site, is there any way to include their projections into the app, like if I was a subscriber to 4for4? Thanks a lot.
Hi Matthew,
We make all projections available at the lowest possible level (e.g., we don’t aggregate different Yahoo projections). To my knowledge, Funston and Pianowski provide rankings not projections (our app aggregates projections).
Re subscription projections: We do have plans to include subscription projections in the app. Otherwise, feel free to download our projections and merge them manually.
Thanks and hope that helps,
Isaac
Isaac,
Two quick comment/questions:
1) When I customize the projection for ESPN PPR, I am getting a weird result where three NE players are ranked 1-3 in the rankings see below for the first couple of rows:
playername position team playerposition vor points overallECR overallRank positionRank
Rob Gronkowski TE NE TE 313.3698 490.66 7 1 1
Julian Edelman WR NE WR 271.9404 463.787 27 2 1
Dion Lewis RB NE RB 259.0552 378.95 34 3 1
Antonio Brown WR PIT WR 230.679 422.5255 1 4 2
LeVeon Bell RB PIT RB 226.6467 346.5414 3 5 2
Julio Jones WR ATL WR 208.6758 400.5224 2 6 3
David Johnson RB ARI RB 179.6552 299.55 6 7 3
Odell Beckham WR NYG WR 177.7154 369.562 4 8 4
I haven’t checked the rest in detail but I am wondering if there is something specific about the NE players that could be causing this?
2) Bonus points are additive correct? (i.e. A receiver who has a +2/ 100 Yards, +3/ 150 Yards bonus and hits 160 yards receiving gets a +3 bonus not a +5 bonus)
Hey Bral,
We’re not able to replicate this issue. Could you provide us a list of all the changes you make (https://fantasyfootballanalytics.net/about-the-site/faq#incorrectProjection)? We get expected rankings when getting ESPN projections with PPR scoring settings. I’ll check on the calculations of bonuses.
Thanks,
Isaac
Best way to get you the info?
Just post a list of all changes here in a comment. Thanks!
Data Settings
Season 2016, Week Full Season, AAV AVG, ADP ESPN, League Scoring ESPN
Positions 19 QB 44 RB 43 WR 12 TE 2 K 6 DST
VOR Baseline 19 QB 44 RB 43 WR 12 TE 2 K 6 DST
Impute Replacement-Level Points for Missed Games? [No]
Calculation Type Average
Analysts CBS: CBS Average YES , ESPN YES, FantasyFootballNerd YES, FantasyPros NO, FantasySharks YES, FFToday YES, RTSports.com YES, Yahoo Sports Yes
Scoring Settings
QB – Passing Yards (pts/yd) 1 per 25, Passing Attempts 0, Passing Completions .5, Passing Incompletions -.25, Passing TDs 4, Passing Ints -2, Pass of 40+yd 2, Passing 300yd Bonus 2, Passing 350yd Bonus 0, Passing 400yd Bonus 4, Rushing Yards (pts/yd) 1 per 10, Sacks -1, Rushing Attempts 0, Rush of 40+yd2, Rushing TDs 6, Two Points 2, Fumbles -2
RB – Rushing Yards (pts/yd) 1 Per 10, Rushing Attempts 0, Rushing TDs 6, Rush of 40+yd 2, Rushing 100yd Bonus 2, Rushing 150yd Bonus 0, Rushing 200yd Bonus 4, Reception 1, Receiving Yards (pts/yd) 1 Per 10, Receiving TDs 6, Reception of 40+yd 2, Return Yards (pts/yd) 1 Per 25, Return TDs 6, Two Points 2, Fumbles -2
WR – Rushing Yards (pts/yd) 1 Per 10, Rushing Attempts 0, Rushing TDs 6, Rush of 40+yd 2, Receptions 1, Receiving Yards (pts/yd) 1 Per 10, Receiving TDs 6, Reception of 40+yd 2, Receiving 100yd Bonus 2, Receiving 150yd Bonus 0, Receiving 200yd Bonus 4, Return Yards (pts/yd) 1 Per 25, Return TDs 6, Two Points 2, Fumbles -2
TE – Rushing Yards (pts/yd) 1 Per 10, Rushing Attempts 0, Rushing TDs 6, Rush of 40+yd 2, Reception 1, Reception Yards (pts/yd) 1 Per 10, Receiving TDs 6, Reception of 40+yd 2, Receiving 100yd Bonus 2, Receiving 150yd Bonus 0, Receiving 200yd Bonus 4, Return Yards (pts/yd) 1 Per 25, Return TDs 6, Two Points 2, Fumbles -2
K – Extra Points 1, Field Goal 0-19 3, Field Goal 20-29 3, Field Goal 30-39 3, Field Goal 40-49 4, Field Goal 50+ 5, Field Goal Miss -1
DST – Fumble Recovery 2, Interception 2 , Safety 2, Sack 1, Touchdown 6, Blocked Kick 2, Return Yds (pts/yd) 1 Per 25, Threshold Points <=0 ;5 , <=6;4, <=7;3,<=14;1,<=18;0,<=28;-1,<=35;-3,<=46;-5,<=99;-5, Points Allowed 0
Isaac, had more time to look at it and the problem mathematically based on the FFA RawStatProjections info is that certain NE players seem to have unusually high rec40 numbers. Sort the Excel sheet by rec40. RGronkowski has 121, JEdelman has 116, DLewis has 88.7, CHogan has 42.1, MBennett has 31.8, D Amendola has 23.4, MMitchell has 18.4. The next highest from ANYONE in the league Julio Jones is credited with 14.8. I think that’s where the issue is…
With 7/4 update the issue with Rec 40 values for NE players has cropped up again. Thanks again for all you do!
Unfortunately this is an issue with the Yahoo data (see http://football.fantasysports.yahoo.com/f1/52880/players?status=A&pos=WR&cut_type=9&stat1=S_PS_2016&myteam=0&sort=63&sdir=1). Once Yahoo corrects the data then it will be updated in the app.
And you are correct that the bonuses are additive. They represent the number of games that a player reaches that threshold. So if a player is projected to break the 100 yard threshold in 5 games and the 150 yard threshold in 3 games then he will have 2 games with more than 100 but less than 150 yards. If he gets 2 points for breaking 100 yards and 3 points for breaking 150 then he will get 19 points total.
Hope that helps!
-Isaac
Thanks for the clarification!
This should be fixed now. Thanks!
Thanks!
Hey Issac,
Not sure if this is intended or not (or the product of a user error) but the optimizer doesn’t seem to be limiting the cost of its suggested lineup to the cap set for starters. Using the default settings shows $189 under “remaining cap for starters”, but the cost for the lineup with the highest points adds up to $200, leaving nothing for the bench.
Thanks, will look into this!
Hey Mike,
Should be fixed now, thanks!
-Isaac
Looks good! Fantastic website by the way!
Isaac,
FYI, when I am looking at data dump into excel for FFA-CustomRankings noticing that certain players (e.g. DeVante Parker, Zach Miller, Mike Tolbert, Ryan Fitzpatrick, Saints, Chargers, Cowboys) are generating a null values under the overallECR column. Not a big deal for me but wanted you to be aware…
Isaac,
I’m trying to make some projections for this coming season and I need average auction values from last season. Do you know where I can find them?
Thanks
Click “Change Data Settings”:
http://apps.fantasyfootballanalytics.net/
A note for kicker projections:
1. The app doesn’t comb XP missed or XP attempt data yet, even though most sites provide this now. This would be helpful.
2. If you can begin using PFF directly, instead of Yahoo’s version, you can get misses-by-distance data as well.
3. When I run composite kicker projections, I get nonsensical totals. This is because you don’t aggregate the Yahoo make-by-distance data into total FGs converted. So if I combine Fantasy Sharks and Yahoo, the make-by-distance data is an average, but the total makes & misses is just fantasy sharks, so the numbers don’t add up. Can you make the “fg” column for Yahoo to just be a sum of the make-by-distance numbers? And then when running a composite, maybe just use the composite FG% to fill in for the misses (until you use PFF directly)?
#3 should be fixed now. #1 and #2 on our to-do list, thanks.
Also, could you please add Coach and Punter projections? Fantasy Sharks has projections for these.
I’ll add to our to-do list.
Could you add a feature where we can create and save our own “League Scoring” settings?
On our to-do list, thanks.
I second this
Bug
Settings: 2015 / Full Season / AVG / AVG / FFAnalytics
DL position only checked; weighted
FFToday weight set to 1; all others 0.
Scoring settings (in order):
2.74, 1.12, 1.98, 0, 5.49, 3.69, 3.83, 0, 2
I get error message.
We can look into fixing it, but this can be handled by using “average” instead of “weighted” (because there is only one source selected).
This is fixed now
Is your “Accuracy by Position” feature not yet up-and-running for previous years? I notice that when I include multiple defensive positions, the only positional category is “All”
We’re looking into this. In the meantime, if you select all positions in 2015, it will show accuracy for IDP positions separately.
Why is Fantasy Sharks data not available for 2015?
We did not scrape Fantasy Sharks data unfortunately in 2015. We’re including them this year, though.
Lots of players are missing now when I select only Fantasy Sharks for IDP Projections. This is just from the few top players (seems about half are missing).
Aaron Donald
Calais Campbell
William Gholston
DeForest Buckner
Robert Quinn
Jabaal Sheard
Brandon Graham
Navorro Bowman
Denzel Perryman
Lavonte David, etc.
Not just Fantasy Sharks either. FF Today is missing Aaron Donald, Connor Barwin, Robert Quinn, Ezekiel Ansah, William Hayes, and others.
Should be fixed now, thanks.
A couple other suggestions:
1. For injury risk, why not include the average # of Projected Games Played Below 16 as a factor? Many sites project games played and this is their way of assessing injury risk.
2. Where are you getting data on AAV?
3. Why no ECR for defensive players? Also will this be expanded to include both redraft and dynasty ECR?
#1 is accounted for by SportsInjuryPredictor estimates that we include. #2 is selected by the user in “Change Data Settings”. #3 is on our to-do list, thanks.
There’s a few IDP players that show up twice in the rankings, mostly due to the fact they have different positions and therefore different VOR baselines / rankings.
(DB/LB)
Deone Bucannon
(DL/LB)
Khalil Mack
Hauoli Kikaha
Jerry Hughes
Trent Murphy
Vic Beasley
Chris Jones (he’s just there twice as DL)
(Different teams)
Ifo Ekpre-Olomu
Chris Clemons
Brandon Williams
Thanks, Josh. Having a row for each player-position combination is expected behavior for dual-position players (because the VOR will be different for each position). We’ll look into what’s going on with the different team designations.
This should be fixed (Ifo now has 1 row). Jones, Clemons, and Williams each have 2 rows because there are 2 different players with the same name.
Thanks!
-Isaac
Hey Issac,
Is there any way to add 1st downs for offensive players, and 3 and outs forced for defense to the scoring settings? I have a couple leagues that use those. I know yahoo projects those, but not sure if anyone else does. I’m not familiar with R at all, or I would try to build it myself…
Hi Matt,
We can add that to our to-do list.
Thanks!
-Isaac
You have two copies of Devin Smith on your WR rankings.
Should be fixed, thanks!
Duplicate players
Alex Smith 2
Antwan Goodley 2
Chris Harper 2
Connor Cook 2
David Johnson 2
Devin Smith 2
Ryan Griffin 2
Zach Miller 2
We fixed the duplicates in that list. Alex Smith, Johnson, and Griffin are actually different players. Thanks!
Can you include a column for Bye Week in your app?
Yes, on our to-do list. Thanks!
I’m not able to scroll all the way to the right.. To input WR custom scoring.. I’m using chrome.. This site is awesome! Thanks
Sorry it’s not working for you! It’s working for me on Chrome–could you send more details on your setup? OS, Chrome version, etc. You might try disabling ad blockers and deleting your cache.
Hope that helps!
-Isaac
Getting an error.
Baselines:
30 QB
44 RB
59 WR
30 TE
18 K
0 DST
68 DL
78 DB
58 LB
Average; all analysts checked (except FantasyPros)
RB, WR, or TE checked
Rush yds: .2 per 1
Rush att -.5
rush td 4
rec -.75
rec yds .2 per 1
rec td 4
return td 0
two pts 2
fum lost -8
This has been fixed. App wasn’t prepared for scoring with negative points for receptions.
Also seem to be missing De’Vondre Campbell.
Campbell shows up for me. Our spelling is without the punctuation: DeVondre Campbell.
-Isaac
Can you make it possible to view multiple seasons at once on the accuracy tab?
Yes, we can add that to our to-do list. Thanks!
Seems to be an error in Imputing points for missed games. Demarcus Lawrence and Mike Pennel skyrocket to the top of my DL rankings if I check that option.
Should be fixed now, thanks!
Will there be any risk of app downtime due to high traffic? The draft app was unavailable during my draft in 2015 due to what I suspect to be back-end bandwidth issues (response time slowed to a crawl, repeated error messages when recalculating optimal draft lineups, etc.). I was fortunate enough to have an excel model already built but I’m trying to decide whether to invest the time in learning R to avoid some risk.
Thanks for everything!
Hi Darren,
Sorry about the difficulty you had last season. We don’t anticipate downtime this season–the apps are more efficient, and we are using higher-capacity servers. We can’t fully predict how much traffic we’ll have, though!
-Isaac
Two questions:
1. Aren’t you double-counting by default if you include both fantasydata and fantasyfootballnerd? Wouldn’t it be cleanest to uncheck fantasyfootballnerd by default, just like fantasypros?
2. Can you include FantasyData’s past projections in the previous years analysis portion of the app? I’m fascinated to compare their accuracy to other sources, since they use a crowd-based approach themselves. This will be especially powerful if/when you make it possible to just compare “Top 50 players” and such among the different sites.
1) It’s unclear which sources of projections FantasyFootballNerd includes. It’s also not clear to me which (if any) sources FantasyData includes. Regardless, the user can easily uncheck them (or any other sources) if they’d like.
2) Yes, we can add this to our to-do list.
-Isaac
Ok, we modified the default weight of FFN to zero.
Thanks!
-Isaac
Awesome, thanks. Also, are there plans to include FFdata’s full dataset by scraping projections team-by-team rather than just Top 50 by position? I realize that’s significantly more programming work.
It looks like FantasyData doesn’t show all stats at the team-level, so we’d have to run every scrapes for every team-by-position combination (32 x 6 = 192 scrapes) to get the info for the bottom players. We’re probably not going to do that at this point.
Ok, cool. Yeah, I realized it would require 192 scrapes; I just have very little conception for how much work that takes on your end or how it affects app performance. But since they claim to be the world’s only weighted composite…thought it was worth testing the whole set for its accuracy.
Also, wanted to suggest a way to get more accurate composite projections for “rate statistics” such as Tackles for Loss (which are a percentage of solo tackles), Receptions (which are a percentage of targets), and Sacks for a QB (which are a percentage of Pass Attempts). Can you get the app to first compute the composite TFLrate, Catch Rate, and Sack Rate for all sources listed, and then back-apply this rate to come up with a composite?
For example, Yahoo is 20% higher on Alec Ogletree’s tackling prospects this year than other sites. However, when composite is made in your app, it reports his TFL score straight from Yahoo, while still giving a composite number for Solo Tackles. This artificially inflates his TFL rate to be 20% higher than any of the sites likely believe it to be. Instead, can you use a composite TFL rate (which, for now, would just be Yahoo’s), and then apply that rate to the composite solo tackles for a much more accurate composite TFL count? This will be significantly more important for most of your users for Targets and Receptions, though (except in that case the more-reported stat is the smaller number). Presently 3 of your sources project Targets (ESPN, Yahoo, FantasyData). Getting a composite catch-rate from them, according to the weights specified, and then multiplying composite receptions by composite catch rate to get an accurate composite Targets figure would be huge for scoring systems including targets.
This rate issue is also making for some nonsensical totals in composite kicker data (makes + misses don’t equal attempts) when compiling sites that don’t include misses with those that do.
Two bugs:
1. I get an error when selecting kicker projections for just ESPN. All other settings default.
2. FantasyData kicker data has FGmiss at 0 for every kicker, rather than subtracting FGmakes from FG attempts.
Why no Fantasy Sharks projections in the 7/12 update?
You should use ggplotly instead of plotly. It takes the nice graphics you made and makes them interactive. Plotly works very nicely with Shiny and it wouldn’t take long to change your code.
https://plot.ly/ggplot2/getting-started/
Hi Sean,
Thanks for the suggestion—I like the idea of more interactive plots. The apps are OpenCPU (not Shiny), so I’m not sure how easy it is to implement plotly. Would be interested, though, if you have suggestions.
Thanks,
Isaac
Getting an error message w/ the default options selected.
Should be fixed, thanks!
This site is seriously awesome. Where do the average auction values come from? I’m seeing much higher figures on, for example, ESPN.com and fantasypros.com. Also, on the lineup optimizer, the AAVs don’t change when I toggle between the different sources. For NFL.com AAVs, do you utilize figures shown here: http://fantasy.nfl.com/draftcenter? If so, I’m not sure these are reflective of true auction values given I believe they’re an aggregate of AAVs from all mock drafts completed on NFL.com. Most of the auction mock drafts on NFL.com aren’t fully subscribed (i.e. for a 10 team standard mock, very rarely are more than 3 or 4 slots utilized, leading to lower auction values than would be the case in full 10 team drafts). Have you considered including fantasypros.com AAVs? I’m unsure how these are aggregated but it could be another useful reference point. Thanks for the awesome resource.
Hi Luke,
Yes, we get NFL.com AAVs from their mock drafts. Agreed that they’re not perfect, so we give users multiple options for AAV. We can add FantasyPros AAV to our to-do list. The issue should be fixed now so you should be able to toggle between the different sources of AAV.
Cheers!
-Isaac
Awesome. It seems to be working fine now on the projections app, but still doesn’t seem to be working on the lineup optimizer app. Thanks again.
How frequently are you going to update the projections? My league will draft next week on friday and it would be cool to get the latest one. Though I dont want to rush you. It can be seen that this site contains a lot of work in it. By far the best site for draft informations.
https://fantasyfootballanalytics.net/about-the-site/faq#updateOften
We just updated and will plan to update again before then!
Great news thank you!
Hi Isaac,
Is Tyler Lockett missing?
Thanks!
He’s definitely there. Type his name into the “Compare Players” box in the default projections.
He’s definitely there. Just type his name into the “Compare Players” box.
Issac … Saw this online and thought it was quite interesting – any comments ?
http://www.johnmyleswhite.com/notebook/2016/07/23/why-im-not-a-fan-of-r-squared/
Yes, there are many problems with R-squared. That’s why we also examine other measures of accuracy!
https://fantasyfootballanalytics.net/2015/07/accuracy-of-fantasy-football-projections-interactive-scatterplot-in-r.html
Hi Isaac,
How can I upload projections under excel and a web query? While saving the league data?
Also, we have briefly talked fantasy football before. I have a draft tool in excel that I would like to talk to you about.
Mike
I’m not sure what you mean by “upload projections under excel and a web query while saving the league data.” Could you clarify?
Excel allows you to auto import data from websites. However, it looks like the website format is not available for this feature. Is there a basic HTML page or something so that Excel can import the data?
Are you asking a) how to use R to scrape data from websites or b) how to export our projections to Excel?
Excel until I learn have time to learn R
https://fantasyfootballanalytics.net/2016/06/download-projections.html
Question, can I save my data set configurations so that I don’t need to fill it in each time I come back to the site?
Hi Paul,
This is on our to-do list.
Thanks,
Isaac
Any chance you guys can incorporate the STATS projections into the app? They are included in the FantasyPros consensus projections, so they can be ascertained by choosing just STATS and one other projection set that you already have (say, ESPN) and then working backward from this “two-source consensus” to find the STATS projections. They have been the most accurate in the industry, so figured they would be of special interest to you.
I am not seeing any data come through when I check the FantasyPros box as my projections.
One of the things I love about your optimizer app is that it allows you to input league size to adjust a player’s cost. It would be great if you were able to utilize that league size dimension within the projection app. Another solution would be to make the info in the optimizer exportable.
Thanks for your hard work!
Jake
Good point, Jake. Will add to our to-do list.
Thanks!
-Isaac
I know how to download the stats with the directions provided. I’m looking to download the projections from within Excel with a web query.
I don’t know how to do that in excel–we use R for scraping.
I just wanted to say thank you for putting all of this together for free and that all of your efforts are very much appreciated. Let us know if there’s anything we can do to be helpful to you.
Is there a “base” or “default” statistics page? Maybe a .html or .php page? The web query could probably be pulled from that.
Isaac, something seems to have gone amiss with your DB; the projections seem to have reverted to a 7/14/2016 update, highlighted by LeVeon Bell rated #1.
Thanks, we’ll look into this!
-Isaac
Should be working now, thanks!
-Isaac
Hi Issac,
Great job as always! Is there a way for me to access the 7.28.16 update that I saw yesterday?
Tom K.
We’re looking into it, thanks!
-Isaac
Should be up now, thanks!
I am getting the error:
Exception: Unable to connect to any of the specified MySQL hosts.
Controller: projections
Action: Index
at MySql.Data.MySqlClient.NativeDriver.Open()
at MySql.Data.MySqlClient.Driver.Open()
at MySql.Data.MySqlClient.Driver.Create(MySqlConnectionStringBuilder settings)
at MySql.Data.MySqlClient.MySqlPool.GetPooledConnection()
at MySql.Data.MySqlClient.MySqlPool.TryToGetDriver()
at MySql.Data.MySqlClient.MySqlPool.GetConnection()
at MySql.Data.MySqlClient.MySqlConnection.Open()
at ffanalytics.Data.Dapper.BaseRepository.GetConnection[T](Func`2 getData)
at ffanalytics.Data.Dapper.Repository.GetPoints()
at ffanalytics.Controllers.ProjectionsController.Index()
at lambda_method(Closure , ControllerBase , Object[] )
at System.Web.Mvc.ActionMethodDispatcher.Execute(ControllerBase controller, Object[] parameters)
at System.Web.Mvc.ReflectedActionDescriptor.Execute(ControllerContext controllerContext, IDictionary`2 parameters)
at System.Web.Mvc.ControllerActionInvoker.InvokeActionMethod(ControllerContext controllerContext, ActionDescriptor actionDescriptor, IDictionary`2 parameters)
at System.Web.Mvc.Async.AsyncControllerActionInvoker.b__39(IAsyncResult asyncResult, ActionInvocation innerInvokeState)
at System.Web.Mvc.Async.AsyncResultWrapper.WrappedAsyncResult`2.CallEndDelegate(IAsyncResult asyncResult)
at System.Web.Mvc.Async.AsyncResultWrapper.WrappedAsyncResultBase`1.End()
at System.Web.Mvc.Async.AsyncControllerActionInvoker.EndInvokeActionMethod(IAsyncResult asyncResult)
at System.Web.Mvc.Async.AsyncControllerActionInvoker.AsyncInvocationWithFilters.b__3d()
at System.Web.Mvc.Async.AsyncControllerActionInvoker.AsyncInvocationWithFilters.c__DisplayClass46.b__3f()
at System.Web.Mvc.Async.AsyncControllerActionInvoker.c__DisplayClass33.b__32(IAsyncResult asyncResult)
at System.Web.Mvc.Async.AsyncResultWrapper.WrappedAsyncResult`1.CallEndDelegate(IAsyncResult asyncResult)
at System.Web.Mvc.Async.AsyncResultWrapper.WrappedAsyncResultBase`1.End()
at System.Web.Mvc.Async.AsyncControllerActionInvoker.EndInvokeActionMethodWithFilters(IAsyncResult asyncResult)
at System.Web.Mvc.Async.AsyncControllerActionInvoker.c__DisplayClass21.c__DisplayClass2b.b__1c()
at System.Web.Mvc.Async.AsyncControllerActionInvoker.c__DisplayClass21.b__1e(IAsyncResult asyncResult)
I’m not seeing that error. Could you let us know what you did to get the error so we can try to replicate?
Hi Isaac, My understanding is that the risk number for each player is a combo of injury risk and projection risk (standard deviation of projections). I’m having trouble understanding the risk gap between Top 20 WR’s and Top 20 RB’s where WR’s are shown to have higher risk. Particularly in the case of the top WR’s who have several years under their belt vs. RB’s who are rookies/sophomores with little data. In addition, RB’s tend to get injured more and can often be boom or bust. On the other hand, WR’s can be more spotty as there are simply more WR’s in a starting lineup than RB’s. Can you help provide some more insight into the underlying mechanics and the drivers between WR/RB risk? Any insight would be appreciated. FYI – my settings are Yahoo/Standard. -JP
Hi Joel,
Risk is standardized within position, so it’s more meaningful to compare risk values within the same position than across positions. That is, higher risk values within a given position (e.g., RBs) reflect a higher combination of injury risk and uncertainty risk. They are less meaningful to compare across positions.
Hope that helps,
Isaac
Yes that does help and makes sense (blocking by position). Now I have to decide who to pick between RB/WR in the first round…its a tough decision. Last year I went with the “low risk” Marshawn Lynch and a “higher risk” Antonio Brown. On aggregate, statistics are helpful but they cannot prevent first round busts.
Hey Isaac. Thanks so much for doing this!
I’m having an issue downloading the file. It seems to time out. I’ve been trying to download on the Edge browser and will try another one this evening. Have you come across any issues with downloading in certain browsers?
Thanks!
Works for me. Might try a different browser, turning off adblockers, and deleting your cache.
Hope that helps,
Isaac
The projections aren’t loading for me either… An unexpected error…
I’m on a Mac running 10.10.5 and I tried it on Chrome and Safari. Adblocker running and paused on your site in Chrome, not running on Safari…
Thanks!
Error
An unexpected error has occurred. Please contact the system administrator.
Exception: Object reference not set to an instance of an object.
Controller: projections
Action: Index
at Dapper.SqlMapper.GetColumnHash(IDataReader reader)
at Dapper.SqlMapper.d__61`1.MoveNext()
at System.Collections.Generic.List`1..ctor(IEnumerable`1 collection)
at System.Linq.Enumerable.ToList[TSource](IEnumerable`1 source)
at Dapper.SqlMapper.Query[T](IDbConnection cnn, String sql, Object param, IDbTransaction transaction, Boolean buffered, Nullable`1 commandTimeout, Nullable`1 commandType)
at ffanalytics.Data.Dapper.Repository.c.b__18_0(IDbConnection c)
at ffanalytics.Data.Dapper.BaseRepository.GetConnection[T](Func`2 getData)
at ffanalytics.Data.Dapper.Repository.GetPoints()
at ffanalytics.Controllers.ProjectionsController.Index()
at lambda_method(Closure , ControllerBase , Object[] )
at System.Web.Mvc.ActionMethodDispatcher.Execute(ControllerBase controller, Object[] parameters)
at System.Web.Mvc.ReflectedActionDescriptor.Execute(ControllerContext controllerContext, IDictionary`2 parameters)
at System.Web.Mvc.ControllerActionInvoker.InvokeActionMethod(ControllerContext controllerContext, ActionDescriptor actionDescriptor, IDictionary`2 parameters)
at System.Web.Mvc.Async.AsyncControllerActionInvoker.b__39(IAsyncResult asyncResult, ActionInvocation innerInvokeState)
at System.Web.Mvc.Async.AsyncResultWrapper.WrappedAsyncResult`2.CallEndDelegate(IAsyncResult asyncResult)
at System.Web.Mvc.Async.AsyncResultWrapper.WrappedAsyncResultBase`1.End()
at System.Web.Mvc.Async.AsyncControllerActionInvoker.EndInvokeActionMethod(IAsyncResult asyncResult)
at System.Web.Mvc.Async.AsyncControllerActionInvoker.AsyncInvocationWithFilters.b__3d()
at System.Web.Mvc.Async.AsyncControllerActionInvoker.AsyncInvocationWithFilters.c__DisplayClass46.b__3f()
at System.Web.Mvc.Async.AsyncControllerActionInvoker.c__DisplayClass33.b__32(IAsyncResult asyncResult)
at System.Web.Mvc.Async.AsyncResultWrapper.WrappedAsyncResult`1.CallEndDelegate(IAsyncResult asyncResult)
at System.Web.Mvc.Async.AsyncResultWrapper.WrappedAsyncResultBase`1.End()
at System.Web.Mvc.Async.AsyncControllerActionInvoker.EndInvokeActionMethodWithFilters(IAsyncResult asyncResult)
at System.Web.Mvc.Async.AsyncControllerActionInvoker.c__DisplayClass21.c__DisplayClass2b.b__1c()
at System.Web.Mvc.Async.AsyncControllerActionInvoker.c__DisplayClass21.b__1e(IAsyncResult asyncResult)
Apps should be up now!
I get the following when clicking the link http://apps.fantasyfootballanalytics.net/projections
Error
An unexpected error has occurred. Please contact the system administrator.
Exception: No database selected
Controller: projections
Action: Index
at MySql.Data.MySqlClient.MySqlStream.ReadPacket()
at MySql.Data.MySqlClient.NativeDriver.GetResult(Int32& affectedRow, Int64& insertedId)
at MySql.Data.MySqlClient.Driver.GetResult(Int32 statementId, Int32& affectedRows, Int64& insertedId)
at MySql.Data.MySqlClient.Driver.NextResult(Int32 statementId, Boolean force)
at MySql.Data.MySqlClient.MySqlDataReader.NextResult()
at MySql.Data.MySqlClient.MySqlCommand.ExecuteReader(CommandBehavior behavior)
at MySql.Data.MySqlClient.MySqlCommand.ExecuteReader()
at MySql.Data.MySqlClient.MySqlDataReader.ClearKillFlag()
at MySql.Data.MySqlClient.MySqlDataReader.Close()
at MySql.Data.MySqlClient.MySqlDataReader.Dispose()
at Dapper.SqlMapper.d__61`1.m__Finally1()
at Dapper.SqlMapper.d__61`1.System.IDisposable.Dispose()
at Dapper.SqlMapper.d__61`1.MoveNext()
at System.Collections.Generic.List`1..ctor(IEnumerable`1 collection)
at System.Linq.Enumerable.ToList[TSource](IEnumerable`1 source)
at Dapper.SqlMapper.Query[T](IDbConnection cnn, String sql, Object param, IDbTransaction transaction, Boolean buffered, Nullable`1 commandTimeout, Nullable`1 commandType)
at ffanalytics.Data.Dapper.Repository.c.b__18_0(IDbConnection c)
at ffanalytics.Data.Dapper.BaseRepository.GetConnection[T](Func`2 getData)
at ffanalytics.Data.Dapper.Repository.GetPoints()
at ffanalytics.Controllers.ProjectionsController.Index()
at lambda_method(Closure , ControllerBase , Object[] )
at System.Web.Mvc.ActionMethodDispatcher.Execute(ControllerBase controller, Object[] parameters)
at System.Web.Mvc.ReflectedActionDescriptor.Execute(ControllerContext controllerContext, IDictionary`2 parameters)
at System.Web.Mvc.ControllerActionInvoker.InvokeActionMethod(ControllerContext controllerContext, ActionDescriptor actionDescriptor, IDictionary`2 parameters)
at System.Web.Mvc.Async.AsyncControllerActionInvoker.b__39(IAsyncResult asyncResult, ActionInvocation innerInvokeState)
at System.Web.Mvc.Async.AsyncResultWrapper.WrappedAsyncResult`2.CallEndDelegate(IAsyncResult asyncResult)
at System.Web.Mvc.Async.AsyncResultWrapper.WrappedAsyncResultBase`1.End()
at System.Web.Mvc.Async.AsyncControllerActionInvoker.EndInvokeActionMethod(IAsyncResult asyncResult)
at System.Web.Mvc.Async.AsyncControllerActionInvoker.AsyncInvocationWithFilters.b__3d()
at System.Web.Mvc.Async.AsyncControllerActionInvoker.AsyncInvocationWithFilters.c__DisplayClass46.b__3f()
at System.Web.Mvc.Async.AsyncControllerActionInvoker.c__DisplayClass33.b__32(IAsyncResult asyncResult)
at System.Web.Mvc.Async.AsyncResultWrapper.WrappedAsyncResult`1.CallEndDelegate(IAsyncResult asyncResult)
at System.Web.Mvc.Async.AsyncResultWrapper.WrappedAsyncResultBase`1.End()
at System.Web.Mvc.Async.AsyncControllerActionInvoker.EndInvokeActionMethodWithFilters(IAsyncResult asyncResult)
at System.Web.Mvc.Async.AsyncControllerActionInvoker.c__DisplayClass21.c__DisplayClass2b.b__1c()
at System.Web.Mvc.Async.AsyncControllerActionInvoker.c__DisplayClass21.b__1e(IAsyncResult asyncResult)
© 2016 – Fantasy Football Analytics version 1.0p
Apps should be up now!
Same thing for me as Reid & wafflestomper. Im on a mac but I used Safari & Chrome to the same end.
Apps should be up now!
This is my first time finding the site, and unfortunately I’m getting the same as wafflestomper, Reid and Ross. I’ve tried two different browsers (Chrome and Firefox) with the ADP turned off. Please help us if you can.
Apps should be up now!
Edits needed for MFL Draft Only scoring settings:
* Pass yards are 1 per 20 yards, not 1 per 25.
* Return TDs count for 0, not 6.
* Blocked Kicks should have no value for DST.
Thanks, we’ll look into this.
Hi Isaac, it looks like the app is having trouble processing half-point PPRs. Whether I use the app’s default Half PPR scoring or make my own, it returns null for all AAVs and $1 for all Costs. Would you mind taking a look? I loved using your tool last year and would love to use it again for my draft this weekend! Thanks!
Should be updated now, thanks!
I don’t draft til September 3, are you constantly updating projections, etc? and should I wait to download right before my draft?
https://fantasyfootballanalytics.net/about-the-site/faq#updateOften
Isaac, I appreciate and love the app. However, I can’t figure out how to download on a Mac computer. The load button simply opens up a new window in the browser and breaks down everything to raw data. I had the same issue last year, and someone posted a solution, but I am unable to find it. Any suggestions.
What browser? You might try this:
https://www.123contactform.com/docs/i-cannot-download-csv-file-on-safari-what-to-do/
If not, try right click, save as. Or try a different browser.
The AAV do not seem to be close to what ESPN or Yahoo have. When I try to pull them individually they come out as zero.
On another note, these AAVs do not account for changes in league type (# of teams/PPR scoring/roster type).
We’re aware of the AAV issue not updating, and we’re working on it. It does account for # of teams in your league and which source, it’s just not currently updating.
Is there an update on the AAV?
And I also want to confirm, if I change the baseline on a position shouldn’t the AAV change?
Thank you for your hard work @ FFA
Should be fixed now, thanks! VOR Baselines are really only useful for snake drafts (not auction drafts). Changing them won’t modify AAV.
How do you figure that VOR Baselines do not work for Auction values? Where would the auction value come from then? If I had a league with no TE and change the TE Baseline to zero, then all the values for TE’s should be zero
Also, I am still getting all zeros for ESPN and Yahoo AAV
You might try ESPN/Yahoo AAV, we’ve been making updates to the app. You might not be able to get AAV for ESPN/Yahoo PPR/0.5 PPR leagues because they don’t provide them. They only provide them for non-PPR leagues.
We scrape AAV and do not edit the AAV because those are defined as the average cost a player is bought for. If you’re talking about the Bid-Up-To value, that depends on many, many factors. Nevertheless, we account for several factors when calculating how much you should be willing to bid for a player. For more info, see here:
https://fantasyfootballanalytics.net/2013/06/win-your-fantasy-football-auction-draft.html
That was the problem. I was using PPR. You would think as popular as PPR is, they would provide AAV for them.
The only work around I can come up with to have an AAV for a PPR league is to take the relative value from the Standard scoring projections and allocate that across adjusted projections for PPR.
Thanks for the feedback
Hey Isaac. Is Yahoo ADP going to return? I see it’s no longer available in the Data Settings. Yahoo XRank would be useful too. Thanks!
Isaac, I have a basic question: in the projection tool, a player has Average, Ceiling (90 percentile) and Floor (10 percentile).
However, if I understand correctly, each source (e.g. ESPN, Yahoo) counts as one data point. The default number of sources are 8. Would 90 percentile and 10 percentile just 1 ignoring the highest and the lowest data source?
How would that change if I select even fewer sources?
It wouldn’t ignore the highest and lowest source if there are fewer than 10 sources. It would estimate/interpolate the 10th and 90th percentiles based on the distribution of available sources. For example, in R:
> quantile(c(0,1,2,3,4,5,6,7,8,9,10), probs=c(0.1, 0.9))10% 90%
1 9
> quantile(c(0,1,2,3,4,5,6,7), probs=c(0.1, 0.9))
10% 90%
0.7 6.3
For the mathematical formula, see here:
https://stat.ethz.ch/R-manual/R-devel/library/stats/html/quantile.html
hey isaac I think the “Impute Replacement-Level Points for Missed Games” tool is broken with the latest update. When I click it and run the app, the VORs are all messed up and brings all the QBs to the top.
Settings: yahoo scoring, AVG aav and adp
Should be working now, thanks!
Dear FFA,
This site is doing an amazing job providing an alternative to some of the more popular sites thought the web; its so cool to see the application of intellectual data driven methods to fantasy football. Moreover, the articles, such as the one comparing R to ECR, are quite interesting.
But one quick, and perhaps obvious, question: obviously were all trying to get the most up to date information that takes into account current event, such as preseason games, so how often are underlying projections being updated and incorporated into the projection app.
https://fantasyfootballanalytics.net/about-the-site/faq#updateOften
I’m running IDP projections and Chandler Jones is not appearing at all. Do you know what’s up?
Works for me when I type his name in the Projections tool.
I should clarify. When I go to export, he shows up in the custom ranking Excel sheet, but not the Raw Projections.
Same problem occurs with Connor Barwin.
Other players not showing:
Dante Fowler
Brandon Graham
Deone Bucannon
Seems to be happening to players who have had different position designations.
Also Yahoo just did their first update in 2 weeks on 8/22, but it wasn’t included in your 8/22 update. Could you run an 8/23 update to include it?
Could you please start including the Targets column of Yahoo’s projections in your scrapes? It doesn’t show up in the “Raw Projections” export.
Yes, that’s on our to-do list. We’re working on lots of to-do list items right now—if you want targets sooner, feel free to make a pull request! For more info, see here:
https://fantasyfootballanalytics.net/2014/10/download-run-r-scripts.html#shareYourScripts
I keep getting the same error when I try to access the projections. I’ve tried on Chrome, Safari, and Edge browsers. All I’m doing is clicking the link. Here’s the message I see:
Error
An unexpected error has occurred. Please contact the system administrator.
Exception: Timeout expired. The timeout period elapsed prior to completion of the operation or the server is not responding.
Controller: projections
Action: Index
at MySql.Data.MySqlClient.ExceptionInterceptor.Throw(Exception exception)
at MySql.Data.MySqlClient.MySqlConnection.Throw(Exception ex)
at MySql.Data.MySqlClient.MySqlConnection.HandleTimeoutOrThreadAbort(Exception ex)
at MySql.Data.MySqlClient.MySqlCommand.ExecuteReader(CommandBehavior behavior)
at MySql.Data.MySqlClient.MySqlCommand.ExecuteDbDataReader(CommandBehavior behavior)
at System.Data.Common.DbCommand.System.Data.IDbCommand.ExecuteReader(CommandBehavior behavior)
at Dapper.SqlMapper.d__61`1.MoveNext()
at System.Collections.Generic.List`1..ctor(IEnumerable`1 collection)
at System.Linq.Enumerable.ToList[TSource](IEnumerable`1 source)
at Dapper.SqlMapper.Query[T](IDbConnection cnn, String sql, Object param, IDbTransaction transaction, Boolean buffered, Nullable`1 commandTimeout, Nullable`1 commandType)
at ffanalytics.Data.Dapper.Repository.c.b__12_0(IDbConnection c)
at ffanalytics.Data.Dapper.BaseRepository.GetConnection[T](Func`2 getData)
at ffanalytics.Data.Dapper.Repository.GetAllFormOptions()
at ffanalytics.Controllers.ProjectionsController.Index()
at lambda_method(Closure , ControllerBase , Object[] )
at System.Web.Mvc.ActionMethodDispatcher.Execute(ControllerBase controller, Object[] parameters)
at System.Web.Mvc.ReflectedActionDescriptor.Execute(ControllerContext controllerContext, IDictionary`2 parameters)
at System.Web.Mvc.ControllerActionInvoker.InvokeActionMethod(ControllerContext controllerContext, ActionDescriptor actionDescriptor, IDictionary`2 parameters)
at System.Web.Mvc.Async.AsyncControllerActionInvoker.b__39(IAsyncResult asyncResult, ActionInvocation innerInvokeState)
at System.Web.Mvc.Async.AsyncResultWrapper.WrappedAsyncResult`2.CallEndDelegate(IAsyncResult asyncResult)
at System.Web.Mvc.Async.AsyncResultWrapper.WrappedAsyncResultBase`1.End()
at System.Web.Mvc.Async.AsyncControllerActionInvoker.EndInvokeActionMethod(IAsyncResult asyncResult)
at System.Web.Mvc.Async.AsyncControllerActionInvoker.AsyncInvocationWithFilters.b__3d()
at System.Web.Mvc.Async.AsyncControllerActionInvoker.AsyncInvocationWithFilters.c__DisplayClass46.b__3f()
at System.Web.Mvc.Async.AsyncControllerActionInvoker.c__DisplayClass33.b__32(IAsyncResult asyncResult)
at System.Web.Mvc.Async.AsyncResultWrapper.WrappedAsyncResult`1.CallEndDelegate(IAsyncResult asyncResult)
at System.Web.Mvc.Async.AsyncResultWrapper.WrappedAsyncResultBase`1.End()
at System.Web.Mvc.Async.AsyncControllerActionInvoker.EndInvokeActionMethodWithFilters(IAsyncResult asyncResult)
at System.Web.Mvc.Async.AsyncControllerActionInvoker.c__DisplayClass21.c__DisplayClass2b.b__1c()
at System.Web.Mvc.Async.AsyncControllerActionInvoker.c__DisplayClass21.b__1e(IAsyncResult asyncResult)
Isaac,
The projections apps seem to be down. Are you aware of this?
The app was working earlier today, but now appears to be down.
Hi Isaac,
Looks like the projections app is down, sadly. (I’m about to draft. Thanks for building such an awesome thing that I’m experiencing loss without it).
Any idea when you will have the app back in working order? I am getting the same error- thanks!
Everything is down!!! Help us Isaac!!!
Y’all really need to download the projections before your draft. Aside from safety, you can also manipulate the data more quickly during the draft. I like to conditionally format the risk scores as well.
Hey Isaac,
When do you think we will be able to see just week 1 specific projections? Thanks.
https://fantasyfootballanalytics.net/2016/06/download-projections.html#comment-18330
Is there a reason that FantasyPros and FanatsyFootballNerd are by default unchecked in the data settings?
See the Analysts section:
https://fantasyfootballanalytics.net/2014/06/custom-rankings-and-projections-for-your-league.html
App is down for me at the moment
I was just curious what the benefits were to the $4.99/month subscription?
The FFA Insider subscription will allow you to use our Lineup Optimizer tool for DFS leagues as well as optimize lineups for any given week in season long leagues.
I appreciate the response. I actually received an email about it this morning after I had posted. I have already paid and subscribed. I have done very well the last couple years with a huge help from this site. Thank you.
So I have used this tool in the past, and its awesome. Greatly appreciate what you guys do here! I had a few comments though.
1) If I recall, fantasypros has the second most accurate projections compared to this tool. The difference was small? Anyways I checked their projections for PPR and compared them to the tool here and there seem to be some large differences. For instance, Jamaal Charles is #2 points while Devonta Freeman is #5… on your site Gurley comes in at #2, where Charles is #5. They have an “average” of sources, as do you… so while I trust your site more, the differences seem to make a large difference when it comes to VOR? Its small differences…. but it does make a difference.
2) The is my larger thought. In looking at VOR, I have a thought process. When you use a VOR baseline, it create inefficiencies in drafts that I would imagine you want to take advantage of. For instance, when I look through rounds 8, you see ~7-8 TEs come off the board. This puts Gronk’s VOR incredibly high… Showing he should be a 1st round pick. However, I look at this as, I am draft pick 10, which means I get:
10/15
34/39
58/63 etc.
When I calculate the # of each position taken up through pick 39 for instance, I get 1 QB, 14 RB, 21 WR, 2 TE. I use this for my baseline and find out, the VOR of RBs in my first 2 picks is actually MUCH higher because the dropoff from RB 2-14 than from WR 5-21 (these are realistically who I could pick). So this tell me, I should go RB/RB or RB/WR depending on who is open. It also tells me the value of Gronk is less and actually almost gets bumped to a round 3 pick, because his VOR over the #2 TE (Reed) isnt as big of a spread…
So that takes care of picks 1/2
I then calculate the # of each position taken using ADP up through pick 63. This gives me 5 QB 25 RB 30 WR and 3 TE and use this for my “baseline” in determining my picks 3/4. With a VOR of just the firs 8 rounds, normally round 3 tells me Rodgers or Reed are solid bets. But not so when I do this, it tells me that (again) I should take WR/RB depending on who is available.
Then I do this step again for picks <#87. I get 9 QB 33 RB 38 WR and 7 TE. So at THIS point it tells me to pick up TE/QB or QB/TE.
What are your thoughts of doing VOR based on expected # of each position taken prior to your next "round" of picks. Since I am pick 10 and I get 2 picks similar to eachother, I used my later pick of each "jump" which is why I used 15/39/63/87 etc considering ill have virtually 2 picks back to back. I feel this also helps build in some margin for error in case there is a significant run on 1 position, allowing me to use my slightly earlier pick to modify.
I think that overall you get the highest VOR doing this, because VOR broadly doesnt show you the dropoff between say TE 1 and TE3, it merely just shows the dropoff from TE 1 to TE 7/8 which is BIG… but in reality, it seems TE4/5 are where the bigger dropoff happens… which means the value of gronk in the first or second isnt much of a value if you can get say walker in the 6th…
3) Using VOR broadly, do you think it wise to try and create shortages or larger VORs? For instance, I notice in PPR leagues Gronk happens to put up the same expected points as some of the top 10 WRs.. So my thought was, what if you DID draft Gronk in the 1st/2nd round. Then in rounds 3-5 ALSO drafted one of the other TEs and played gronk as a "flex". In theory you lose an optimal number of points, but you also forced a larger shortage on your opponents, where only the ones who picked up TE 2-4 would get a net "gain" on you, however the other 8 teams all lost VOR since now the "worst starter" is TE 13, not 12…. I guess overall this wont make much difference and probably can only be done in PPR formats… my way to excuse this was suggesting that my WRs will each have buy weeks, where TE will only have 1, so I lose 2 weeks of VOR at WR and only 1 at TE, which might moot out the difference I gained…. so its better to pick a WR flex, that can replace my WRs during each buy week affording me less a loss of points. IDK, just something i was pondering.
1) We include more sources than FP and have been more accurate. There will be differences to the extent that our sources differ.
3) Interesting idea. Risky maneuver but it could pay off potentially. Never know how injuries will play out though.
When I load the data it just goes back to the baseline scoring and doesn’t apply any of the changes. How can I change this?
We’d need more info to reproduce. Which tool, what changes, etc?
I’m using the auction draft optimizer, I’m attempting to change scoring to PPR, but nothing changes when I click load data it just goes back to default and disregard any of the parameters I have set. Almost like the whole page is refreshing instead of just the data.
Hi Isaac,
I am having a similar issue as Joe…I am just changing the data setting ‘League Cap’ from 200 (default) to 300. When i press load, the tables show ‘Generating Optimal Lineup’ but, the ‘Remaining Cap for Starters’ resets back to $189 instead of $289. I have tried multiple browsers and tried clearing my cache. Any ideas?
BTW…this site helped me take home the trophy last year…thanks.
Looks like the “Impute Replacement” checkbox is causing issues with the QB VOR values again 🙂
Hmm… looks like the projections page is down… runtime error on load. Anyone else having issues?
Down here too.
The apps are down at the moment due to heavy traffic during peak hours of peak draft season. We’re hoping to generate revenue to cover increased server capacity. In the meantime we recommend a) downloading the projections to use offline, and b) using the apps during non-peak hours.
Hope that helps,
Isaac
Hi Eric,
The apps are down at the moment due to heavy traffic during peak hours of peak draft season. We’re hoping to generate revenue to cover increased server capacity. In the meantime we recommend a) downloading the projections to use offline, and b) using the apps during non-peak hours.
Hope that helps,
Isaac
I’m having the same “runtime error” every time I try to access the projections. It’s been happening since at least 9/3 across multiple browsers. Please help, Isaac. Really appreciate all the work you do.
The apps are down at the moment due to heavy traffic during peak hours of peak draft season. We’re hoping to generate revenue to cover increased server capacity. In the meantime we recommend a) downloading our projections to use offline, and b) using the apps during non-peak hours.
Hope that helps,
Isaac
I’m getting this error for both apps:
Server Error in ‘/’ Application.
Runtime Error
Description: An exception occurred while processing your request. Additionally, another exception occurred while executing the custom error page for the first exception. The request has been terminated.
Apps should be up now. For more info on why they were down, see here:
https://fantasyfootballanalytics.net/app-down-heavy-traffic
App is back up but a few sets of data do not appear in the output: VOR, ECR Rank, ADP, ADP Diff, AAV. Your site/tool is great. Thanks for sharing with the world!
https://fantasyfootballanalytics.net/2016/06/download-projections.html#comment-18505
Is there a way to sort the app by actual statistical categories like TDs or passing yards?
Not currently, but we can add that to our to-do list, thanks!
Yes there is. If you want to sort by projected TD’s, for example, just set the scoring system to TD = 1 pt, and everything else is 0. I’ve done this a number of times with the app, especially to analyze historical projections and answer questions like, “Which source is best at projecting yards? Which at TD’s? Which at completion Pct?, etc.” I wonder if there’s a way that this site could run a comprehensive analysis of this in order to improve their own FFA Composite projection. Perhaps there are statistically significant findings in the historical data that lend themselves to a stat-by-stat composite from various sources.