Which DFS Projections Are Most Accurate?

In this analysis, we dive into the performance of various fantasy football projection sources for daily fantasy sports (DFS). Using Mean Absolute Error to measure accuracy and Coefficient of Variation to measure week-to-week consistency over the last 11 seasons, we explore which sources consistently provide the most accurate and stable weekly projections. We break down the results by position and offer final thoughts and recommendations for which sources DFS players should rely on moving forward.

Methods to Evaluate Projections

To assess the accuracy of DFS projections across various sources, we used the Mean Absolute Error (MAE) as the primary metric for evaluating projection performance. The MAE provides a straightforward and interpretable measure of accuracy by calculating the average absolute difference between projected and actual player performance on a weekly basis.

Because DFS requires making lineup decisions every week, we also evaluated consistency of projection accuracy using the Coefficient of Variation (CV). The CV measures week-to-week volatility by dividing the standard deviation of weekly MAE by the mean MAE, expressed as a percentage. A lower CV indicates more stable, predictable performance from week to week.

The analysis covers the 2015 to 2025 NFL seasons and focuses on weekly projections starting from week 1 through the end of each season. To focus on fantasy relevant players, we limited the analysis to the top 20 quarterbacks and tight ends and the top 40 running backs and wide receivers each week, based on projected points. We evaluated two different sets of projections from Fantasy Football Analytics (FFA):

  • FFA Average: A simple average of projections across all available sources.
  • FFA Weighted Average: A weighted average based on the accuracy of each source in the prior seasons. Sources with lower MAE in previous seasons were given greater weight, ensuring the most accurate projections had a stronger influence. Specifically, weights were calculated as the inverse of each source’s historical MAE (1/MAE), then normalized within each position to sum to one. Each season’s weights are calculated using all available prior data. For instance, 2017 weights were based on 2015–2016 (the only seasons available), while 2025 weights incorporated the full 2015–2024 history. This method prioritizes sources with a strong track record of accuracy, giving them greater influence on the weighted average.

To ensure fair comparison, we only included sources with at least 8 weeks of data per season. Overall accuracy figures represent the average of seasonal MAE values across all qualified seasons. Additionally, we focused our analysis on sources with at least seven seasons of data during the 2015 to 2025 period, ensuring consistency and fairness in the evaluation.

The nine sources included in our analysis were CBS, ESPN, FFA Average, FFA Weighted, FFToday, FantasyPros, FantasySharks, NFL, and NumberFire. Two sources were excluded due to limited data availability: FleaFlicker (4 seasons of data) and Yahoo (4 seasons). This methodology allowed us to evaluate the consistency and accuracy of each source while considering different strategies for combining projections, both simple and weighted.

Who Has the Best Projections?

Quarterback Projections

FantasyPros led QB projections over the full 11 seasons with the best overall accuracy (6.20 MAE). FFA Weighted ranked second (6.22 MAE), followed closely by CBS (6.23 MAE) and FFA Average (6.24 MAE). NumberFire ranked fifth (6.27 MAE), while ESPN posted a 6.38 MAE, slotting in seventh overall. In the most recent three seasons (2023–2025), CBS moved to first place (6.13 MAE), while NumberFire climbed to second (6.20 MAE). FFA Weighted and FFA Average remained competitive, finishing third (6.21 MAE) and fourth (6.22 MAE) respectively. FantasyPros fell to sixth in recent seasons (6.28 MAE), while ESPN remained at the bottom (6.44 MAE). FantasySharks showed slight improvement in recent seasons, moving from sixth overall to fifth in the last three seasons (6.26 MAE).

Running Back Projections

FFA Average led RB projections over the 11-season average with the best overall accuracy (5.20 MAE). FantasyPros ranked second (5.22 MAE), while ESPN (5.23 MAE), NumberFire (5.24 MAE), and FFA Weighted (5.25 MAE) followed closely behind. In recent seasons, FFA projections continued to dominate, with FFA Weighted, FFA Average, and FantasyPros finishing in a virtual three-way tie for first (5.06 MAE each). Both versions of FFA delivered the most accurate RB projections in 2023 (4.78 MAE and 4.79 MAE) and 2024 (5.07 MAE each). In 2025, NumberFire took the top spot (5.27 MAE) with FantasyPros close behind (5.28 MAE), while FFA Average (5.31 MAE) and FFA Weighted (5.32 MAE) finished fifth and sixth. FFToday struggled with RB projections (5.56 MAE overall), and ESPN posted the weakest recent performance at the position (5.31 MAE).

Wide Receiver Projections

FFA Weighted led WR projections over the full 11 seasons with the best overall accuracy (4.94 MAE), closely followed by NumberFire (4.95 MAE) and FantasyPros (4.95 MAE). CBS ranked fourth (5.01 MAE), while FFA Average finished sixth (5.03 MAE). ESPN slotted in seventh (5.05 MAE), with NFL and FFToday rounding out the bottom. In recent seasons, FantasyPros took the top spot (4.84 MAE), delivering the most accurate WR projections from 2023 to 2025. FFA Average and FFA Weighted improved their recent performance, tying for second (4.86 MAE each). NumberFire and FFToday tied for fourth (4.90 MAE). CBS ranked sixth (4.93 MAE), while ESPN (5.18 MAE) struggled.

Tight End Projections

ESPN was the most accurate source for TE projections over its seven qualifying seasons (3.85 MAE), edging out NumberFire (3.86 MAE) in second. FantasyPros ranked third overall (3.90 MAE), while FFA Weighted (3.91 MAE) and FantasySharks (3.92 MAE) followed closely behind. FFA Average finished sixth (3.93 MAE). In recent seasons, ESPN maintained its strong projections (3.71 MAE), with FFA Average moving to second (3.74 MAE) and FFA Weighted third (3.76 MAE). FantasySharks (3.77 MAE) performed well, including the best TE projections in 2024 (3.50 MAE). FantasyPros (3.81 MAE) remained consistent, while FFToday continued to struggle at the position (4.47 MAE overall).

Week to Week Consistency

For DFS players, consistent performance from week to week is just as important as overall accuracy. Volatile projections can lead to poor lineup decisions even if the source is accurate on average. We measured consistency using the Coefficient of Variation, with lower percentages indicating more stable weekly performance.

Quarterback Consistency

FFA Weighted led QB consistency over 11 seasons with the most stable week-to-week performance (16.64% CV). ESPN ranked second overall (16.87% CV), while FantasyPros finished third (17.12% CV). In recent seasons, ESPN improved to first place (15.77% CV), with FFA Weighted finishing a close second (15.81% CV) and FFA Average third (16.06% CV). This demonstrates that FFA projections maintain stable QB performance across the entire season. NumberFire (17.14% CV) and FantasySharks (17.17% CV) showed more volatility in recent seasons, ranking near the bottom for consistency.

Running Back Consistency

NumberFire led RB consistency across all seasons (12.78% CV), followed closely by NFL (12.85% CV) and FFA Average (12.93% CV). FFToday ranked fourth (12.95% CV). In recent seasons, ESPN posted the most stable performance (12.40% CV), while NumberFire held second (12.84% CV). FFA Weighted finished third (13.13% CV), with FFA Average close behind in fourth (13.18% CV). CBS (13.90% CV) consistently ranked near the bottom for RB consistency, while FantasyPros (13.95% CV recent) and FFToday (14.11% CV recent) struggled in the last three seasons.

Wide Receiver Consistency

FFToday led WR consistency over all seasons with the most stable projections (11.74% CV). FFA Weighted ranked second (12.02% CV), closely followed by FFA Average and NFL (both 12.06% CV). In recent seasons, NFL took the top spot (11.80% CV), while FantasySharks moved to second (11.94% CV). FFA Average ranked third in recent seasons (11.99% CV) and FFA Weighted finished fifth (12.26% CV). ESPN (14.12% CV recent) and CBS (14.43% CV recent) struggled with consistency throughout both periods, particularly in recent seasons.

Tight End Consistency

NFL led TE consistency (18.80% CV), followed by CBS (19.06% CV) and FFToday (19.23% CV). TE projections showed more volatility overall compared to other positions, with all sources having higher CV percentages. In recent seasons, FFToday ranked first (22.41% CV), while FantasySharks finished second (22.55% CV). FFA Average ranked third (23.20% CV), while FFA Weighted placed seventh (24.02% CV). ESPN showed by far the most volatility in recent seasons, finishing last (35.50% CV), a dramatic outlier driven by inconsistent week-to-week TE performance.

Variance Explained (R²)

While MAE measures the average size of projection errors, R² tells us how much of the variance in actual player performance is explained by the projections. Higher R² means the projections better capture the week-to-week fluctuations in player scoring.

Across all sources, R² values are notably low, typically between 3% and 23% depending on position. This isn’t a flaw in any particular source, but it reflects the unpredictability of weekly fantasy outcomes among top players. Predicting that a backup running back will score few points is easy. Distinguishing which RB1 will outscore another in a given week is much harder.

R² varies considerably by position. RBs show the highest R² (15–20%), likely because their production is more volume-dependent and workload is somewhat predictable. QBs fall in the middle (4–10%), with significant week-to-week variance driven by game script and matchups. WRs (4–9%) and TEs (3–9%) show the lowest R², reflecting their boom-bust nature and dependence on targets and touchdowns.

FFA projections rank among the top sources for R² at most positions. Over the full 11 seasons, FantasyPros posted the highest R² at QB (8.9%), while NumberFire led at WR (8.9%) and TE (9.0%). FFA Average led at RB (19.1%) and performed consistently well across all positions. Over the last three seasons, FFA Average and FFA Weighted posted the highest R² for RBs (23.3% each), and both ranked among the top sources at every position. The relatively low R² values across all sources reinforce why consistency and smaller error matters because even the best projections explain only about 20% of weekly variance.

Projection Bias by Position

In addition to MAE and R², we measured directional bias using Mean Error (ME), calculated as predicted minus observed. A positive ME indicates the source overprojected (players underperformed their projections on average), while a negative ME means the source underprojected (players outperformed their projections). In the DFS weekly context, the directional biases are small, especially compared to seasonal biases observed in our article here, because weekly projections capture shorter-term expectations where extreme outcomes in either direction partially wash out across weeks.

Quarterbacks show the most consistent negative bias across all sources. Over 11 seasons, the average QB ME across all sources was −0.76 points, meaning the top 20 projected QBs each week slightly outperformed their projections. In recent seasons (2023–2025), this bias has held steady at −0.78 points. The calibration plot reveals that this bias is not uniform across the projection range. QBs with lower projected points tend to outperform their projections by a wider margin, while the highest-projected QBs come much closer to their projected values or slightly underperform them. This pattern, where the low end outperforms more than the high end underperforms, is what produces the overall negative ME at the position. ESPN showed the most negative QB bias (−1.28 overall, −1.38 recent), while FFToday was the most neutral (−0.20 overall). FFA Average posted a moderate QB bias of −0.90 overall and −0.98 in recent seasons.

Running backs show the smallest directional bias of any position. The overall average RB ME was just −0.16 points, with a slight increase to −0.30 in recent seasons. Some sources show near-perfect directional calibration at RB. NumberFire (+0.01 overall) and NFL (+0.09 overall) essentially split the difference between over and under-projection. The calibration analysis shows that while RBs are the best-calibrated position overall (slope of 0.91), the highest-projected RBs still tend to fall short of their projections. The near-zero overall ME reflects the fact that over-projection at the top of the range is largely offset by under-projection at the bottom. ESPN (−0.55 overall, −0.82 recent) showed the most negative RB bias, consistently projecting slightly below actual points. FFA Average was near the average at −0.25 overall and −0.39 recently.

Wide receivers are the only position where some sources historically showed positive bias (under-projection). FFToday posted the most positive WR bias (+1.78 overall), while NFL (+0.63 overall) also tended to under-project WRs. In recent seasons, however, the direction has shifted. All sources now show negative WR bias, with an average of −0.29 across sources. The calibration plot confirms both patterns, lower-projected WRs tend to outperform their projections, while the highest-projected WRs tend to fall short, consistent with the exaggerated spread observed at other positions. The relatively steep calibration slope (0.75) suggests that WR projections overstate the gap between top and bottom options more than any other position. ESPN showed the most negative WR bias (−1.16 recent), while FantasySharks (+0.22 recent) was the only source still slightly under-projecting WRs.

Tight end bias has grown more negative in recent seasons. Historically, the average TE ME was −0.12 points, but in the last three seasons it has fallen to −0.58 points. The calibration slope at TE (0.72) is the lowest of any position, indicating the most exaggerated spread between projected and actual outcomes. Projections overstate the difference between top and bottom tight ends more than at any other position, which is consistent with the touchdown-dependent, high-variance nature of TE scoring. ESPN showed by far the most extreme TE bias (−0.97 overall, −1.89 recent), consistently under-projecting TE production. FFToday was the most neutral historically (+0.77 overall), while CBS (−0.03 overall, −0.05 recent) maintained near-perfect directional calibration at the position. FFA Average posted moderate TE bias at −0.23 overall and −0.67 recently.

Balancing Accuracy and Consistency

The most valuable DFS projection source delivers both accuracy and consistency. To evaluate this balance, we combined each source’s accuracy ranking with its consistency ranking across all four positions. Lower combined scores indicate better overall performance.

Over the last three seasons, FFA Average ranked first when combining accuracy and consistency, with a combined score of 2.9 (2.5 average accuracy rank, 3.2 average consistency rank). FFA Weighted finished second with a score of 3.4 (2.5 in accuracy, 4.2 in consistency). FantasyPros ranked third with a combined score of 4.6, achieving strong accuracy (4.0) with moderate consistency (5.2). NumberFire placed fourth (4.9 combined), while FantasySharks (5.4) balanced middle-tier accuracy (6.2) with strong consistency (4.5). ESPN posted a 5.9 combined score, ranking near the bottom for accuracy (7.0) but showing surprisingly strong consistency (4.8). CBS ranked last among qualifying sources (6.2 combined), with middle-tier accuracy (5.8) but the weakest consistency (6.8).

This analysis demonstrates that FFA projections provide the optimal balance for DFS players. While other sources may rank higher in a given metric for a given position during a given week, FFA consistently delivers top-tier performance in both accuracy and consistency across all positions.

The Value of FFA Projections for DFS

The true strength of FFA projections lies in their consistency across multiple dimensions. While other sources may occasionally produce the best projections for a specific position or week, they often experience significant fluctuations in both accuracy and consistency. For example, ESPN posted strong TE accuracy (3.85 MAE overall, best among all sources) but finished last in recent TE consistency (35.50% CV). NumberFire ranked near the top for WR accuracy (4.95 MAE overall) but struggled with WR consistency in recent seasons (13.08% CV). This kind of volatility can leave DFS players on the wrong side of a source change at the worst time.

In contrast, FFA projections remain reliably near the top in both metrics. Whether using the simple or weighted average, they consistently rank among the top sources across all positions and time periods. FFA Average achieved the best overall accuracy across all positions in recent seasons (4.97 MAE), tying with FFA Weighted, while also placing first in overall consistency (16.11% CV). This dependability is what sets FFA apart for DFS players. The projections may not always claim the number one spot in a single metric, but they don’t experience the drastic swings that other sources do.

A key takeaway from our analysis is that both FFA Average and FFA Weighted excel at balancing accuracy and consistency. FFA Average ranks first when combining both metrics (2.9 combined score), while FFA Weighted ranks second (3.4 combined score). This highlights the power of aggregating projections to minimize risk and volatility.

For running backs specifically, FFA projections have dominated in recent seasons, finishing first and second in accuracy for 2023 (4.78 and 4.79 MAE) and 2024 (5.07 MAE each). While 2025 saw NumberFire and FantasyPros edge ahead at RB, FFA Average (5.31 MAE) and FFA Weighted (5.32 MAE) still finished in the top six. This consistency at the RB position, which carries significant weight in DFS scoring, makes FFA projections particularly valuable for weekly lineup decisions.

By aggregating projections, FFA minimizes the risk of relying on a volatile source, providing a stable and dependable tool for DFS decision-making. Even if an individual source rises to the top in a given season or position, it rarely maintains that performance across all positions or sustains it week to week. FFA projections, however, consistently deliver top-tier accuracy and stability across every position and throughout the entire season. This makes them invaluable for DFS players who need reliable projections they can trust week after week.

In any given week, you can expect FFA projections to perform at or near the top for all positions. No other source can match this level of consistent, cross-positional excellence combined with week-to-week stability.

Interesting Observations

The average of sources is more accurate than individual sources. This is perhaps the most robust finding in our analysis and consistent with the principle of the wisdom of the crowd. FFA Average outperformed individual sources in 63% of head-to-head comparisons across all positions and seasons, finishing with an average rank of 3.7 out of 9 sources. It posted the most top-three finishes of any source (23 out of 43 position-season combinations). The edge is somewhat narrower than what we observe in seasonal projections, likely because weekly projections leave less room for individual source biases to compound over time. Still, aggregating sources remains one of the most reliable strategies for improving weekly projection accuracy.

The weighted average and simple average are essentially interchangeable. FFA Weighted posted a marginally lower overall MAE (5.05) compared to FFA Average (5.07), but FFA Average won 54% of head-to-head position-season comparisons. The two deliver nearly identical performance in the weekly context. This suggests that source accuracy does not persist reliably enough from week to week for historical weighting to provide a consistent edge. In a world where individual source accuracy fluctuates unpredictably, equal weighting may be the safer and simpler bet.

Weekly projections explain far less variance than seasonal projections. Within-position R² values for weekly projections are typically between 3% and 23%, depending on position. Running backs show the highest within-position R² (up to 23% over the last three seasons for FFA projections), while QBs, WRs, and TEs often fall into single digits. This is expected, as weekly outcomes are noisier than seasonal totals, where extreme weeks will partially cancel out over 17 games. The low weekly R² values underscore just how unpredictable individual game outcomes are among top players and reinforce why aggregating sources is especially valuable for DFS.

Projections are more accurate for some positions than others. In terms of raw MAE, tight ends have the lowest error (3.85 for the best source) and quarterbacks have the highest (6.20), but this largely reflects differences in scoring scale. When expressed as relative error (MAE divided by average projected points), quarterbacks show the lowest relative error at 37%, followed by running backs (51%), wide receivers (54%), and tight ends (63%). QBs benefit from a higher scoring baseline that makes projection misses proportionally smaller. Tight ends are the hardest position to project on a relative basis, reflecting their heavy dependence on touchdowns and low target volumes.

Weekly projections are well-calibrated directionally. The average ME across all sources and positions is just −0.10 points per week, meaning projections are nearly perfectly balanced between over and under-projection. If anything, top projected players slightly outperform their weekly projections on average. Quarterbacks show the most consistent negative bias (−0.76 ME), meaning QBs modestly outperform their projections, while wide receivers are the only position with a slight positive historical bias (+0.29 ME). For DFS purposes, the near-zero weekly ME means projection sources are not systematically steering you wrong in either direction.

Wrapping Up

The most reliable DFS projection sources are those that consistently produce accurate projections with stable week-to-week performance across all positions. While some sources may shine in accuracy or consistency individually, FFA projections stand out for delivering the optimal balance of both. Our bias analysis confirms that weekly projections are remarkably well-calibrated directionally, with near-zero average error across all positions and sources. This combination of accuracy, consistency, and structural insight makes FFA projections an excellent tool for DFS players seeking dependable projections that help minimize risk and volatility from week to week. But don't take our word for it. You can use the web app to evaluate historical accuracy for yourself.

Stay Connected with Fantasy Football Analytics
For more data-driven fantasy football analysis and projections, follow us on social media:

Visit FantasyFootballAnalytics.net for our latest tools and articles.