The question of SoS is one of the topics I've tried to understand-- obsessively.
It's up there with "Where does each Position get most 'Upside' in the draft?" And "When does it pay off to stream QB?"

What's Different This Time?

It's hard to judge SoS the "right" way, because you need 3 hard-to-get things :

1. Blind multi-week predictions. Points-allowed forecasts for the 17 weeks-- as seen "from pre-season". We want quality ones, using predictive modeling.

2. Expected 'Baseline' Scores, on a weekly basis. The expected team RB fantasy scores, if they were going up against an "average" opponent. We need to represent the team's "average RB score"-- as seen in the moment of each week. Same for WR and TE.

3. Multiple seasons. All the above.

Here's a summary of methods I've published before-- and why they weren't enough:
[Link to 2025 Reddit Article]

But these studies got us pretty far!

1. We understood there must be SOME positive benefit of SoS from pre-season.

2. We understood that DURING the season, the charts definitely work.

To analyze things more-correctly I created new "multi-week ahead" predictive models.
These can simulate "what points-allowed projections would have been" for all 17 future weeks as seen from week 1. I prepared these for 11 seasons, 2015 - 2025.

Methodology

This part's simple! The hard part is done-- generating the models and the baseline expectations. Now we just measure the correlations between two types of numbers:

1. Actual points above expectation. (Actual data of points-allowed to RBs each week) minus (the opposing RB's "expected average" score, at that week)

2. Expected fantasy points difference. These are the predictive models' forecasted "team points above expectation"-- forecasted from before week 1.

Results

I'm choosing to depict a "cartoon-y" graph, that cuts through random jumps in the patterns.
The data show there are some real trends you can expect, on average, when trying to forecast different weeks.

TLDR: Overall, we're looking at LOW PREDICTABILITY, when trying to forecast Strength of Schedule FROM PRE-SEASON. The overall level is somewhere between Betting Line gambles and Kicker predictions. It's not useless, but it's also not reliable.

My takeaways:

• Tight End. Not reliable.

• Wide Receiver. There might be some small advantage of targeting WRs with a good, early schedule. The trend is still better than a coin flip, and you might increase chances of finding a breakout.

• Running Backs. I'm surprised as anyone, but it does seem RB's can be viewed with an eye on their Playoffs schedules. But the entire mid-season is unpredictable for RBs.

It doesn't matter for me-- you can freely use website's special predictive tables, to supplement your draft if you want.

This is a copy of the 2025 Study of Strength of Schedule.

It is in the process of being updated for 2026 data and for the new, more definitive analysis.

[No-- These SoS charts are NOT integrated into my free TapThatDraft cheat sheets. SoS charts are an EXTRA tool you can use FREELY, as a Draft Supplement. Meanwhile, I hope everyone has got the chance to try the customized 1-pager "TapSheets"-- or watch the VIDEO answering FAQs! ]

Fresh new analysis! I think you'll enjoy this fresh take on SoS. And other analyses haven't covered it quite like this, from different angles. I wanted to get as close as I personally could to "the definitive true answer to whether SoS is useful / useless, and by how much". And I really want it to be understandable, because I feel like we've been lacking numbers we can see, feel, and conclude on. We're gonna try and get to the root of this fundamental question. I hope you're ready to take the dive with me!

The final section here has the concrete assessment of last year's "real" table and "real" results.

Introduction

Strength of Schedule (SoS) is the forecasted difficulty for a team or player, for future weeks, especially with consideration for the prowess of weekly opponent defenses. But the usefulness of "Strength of Schedule" is hotly debated.

I wanted to challenge the full range of Opinions:

1. "SoS is one more strategic way to gain an edge."

2. "SoS is useful later in the season, but useless before week 1."

3. "Unreliable fluff, scam, clickbait."

As usual, my task is to dig into the data, to help you understand what historical results actually say. Here's what I'll cover:

1. My RB / WR / TE Charts exist.

2. Year-on-Year changes to teams' Fantasy Points Allowed

3. Week-on-Week accuracy of (my) models, and how accuracy improves as the season progresses.

4. How accurate did my ACTUAL 2024 SoS forecast turn out to be?

TL;DR

Strength-of-Schedule predictions, for individual weeks as seen before week 1, are usually correct about 60% of the time, therefore a 40% chance of wrong prediction. (This only applies to my predictive models, not usual "fantasy points allowed" type charts.) Even though that might seem low, you should recognize that it's similar to many other rates of predictability involved when you play Fantasy Football. That doesn't mean it's reliable, it just means it's significantly better than a coin-flip, if you're trying to decide between two players.

Free RB/WR/TE Charts for the Draft

These "17-week charts" are free and open for you to use, for the draft! (And for the next 4 weeks too.)
We sadly ran out of space to include them in my minimalist TapThatDraft cheat sheets. But that shouldn't stop you from referring to them:

• RB : 0ppr | 0.5ppr | 1ppr

• WR: 0ppr | 0.5ppr | 1ppr

• TE : 0ppr | 0.5ppr | 1ppr

(And yes-- somebody always asks-- of course these update all season, after absorbing each week of new information.)

The main thing I want you to know is: These aren't your Daddy's "fantasy points allowed" charts. And they're not even your uncle's "Adjusted- fantasy points allowed" charts. I specifically made these to perform much more reliably (because I think those single-dimensional charts are kinda dumb). There's a longer description from 2 years ago.

Instead, they're all-original, from-scratch predictive models ("guided AI"), which goes beyond just fantasy points and basic adjustments. They carefully calibrate all kinds of inputs: not just passing/rushing yards, but also including things like the opposing defense pace of game, or what day of week, or whether they've had a bye, or had multiple weeks on the road, etc. They're also heavily dependent on my NFL game score forecasts, which work the same way. Just like all my models (D/ST, Kicker, QB, etc.), I've tested hundreds of parameters, and I only keep the dozen or so that survive meticulous cross-validation.

I update for off-season changes as well. Having said that, you're always welcome to help out if you notice the offensive "baseline" numbers seem off due to roster + coaching changes.

Year-On-Year Change in Team Defense "Points-Allowed"

Year-on-year team differences can tell us how consistent/inconsistent metrics are. IMO, this is a quick 'n' dirty analysis: it's one-dimensional, and it doesn't uncover how much you can trust the prediction of individual weeks. So I'm just linking here to last year's post, which displayed it well.

TL;DR correlation coefficients for the 3 positions: 0.30 for vsRB, 0.24 for vsWR, 0.13 for vsTE.

Week-on-week Accuracy throughout the Season

OK, this is where things start to get cool! Because what we really care about are the weekly matchups, not necessarily the full season. I've never seen anything like this breakdown before, and I'm excited to show it because it seems to reveal some guidelines.

In short, we're looking at my predictive models, and we're measuring how correlated the predictions are to the actual fantasy point results. I'm re-simulating what historical predictions would have been for 0-PPR (in this case) for the 11 years 2014-2024.

Just one modification: My models are trained to predict "total fantasy points allowed", which means the numbers usually include the offensive strength of the position we're looking at. Like, the RB number will be bigger when the offense is the Lions. We just need to delete that out, to evaluate true Week-on-Week effectiveness of SoS-- evaluating mostly the defense only. So just like my SoS 17-week charts, I've subtracted that "baseline" to make this analysis.

Here's how the weekly correlations of matchup-strength look for each position:

There are 2 observations I think are surprising and worth discussing:

1. Correlations are usually close to 0.20 (TEs lower at near 0.15). This is about the same level of predictivity in trying to choose a kicker each week.

   1. In rough terms, this corresponds to guessing Strength-of-Opponent correctly 60% of the time.

   2. Of course, you will feel and remember the other 40% of the time that you're wrong.

2. Accuracy doesn't clearly improve over the season.

   1. I suspect this is due to (1) D-line injuries that are not in the simulation (but which I do include in my charts through the season). (2) Probably something about changing offensive schemes.

   2. Note: total fantasy point predictions do get clearly more predictable over the season. It's only when we look at predicted deviations that the clear positive trend goes away.

What Actually Happened after the 2024 SoS Forecast

To top it all off, I want to give a really concrete example. Data you can feel. So I looked at last season, to see what matchup-adjustments I'd projected, and then compare against what real results were. By position, for every team and every week. I'm not re-running any simulations here, I'm just using the original SoS tables, as published from when we started last September. I wanted to know how good these outlooks were from DRAFT DAY. So... what happened afterward?

The tables below show the results of all 17 weeks. The weeks which were "Wrong" predictions are all "x"s. For the "Correct" predictions, the table shows the actual number of fantasy points that the defense allowed above the baseline (baseline production of the offensive position). So negative numbers mean the defense allowed fewer points than average.

These charts let us count how many "correct" predictions were made each week. And those percentages are plotted below each chart. On average the pre-season view ended up with these success rates over the season.

• RB Matchup accuracy: 61% correct weeks

• WR Matchup accuracy: 58% correct weeks

• TE Matchup accuracy: 57% correct weeks

Of course, these would have been better if we looked at the tables as they updated each week, to incorporate new data.

Finally, I have labeled for you, in the tables above, how right / wrong it was for each team. For examples,

• the Chargers RBs apparently faced an incredibly predictable strength of schedule (88%),

• and in contrast, the pre-season model apparently flunked in predicting for Bengals WRs (31%).

(Just for those who are nerdiness-inclined, there was one assumption I had to make, in order to make any sense of the above charts. When do we label a prediction "right" or "wrong"? If the prediction and outcome were both the "same sign"? I say no, because if the outcome was "-20" and the prediction was "-0.5", that's too far apart to say I predicted such a large swing. But if the outcome is "5" and the prediction was "-0.5", then yeah that's pretty close-- even though the signs are opposite. I overcame these by mapping the quantiles of projections to the quantiles of outcomes By calibrating against the correlations found in the earlier section, I could see it made most sense to divide the number of data points by 3 (think "low / medium / high"). So for the 512 data points, a projection is called "correct" if it is within 170 ranks of the outcome. I hope it's intuitive that this is close to the "same sign" idea, but improves upon it by giving a sliding range that better covers the "non-extreme" region.)

What's Different This Time?

There are different ways to set up the question.
And most of methods don't REALLY get to an indisputable answer.

If you want to judge SoS the "right" way, it requires 3 things hard to get:

1. Blind multi-week predictions. Points-allowed forecasts for the 17 weeks, as seen "from pre-season". More specifically: QUALITY points-allowed projections, using predictive modeling.

2. Expected 'Baseline' Scores, on a weekly basis. The expected team RB fantasy scores, if they were going up against an "average" opponent. (You'll want to subtract these values from the actual outcome data, to get the deviation.) The expected scores need to represent the team's "average RB score", updated for the best information of each week. Same for WR and TE.

3. Multiple seasons. All the above.

Conclusion

Once again, it's up to you if you consider the "60% rule of thumb" an interesting gamble to play, as part of your weekly strategy or draft strategy. I'm not here to sell the idea of Strength of Schedule to you, I'm just trying to put out as accurate and true numbers as I can! But I hope it makes sense to you that you're making a 60% gamble, which is really similar to the levels of uncertainty that you're making in fantasy football all the time. I also need to warn again, the conclusions would be worse if we didn't use these highly optimized models. Meaning plain old adjusted-historical-fantasy-points aren't gonna make the cut!

Thanks for reading, and Good luck! You gotta go an' TAP THAT DRAFT! :-D

/Subvertadown