It is quite shocking to me that you are willing to go on record in front of the whole community and say you are fine with an 8 to 0 turnover margin leading to an obviously much better team losing a playoff game that they dominated statistically 19-16 and ending their season.

This is unbelievably dismissive and gaslighting. Can you at least acknowledge that this is not the kind of result that you would want or hope for from the sim? Or do you want this kind of unpredictability and chaos?

I'd like to take a step back and clarify what I'm looking for from the beta.

The purpose of this league is to identify issues with the simulation and improve it. If the interception changes have introduced a problem, I absolutely want to find it and fix it.

What I need, though, is actionable feedback. If you think something is wrong, point to specific plays or situations that illustrate the issue. For example, are you seeing the same type of interception happen repeatedly? Is there a particular coverage, route, quarterback decision, or game situation that consistently produces results that don't make football sense? Those kinds of observations help me identify patterns and determine whether there's something in the simulation that needs to be adjusted.

I also want to hear from everyone else in the beta. Are others seeing interception rates that seem out of line? Are you seeing the same types of plays occurring repeatedly, or are the results generally matching your expectations? Hearing from multiple users helps me determine whether I'm looking at a broader tuning issue or something specific that needs further investigation.

I welcome criticism, but comments like "Why would anyone pay for this?" or accusations that I'm being dismissive don't help me identify or solve a problem. If you believe there's an issue, help me understand it by pointing to specific examples and recurring situations rather than focusing only on the final result.

At the end of the day, we're all here for the same reason—to make the simulation better. Let's keep the discussion focused on gathering evidence so I can make informed improvements.

Nothing about this makes any logical sense. My DB's were constantly in front of their receivers when passes were thrown. Their QB went 11/29 with zero interceptions. When you look at knockdowns, we had 10 on 29 throws, they had 9 on 46. Their DB's who got picks have pass catching attribute: 21,15,13(2 picks),13,5. My DB's who had knockdowns have pass catching attribute: 23,21,17,15,10. We were in position more often, and have higher pass catching attributes, yet zero picks to their 6.

This is a common theme that certain attributes show no (or even opposite) correlation to results. I looked at avoid fumble in the past, and found that statistically, higher avoid fumble attribute predicted the player would have a higher (per touch) fumble rate. This was on a full season of data from one entire league.

Pass blocking interactions, pressure, and sacks is something that certainly needs a close look right now. There is no excuse for my Tackle Denson giving up 20 sacks this regular season. Something is clearly out of whack there.

In all one on one interactions (be it blocking, tackling, or route running vs. coverage), attributes need to more clearly decide who "wins" the interaction.

The foot race problem has been a persistent issue that desperately needs addressing. I can't stand seeing my 90+ speed DB roasted for a 75 yard TD by a 73 speed WR, and then later in the same game, my 90+ speed receiver chased down by a 68 speed DB when he had a 10 yard advantage when he caught the ball. Then I post about it and get more gaslighting about "fatigue" and other such nonsense.

Thank you. This is much more helpful because you've pointed to specific areas of the simulation that you believe aren't behaving as expected.

The goal of the beta isn't to determine whether a single game result was "right" or "wrong." It's to identify recurring situations that suggest the simulation isn't behaving as intended. That's the kind of evidence I can investigate and either confirm as an issue or determine is working as designed.

For example, regarding the speed examples, if you can point me to several plays where a significantly slower defender consistently runs down a much faster ball carrier (or vice versa), I'd like to look at them. One play by itself doesn't necessarily indicate a problem, but if there's a recurring pattern, that's something I can trace through the engine and determine whether the logic is producing the intended result.

The same goes for interceptions, pass blocking, sacks, or any of the other areas you mentioned. If there's a pattern, I want to find it. If attributes aren't influencing outcomes the way they're intended to across many interactions, that's something I absolutely want to address.

I also want to hear from everyone else in the beta. Are you seeing similar patterns, or are your results generally matching your expectations? The more data points I have from different teams, the easier it is to determine whether I'm looking at a broader tuning issue or something more isolated.

One thing I do want to point out is that I don't expect every individual play or every individual game to produce the "expected" result. Football has outliers, and I want the simulation to have them too. If the better player always won every matchup or the better team always won every game, the simulation would become predictable rather than realistic.

What I care about is whether those outcomes are occurring at the right frequency. Football occasionally produces surprising games where everything seems to go wrong for one team or right for the other. I expect a simulation to occasionally produce those kinds of outlier results as a natural consequence of many interactions, not because they're intentionally scripted. What I don't want is for those outcomes to occur so often that they suggest player attributes aren't influencing results the way they should across many interactions. If that's happening, I want to find it and fix it.

That's ultimately what this beta is for. The more we can focus on identifying recurring patterns and providing concrete examples, the easier it is for me to make meaningful improvements to the simulation.

First piece, we may have to agree to disagree, but I think you are 100% wrong here. I am talking foot race, straight line sprint. There can not reasonably be any chance of a significantly slower player beating a faster one in this situation. When it does come down to this, speed has to win every time. One instance absolutely does show that this is completely out of whack.

I have every reason to believe this is happening at large scale. It is incredibly cumbersome to analyze the data. Is there any way we could get CSV files per position that contains every player at the position, their attributes, and their stats for the season? That would make it quick and easy (enough) to check what the correlations between attributes and performance are looking like.

There is one thing that I have theorized and observed anecdotally that could be skewing the results for me.

I truly believe there is plenty of hard evidence to say that there is code in place to limit the damage in blowout victories. That is, when you go up by too much too quickly, something kicks in that suppresses your team and/or boosts the opposing team for the rest of the game, regardless of whether they come back or not. It happens like clockwork when I'm getting too far ahead, and the opposing team almost always closes the gap over the rest of the game after showing complete ineptitude from the start.

Much more vaguely, and less evidence, it seems that within a season, when a team has too many wins racked up, essentially the same thing happens, they will be slammed with a horrendous game and a shocking loss.

Are you able to deny that such code exists?

For the first point: yes, there is a "slow down mode" option in the game plan settings. However, this is not a hidden catch-up mechanic or something the engine activates to keep games close. It is a user-configurable strategy option that changes how your team approaches the game after reaching a specified lead. The team exits that mode when the conditions in the game plan are no longer met.

For the second point: no, there is no code in the simulation engine that looks at a team's record, number of wins, playoff position, or previous game results and adjusts outcomes to create losses or maintain league balance. Each game simulation is isolated and has no knowledge of what happened in previous games.

There are situations where results can appear to follow a pattern because football is a game of probabilities and because many different systems interact with each other. A team can have an unusually poor performance, turnovers can swing a game dramatically, or an opponent can have an unusually strong performance. That doesn't mean the engine is intentionally creating that result.

Regarding the speed examples, remember that the speed attribute alone does not determine how fast a player moves in every situation. Acceleration, player weight, and fatigue all play a role. A fresh 75 speed player may be able to catch a fatigued 90 speed player depending on the circumstances. If there are examples where the result seems inconsistent even when accounting for those factors, those are exactly the types of plays I'd like to review.

You will never convince me that your approach to the foot race scenario is correct. Here is an AI analysis of NFL data pertaining to this question:

The short answer is: less than most people think for most players, but a lot for players who never come off the field or who repeatedly run deep routes or chase plays.

Based on GPS tracking, wearable sensors, and studies of repeated sprint performance in football, here's a reasonable estimate:

Player Usage Speed drop by late 4th quarter
Frequently rotated (DL, WR, RB) 0–2%
Typical every-down player 2–4%
Very high snap count (>90%) 4–7%
Exhausted after long drives or hurry-up offense 5–10% temporarily
What does that mean in mph?

Suppose a receiver's maximum game speed is 21.5 mph.

2% drop: 21.1 mph
4% drop: 20.6 mph
6% drop: 20.2 mph
8% drop: 19.8 mph

That may not sound like much, but in football it's huge.

A 1 mph difference over a 40-yard route can amount to roughly 0.08–0.12 seconds, which is enough to create about 2–3 feet of separation—often the difference between a completion and a pass breakup.

Why the drop isn't larger

NFL players are exceptionally conditioned, and several factors help preserve speed:

Rotations keep linemen and many skill players fresh.
TV timeouts and quarter breaks provide recovery.
Offensive huddles allow short rests.
Players rarely sprint all-out every snap.

Instead of gradually slowing all game, players tend to show:

nearly full speed early,
occasional slower bursts after long drives,
recovery after a rest,
then another decline if the next drive is prolonged.
Position matters
Wide receivers: Usually maintain top speed well because they often have plays off or routes that aren't full sprints.
Cornerbacks: Can fatigue more because they must react every snap and often sprint in response to the receiver.
Running backs: Speed is maintained fairly well, but explosiveness and acceleration tend to fade more than outright top speed.
Defensive linemen: Peak power declines faster than straight-line speed, which is why teams rotate them heavily.
Linebackers and safeties: High snap counts can noticeably reduce closing speed late in games.
The bigger loss is acceleration

One interesting finding is that top speed declines less than acceleration. A player might still be capable of hitting 21 mph if given enough distance, but it takes longer to get there because of:

reduced explosive force,
depleted phosphocreatine stores,
neuromuscular fatigue.

That's why tired defenders often look "a step slow" even if their maximum speed hasn't fallen dramatically.

For someone like an NFL wide receiver who runs a 4.35-second 40-yard dash, a full game's fatigue might change the equivalent performance to roughly:

Fresh: 4.35 s
Typical late game: 4.42–4.50 s
Very fatigued: 4.55–4.65 s

Those differences are enough to affect nearly every contested play, which is why coaches pay so much attention to snap counts and rotations, especially at the skill positions.

The gaps that some corners make up on faster WR's in MFN is absolutely absurd.

The interaction between fatigue, acceleration, speed, weight, and pursuit is an area where there are a lot of moving pieces, and it's worth making sure those pieces are interacting the way they should.

However, I also don't want to delay the pass play improvements if the overall benefit of this release outweighs issues that may have become more noticeable as a result of these changes.

One of the things I've been trying to emphasize with the beta process is that we want to make small, iterative changes between releases rather than trying to overhaul multiple systems at once. This release has been primarily focused on improving pass play simulation, and it's possible that some of those changes have made certain areas, like pursuit and foot races, more noticeable.

For this release, the question is whether this issue is severe enough that it outweighs the improvements we've made to pass play simulation. If it is, then that's something we need to discuss before moving forward. If not, then this is the type of area we can continue to investigate and address iteratively in a future update.

Let's keep this in mind as we evaluate the current release and as we discuss priorities for the next iteration. For now, I'd like to keep the focus on identifying any issues that would prevent this release from being ready for general release.