No more guesswork: Unveiling new threshold calibration method

[Kevin McKoy]

The human brain is not a precision instrument. Our reaction times can vary significantly and are influenced by numerous factors, including fatigue, distraction, and even the angle of view. This means that hand timing can introduce a substantial margin of error into the measurements. Even the most focused and well-intentioned person can't achieve the level of precision that electronic timing can.

This is where my method comes in. I employed an app that allows you to track a moving object and time it precisely, much like the chips used in NFL's

Moreover, this method allows for the precise tracking of the player's movement, enabling you to measure the exact moment the player crosses the 40-yard line. This level of precision is simply unattainable with hand timing.

There has long been a debate between 75 threshold and 95, measuring the time of the average OL and average WR, the debate ends today.

https://youtu.be/QB807C0DF7o this is at 75 threshold

https://youtu.be/O3wNZ75UQic this is at 95

There is a delay between the stop watch and the moments before he begins his running motion, so if you break it down in slow mo, just deduct the seconds prior to him running and boom.

The olineman at 75 runs about a 5.44 and a 5.14 at 95, so if 95 felt too fast for some of you, it's because it was.

The ratings I used were 62 for agility, 77 for acceleration and 65 speed, average lineman speed in the game.

the site I used for the stopwatch you see in the videos is https://www.kinovea.org/ the software is completely free.

P.S. it was 75 for Madden 22 as well, don't believe me, use the software.

This is T.J. Watt speed 83, https://youtu.be/8Mxc5AD7wDw after deducting the seconds that pass by prior to him running, he ran about a 4.81, at the combine with no pads and a runners stance, he ran a 4.69. So he is well within the .02 to 0.3 range you heard about earlier when you include pads. 90 spd, which is the average speed for wr in Madden, runs about a 4.6ish, so also within the range.

As far as tight ends, Kelce's speed is 86, so you know if Watt ran a 4.81, he is in the 4.7's, at the combine he ran a 4.61, again, still not out of bounds in anyway.

Now, if someone else wants to run a different experiment, heck, maybe one of you is a sports scientist and you've got a better model and it thrashes the 40 etc, cool. Show me the literature and i'm good.

But if the 40 is the standard, it's case closed. We've run out of positions, and really it's the speed that matters anyway, I've tested wide receivers with 65 speed, 62 agility, 77 accel and there was no difference, so. This is where i'm at.

[ForUntoOblivionSoar∞]

Re: No more guesswork: Unveiling new threshold calibration method

This is pretty awesome, so don't take this as criticism. But:







Wouldn't you need to do more tests, such as WRs who ran similar 40 times versus very dissimilar ones?


Also, I imagine it would be worth it to adjust a WR's ratings so they actually match their 40 time, and then compare it to a WR with a little slower speed, and then a lot slower, to see how things stack up.



I'm assuming you've actually done some if not all of what I'm talking about and just haven't posted the videos. Or you plan to.

[Kramer5150]

Re: No more guesswork: Unveiling new threshold calibration method

This was discussed quite extensively during M22, and I am strongly in the camp that the higher threshold best represents the "tight windows" you see on Sunday's.

https://forums.operationsports.com/f...threshold.html

Since this would otherwise be buried in my User v User slider thread which isn't applicable to most here, I'm re-posting some player speed testing results for the purposes of establishing a SPD threshold baseline for your games:

Spd Threshold: 50
Avg WR (90 spd, 90 acc, 90 agi): 4.80s to run 40 yds
Avg OL (65 spd, 77 acc, 62 agi): 5.87s (22.2% slower)


Benchmarked against NFL combine #s over a 10y stretch, we ideally want to see an avg OL run a 16.9% slower time than the avg WR. In other words, the gap between fast & slow players on default 50 is too high (could have told you that just from eyeballing it though).

After testing across various spots, tentatively I find 95 to be the closest to the NFL:

Spd Threshold: 95
Avg WR (90 spd, 90 acc, 90 agi): 4.70s
Avg OL (65 spd, 77 acc, 62 agi): 5.50s
17.0% gap

While the timing isn't exact science (reliant on my hand-timed self), keep in mind I timed each sprint 5-10x to make sure I was taking the midpoint of my hand times instead of something flukishly fast/slow. So I feel this is overall fairly reliable even if there could be variance by a few hundredths of a second.

For context, last year we landed around 70-75. So, 95 would be the highest setting we've ever used. A few years ago I think 85-90 ended up being a sweet spot but recently it's been closer to 70-75.

One note on the 'raw' time, namely an avg WR taking 4.7s to get 40 yds. That doesn't mean the game is too slow. We don't have exact translation from 40 time in shorts with a sprinter's release (which makes a huge diff btw, prospects train JUST on their sprint release technique which of course has nothing to do with an NFL skill), but the estimates I've found online range from a ~4.45 probably being closer to 4.65ish in game. So ballpark, a 4.7 in-game time a 90 spd WR is in the ballpark of what an NFL target might be, based on my research. Obviously the game is moving more toward MPH than 40 times, but let me know if you have read otherwise RE: the above assumptions!


I will continue to think/observe to make sure I'm not missing anything wonky that negatively impacts gameplay in a way that doesn't show in these controlled tests, but most likely this is what I'll be rocking with.

[canes21]

Re: No more guesswork: Unveiling new threshold calibration method

While more accurate testing is always welcome, I do agree with others that more testing with different positions and scenarios is needed.

[ForUntoOblivionSoar∞]

Re: No more guesswork: Unveiling new threshold calibration method

This was discussed quite extensively during M22, and I am strongly in the camp that the higher threshold best represents the "tight windows" you see on Sunday's.

https://forums.operationsports.com/f...threshold.html

How is the average WR speed in the 4.70s accurate? Pretty sure the average WR speed is in the 4.50s. According to NFL.com's combine results, the average speed (manually calculated) was 4.496. I suppose you can assume some not at the combine were slower, but there were certainly some faster who didn't go to the combine as well.


If speed differential slows WRs to TE speed, that's probably a problem.


EDIT: and that's assuming 90s for spd, acc and agi are averages. If they're higher than average, that's a problem. Of course if they are slower than average, then this is probably good.


And what does this "but the estimates I've found online range from a ~4.45 probably being closer to 4.65ish in game." mean?

[Kramer5150]

Re: No more guesswork: Unveiling new threshold calibration method

How is the average WR speed in the 4.70s accurate? Pretty sure the average WR speed is in the 4.50s. According to NFL.com's combine results, the average speed (manually calculated) was 4.496. I suppose you can assume some not at the combine were slower, but there were certainly some faster who didn't go to the combine as well.


If speed differential slows WRs to TE speed, that's probably a problem.


EDIT: and that's assuming 90s for spd, acc and agi are averages. If they're higher than average, that's a problem. Of course if they are slower than average, then this is probably good.

Oh Lord, please help me....

[Playmakers]

Re: No more guesswork: Unveiling new threshold calibration method

How is the average WR speed in the 4.70s accurate? Pretty sure the average WR speed is in the 4.50s. According to NFL.com's combine results, the average speed (manually calculated) was 4.496. I suppose you can assume some not at the combine were slower, but there were certainly some faster who didn't go to the combine as well.


If speed differential slows WRs to TE speed, that's probably a problem.


EDIT: and that's assuming 90s for spd, acc and agi are averages. If they're higher than average, that's a problem. Of course if they are slower than average, then this is probably good.


And what does this "but the estimates I've found online range from a ~4.45 probably being closer to 4.65ish in game." mean?

This is why I still prefer the data from FBG ratings for madden.

Realistically very few players will have 90 agility or speed.

And most of the OL isn't going to have 65 speed or higher.

That's just EA taking the initiative to inflate the speed ratings In recent years particularly on the OL

Most of them arent clocking 5.0 or better which means they should be in the 50-60 speed range as opposed to 65 and higher

Having said all that this is still madden and trying to figure out EA's football game especially over the last decade is just pointless because I don't think they even know at times what happens under the hood of their game.

EA themselves has never showed a testing of their own game thresholds at least not to my knowledge.

[ForUntoOblivionSoar∞]

Re: No more guesswork: Unveiling new threshold calibration method

Oh Lord, please help me....

There's nothing unreasonable about that. The clause, "but the estimates I've found online range from a ~4.45 probably being closer to 4.65ish in game." Is simply ambiguous. Generally speaking, the phrase, "in game" means in a VIDEO game (which could indicate that Madden's speed scale origin has been translated away from the NFL's true origin). But this passage could also be arguing that NFL players lose two tenths of a second on their speed when they put on pads and play in games—and is of course stated with zero evidence. Never mind the fact that some players actually play faster than their 40 times for various reasons.



This sounds to me like it was numbers pulled from the ether, and again, it's unclear precisely what is meant by this clause.

[ForUntoOblivionSoar∞]

Re: No more guesswork: Unveiling new threshold calibration method

This is why I still prefer the data from FBG ratings for madden.

Realistically very few players will have 90 agility or speed.

And most of the OL isn't going to have 65 speed or higher.

That's just EA taking the initiative to inflate the speed ratings In recent years particularly on the OL

Most of them arent clocking 5.0 or better which means they should be in the 50-60 speed range as opposed to 65 and higher

Having said all that this is still madden and trying to figure out EA's football game especially over the last decade is just pointless because I don't think they even know at times what happens under the hood of their game.

EA themselves has never showed a testing of their own game thresholds at least not to my knowledge.

That former dev guy Adam something mentioned that a higher speed differential was more accurate.

[ForUntoOblivionSoar∞]

Re: No more guesswork: Unveiling new threshold calibration method

Oh Lord, please help me....

There's nothing unreasonable about that. The clause, "but the estimates I've found online range from a ~4.45 probably being closer to 4.65ish in game." Is simply ambiguous. Generally speaking, the phrase, "in game" means in a VIDEO game (which could indicate that Madden's speed scale origin has been translated away from the NFL's true origin). But this passage could also be arguing that NFL players lose two tenths of a second on their speed when they put on pads and play in games—and is of course stated with zero evidence. Never mind the fact that some players actually play faster than their 40 times for various reasons.



This sounds to me like it was numbers pulled from the ether, and regardless, it's unclear precisely what is meant by this clause.

[Kramer5150]

Re: No more guesswork: Unveiling new threshold calibration method

Okay, all the "scientific BS" outta here, the slider simply states a lower threshold give bigger seperation while a higher threshold tightens up the speed differential....

NOW with that said contrary to popular opinion, the slider has NO bearing on top end speed imo...it's all about separation, and believe it or not at a higher threshold setting separation STILL happens.

I read during last years beta that folks were worried that dudes like Tyreek were going to be "caught from behind" by a lb using a higher threshold....this is simply not true.

[ForUntoOblivionSoar∞]

Re: No more guesswork: Unveiling new threshold calibration method

Okay, all the "scientific BS" outta here, the slider simply states a lower threshold give bigger seperation while a higher threshold tightens up the speed differential....

NOW with that said contrary to popular opinion, the slider has NO bearing on top end speed imo...it's all about separation, and believe it or not at a higher threshold setting separation STILL happens.

I read during last years beta that folks were worried that dudes like Tyreek were going to be "caught from behind" by a lb using a higher threshold....this is simply not true.

I hate to be all "scientific" here, but a linebacker catching Tyreek Hill doesn't inherently have anything to do with the top speed. The slider could lower the top speed and the lb still won't necessarily get him, because all other speeds could be lowered as well.

I would be interested to know if it's a linear transformation of all speeds from one set to another, or is it just a changing of the lower extreme.


What I mean is, does it do this:

{1, 2, 3, 4, 5} → {3, 4, 4, 4, 5} (bad)

or does it do this:

{1, 2, 3, 4, 5} → {3, 3.5, 4, 4.5, 5} (good).
.

But note for the sake of argument, it could also do this:

{1, 2, 3, 4, 5} → {3, 3.4, 3.8, 4.2, 4.6}

or this

{1, 2, 3, 4, 5} → {3, 3.6, 4.2, 4.8, 5.4}


.
.
.
.

Also, since we're here, Tyreek Hill's top speed last year was the equivalent of a 3.52 40 yard dash, with pads.* Acceleration is the thing here. He could never get that fast in the first 40 yards. Watching the play it looked like he got there by the time he was somewhere between 50 and 60 yards nearly straight running. We need an acceleration differential. Or does the speed differential affect that too?


*40 yards/[23.24 mph × (1 hour/3600 seconds) × (1760 yards/mile)]

[ForUntoOblivionSoar∞]

Re: No more guesswork: Unveiling new threshold calibration method

Assuming the speed differential doesn't lower maximum speed, however, I think the most realistic way to do it is a very high differential setting, but also manually alter speed/acc/agi settings so that the rare super humans like Hill stand out, but the overwhelming majority of everyone else is closer together.

[RogueHominid]

Re: No more guesswork: Unveiling new threshold calibration method

Trying to distill the main point of the OP here--is it that 75 threshold is a more realistic representation of player speed differentials than 50 or 95?

[ForUntoOblivionSoar∞]

Re: No more guesswork: Unveiling new threshold calibration method

Trying to distill the main point of the OP here--is it that 75 threshold is a more realistic representation of player speed differentials than 50 or 95?

Given that these two sides seem to be in direct opposition, maybe the better move is to just go with 85. :o