In-game 40 times vs. combine/pro day data.

[DCEBB2001]

Quote: Originally Posted by garret2

in my mind this kind of changes the whole game. so here is what we know about in game speeds now. 1) all players timed slower in pads than in shorts, which only seems logical and negates the concern that players were faster with pads than his fastest time at the combine. this was the main concern of in game 40s. those odd examples like brett swain's are now better understood. 2) injuries and age are accounted for but if a player was injured he may not have performed optimally for the season. to counter this, we could take the in game 40 speed averages of the last 2 or 3 years to get a more clear look. if a player like driver or tomlinson was hurt last year, his speed will be helped by averaging in the years before when he was healthier. 3) averaging in 2-3 seasons of in game 40s may lessen the fear that the sample size for each player is too small and doesn't show his true potential.

The only problem, is that the in-game 40s get overwritten, so there is no record of past 40 times.

[garret2]

Quote: Originally Posted by DCEBB2001

The only problem, is that the in-game 40s get overwritten, so there is no record of past 40 times.

really? that sucks... well does the new info easy your concern for using in game 40s? i think the validity has increased dramatically, now we dont have to wonder why players were running faster in pads than in shorts. and i love the fact that they are updated regularly so players like mcgahee, tomlinson, driver, etc will perform better assuming they are healed this season.

[DCEBB2001]

Quote: Originally Posted by garret2

really? that sucks... well does the new info easy your concern for using in game 40s? i think the validity has increased dramatically, now we dont have to wonder why players were running faster in pads than in shorts. and i love the fact that they are updated regularly so players like mcgahee, tomlinson, driver, etc will perform better assuming they are healed this season.

To me, the validity is still in question because a lot could be lost in the translation from a 10yd run to determining a 40 time based off of it.

[garret2]

Quote: Originally Posted by DCEBB2001

To me, the validity is still in question because a lot could be lost in the translation from a 10yd run to determining a 40 time based off of it.

so are you still looking for the right mix of both methods or have you decided at this point?

[guaps]

Quote: Originally Posted by garret2

what about a player like brett swain? his official 40 was in the 4.6 range but his in game 40 was in the 4.3 range. so unless the guy calculating these times simply messed up the calculation we would really be underestimating his speed by just using the 40 yard dash. if a player cant run faster than he has ever run before, we will always have to take the higher of the two speeds then to represent their sprint potential. but this is the whole reason why dan doesnt want to use in game 40's exclusively, because he said in the example of swain and others that his speed was overestimated in his opinion....perhaps the way to settle all of this is to define what we want speed to mean in madden. is it... A) the full sprint potential a player had when he ran at the combine/pro day B) about how fast he runs, on average, currently on gameday (and updated regularly) theres no perfect answer here but what do you guys think it should be??

The whole aspect relies on the fact that you have reliable in-game 40s. As I've stated from the beginning, I prefer the hard data approach, but if you have reliable in-game 40s, then you could use them to correct anomalies in the data set (instances were a player has a suspiciously slow 40 time).

However, after reading about how the in-game 40s are calculated, they seem way too unreliable to be used.

[guaps]

Quote: Originally Posted by DCEBB2001

CLARIFICATION: This was brought up to me in an IM: All of the "in-game" 40 times are not actual 40 times. In actuality, all players ran slower in-game than when the timed. So if Chris Johnson ran a 4.24 at the combine, his in-game ACTUAL calculated time was say 4.54. What then happens is ALL of the times are adjusted to match the 40 time taken at the combine/PD. So, Adrian Peterson ran an official 4.40, ran a 4.58 in real time in a game, but was adjusted the .20s to 4.38. Just to clarify. The "in-game" 40 times are thus, then directly comparable to that of the combine/PD 40 time.

Quote: Originally Posted by DCEBB2001

To me, the validity is still in question because a lot could be lost in the translation from a 10yd run to determining a 40 time based off of it.

How can they be comparable to the combine/pro-day 40s, if they're calculated and then afterwards ''adjusted''?

First of, how does this adjustment work? Is there any logic behind the adjustments or is just to make the results seem more probable?

Second, you cannot calculated 40 times using 10 or 20 yard split times. Just imagine if you had to calculate Chris Johnson's combine 40 time using only his 10 and 20 splits. You wouldn't be able to get any realistic results no matter how you did it.

• If you use his 10 yard split and multiply by 4 you get: 5.6 seconds
• If you use his 20 yard split and multiply by 2 you get: 4.82 seconds
• If you use his 10-20 yard split and multiply by 4 you get: 4.04 seconds
• If you use his 20-40 yard split and multiply by 2 you get: 3.66 seconds
• If you use multiply his 10 yard split by 2 and add his 20-40 yard split you get: 4.63 seconds
• etc.

My point is that you cannot get any usable estimate of a players 40 time without measuring a player running 40 yards in pads. And if you arbitrary calculate a 40 time using split times and then adjust them to make them seem more probable, you have do dismiss the data as not being valid.

[DCEBB2001]

Quote: Originally Posted by guaps

How can they be comparable to the combine/pro-day 40s, if they're calculated and then afterwards ''adjusted''? First of, how does this adjustment work? Is there any logic behind the adjustments or is just to make the results seem more probable? Second, you cannot calculated 40 times using 10 or 20 yard split times. Just imagine if you had to calculate Chris Johnson's combine 40 time using only his 10 and 20 splits. You wouldn't be able to get any realistic results no matter how you did it. If you use his 10 yard split and multiply by 4 you get: 5.6 seconds If you use his 20 yard split and multiply by 2 you get: 4.82 seconds If you use his 10-20 yard split and multiply by 4 you get: 4.04 seconds If you use his 20-40 yard split and multiply by 2 you get: 3.66 seconds If you use multiply his 10 yard split by 2 and add his 20-40 yard split you get: 4.63 seconds etc. My point is that you cannot get any usable estimate of a players 40 time without measuring a player running 40 yards in pads. And if you arbitrary calculate a 40 time using split times and then adjust them to make them seem more probable, you have do dismiss the data as not being valid.

The formula is far more complicated than just multiplying. I think it follows the acceleration shown on the film and then the average acceleration of each player. The rate of acceleration can be measured so you can calculate when acceleration will plateau. I believe that is what is used to calculate the 40 from a dead stop.

[DCEBB2001]

Quote: Originally Posted by garret2

so are you still looking for the right mix of both methods or have you decided at this point?

No decision has been made yet.