What should attributes be based off of?

[DCEBB2001]

Quote: Originally Posted by guaps

Just like you need to setup the run game to preform effective play-action plays - and not have the rely on play-action and PRC ratings, most things are about setting up a certain thing to be more effective. Take a defensive end. Basically, you have the speed rush and the bull rush. The bull rush is effective if the DE is strong, explosive and has a low center of gravity (James Harrison at 6'0" is a perfect example). The main aspect of the bull rush is gaining leverage to get the OT off balance. No matter how strong the OT is, he is gonna lose if doesn't have balance. A speed rush on the other hand is normally about attacking the OTs outside shoulder and beat him upfield before he can set up or run the edge with the DE (Julius Peppers is a prime example of this). Here the key attribute is explosiveness/acceleration. Now to get back to my play-action example above, a DE must be effective in the outside speed rush or bull rush before he can use effective inside counter moves (finesse moves). Otherwise he is not going to force the OT to open his hips, which will reduce his strength and balance enough for the DE to win the battles. Again I argue that this cannot be boiled down to a single power move and finesse move rating - actually I believe those ratings would be redundant if we had proper OL/DL interactions. As for the pass blocking you should see that slower and less athletic OTs will open their hips up too early to attempt to compensate for their lack of agility. This has nothing to do with blocking skills, but the lack of physical abilities. This is why NFL scouts LOVE tackles that have a rare combination of size, strength and athleticism like, say, Jonathan Ogden. Final note: Arm length plays a huge factor in OL/DL interactions (especially in the pass rush), yet Madden does not have individual arm lengths for their player models (but FIFA 11 does with individual hand size too and it's a soccer game!!). One of the things that makes Mario Williams so effective in the speed rush is that he has exceptionally long arms and strength to keep the OTs away from his body.

More good info here. So you would argue that the FMV is more closely correlated to a player's ACC? What about the PMV? Low center of gravity? That can be calculated as well. What other correlations or scientific explanations could you gather or would use for rating player attributes?

[guaps]

Quote: Originally Posted by DCEBB2001

More good info here. So you would argue that the FMV is more closely correlated to a player's ACC? What about the PMV? Low center of gravity? That can be calculated as well. What other correlations or scientific explanations could you gather or would use for rating player attributes?

No, what I argue is that a speed rush is not a move. It's speed and leverage. Just like a bull rush is strength and leverage. The inside counter moves are the finesse and power move that only happen, when the DE is able to run the speed rush and bull rush effectively - just like a play-action play.

Again, the ratings shouldn't be in the game, IMO. A speed rush (or sometimes just the threat of a speed rush) forces the OT to open up for an inside move, thus allowing for a quick change of direction or a spin move (which should just be the agility rating) or for pushing the OT out of the way (which should just be the strength rating).

The bull rush works in a similarly fashion. Here the DE uses his strength and leverage to force the OT backwards and into the quarterback. If the quarterback tries to scramble, then the DE will use his strength to disentangle from the bull rush and his agility to change direction.

[guaps]

Oh, and for a quick overview on how weight should matter with trucking, you should look at this video from sports science. This video shows the reason why defenders almost always try tripping or wrapping up (and waiting for help) a ball carrier that weights more than they do.

[DCEBB2001]

Quote: Originally Posted by guaps

Oh, and for a quick overview on how weight should matter with trucking, you should look at this video from sports science. This video shows the reason why defenders almost always try tripping or wrapping up (and waiting for help) a ball carrier that weights more than they do.

I understand that blockshedding takes a lot of technique work for just about any position, but what other factors matter in shedding a block? Strength would be my guess. Thoughts?

[cheapargentina]

catching in my mind is more of a rating that is based on stats than actual science (sorta like NBA games do the free throw shooting percentages)

a great stat for catching would be the amount of balls a player catches with his hands and the amount he catches with his body. this could be tied directly into the possession and spectacular catch ratings

if we had a way of measuring hand eye coordination and maybe hand size(shouldnt matter too much but it does affect a players ability)

pass rush is just a game of rock, paper, scissors (speed, bull, or finesse) u get ur opponent thinking one way and u go the other
(what i am confused about is the fact that u think ratings shouldnt be there. are you saying the should remove the ratings they have and replace them or just have it be some sort of mini game where u beat ur opponent if u choose the correct move regardless of the type of player u have?)


i would say that balance and weight distribution would be big factors when it comes to blockshedding

[DCEBB2001]

Quote: Originally Posted by cheapargentina

catching in my mind is more of a rating that is based on stats than actual science (sorta like NBA games do the free throw shooting percentages) a great stat for catching would be the amount of balls a player catches with his hands and the amount he catches with his body. this could be tied directly into the possession and spectacular catch ratings if we had a way of measuring hand eye coordination and maybe hand size(shouldnt matter too much but it does affect a players ability) pass rush is just a game of rock, paper, scissors (speed, bull, or finesse) u get ur opponent thinking one way and u go the other (what i am confused about is the fact that u think ratings shouldnt be there. are you saying the should remove the ratings they have and replace them or just have it be some sort of mini game where u beat ur opponent if u choose the correct move regardless of the type of player u have?) i would say that balance and weight distribution would be big factors when it comes to blockshedding

Then again you have to have strength to shed a player off of you.

Another thing I am really trying to dive into is how do you give value to attributes that lie outside of a player's core positional attributes?

For instance, we all know that a QB is rarely going to use his PMV attribute, but how would you rate it for a QB? Would you just use a random number somewhere near the population or sample average? Or would you try to correlate it to another value. For instance, since the PMV requires a bit of strength to pull off, perhaps make the stronger QBs have a higher PMV value and the weaker QBs have a lower PMV. The same could be said for a QB's FMV. Since this requires more agility to pull off well, QBs with higher agility ratings should have higher FMV ratings than QBs with lower agility ratings.

Now all of this goes out the window when you get to QBs that played S as well in college or situations like that, but I have all of that info on-hand. So, instead of basing it on an average, I can instead base it on scouting data to fill in those holes.

How do you think catching for a QB should be quantified? In theory (or a guess), QBs with higher catch ratings likely played another position that is likely based on speed such as RB or WR. Should a QB with higher speed have better catch than a QB with lower speed?

How do you think non-position specific ratings should be calculated realistically?

[cheapargentina]

Quote: Originally Posted by DCEBB2001

Then again you have to have strength to shed a player off of you. Another thing I am really trying to dive into is how do you give value to attributes that lie outside of a player's core positional attributes? For instance, we all know that a QB is rarely going to use his PMV attribute, but how would you rate it for a QB? Would you just use a random number somewhere near the population or sample average? Or would you try to correlate it to another value. For instance, since the PMV requires a bit of strength to pull off, perhaps make the stronger QBs have a higher PMV value and the weaker QBs have a lower PMV. The same could be said for a QB's FMV. Since this requires more agility to pull off well, QBs with higher agility ratings should have higher FMV ratings than QBs with lower agility ratings. Now all of this goes out the window when you get to QBs that played S as well in college or situations like that, but I have all of that info on-hand. So, instead of basing it on an average, I can instead base it on scouting data to fill in those holes. How do you think catching for a QB should be quantified? In theory (or a guess), QBs with higher catch ratings likely played another position that is likely based on speed such as RB or WR. Should a QB with higher speed have better catch than a QB with lower speed? How do you think non-position specific ratings should be calculated realistically?

i agree i didnt mean to say that strength isn't required to shed a block

using ur example of a qb that played S in college couldnt u use the data for players u have and find similarities between those that did play other positions and those that didnt and maybe rate them similarly?

this could maybe apply for all positions say a DT around the same size as eric foster or terrance cody could maybe get similar ratings to them in run blocking since they play a little fullback. i wouldn't rate them as high because there is a reason they arent playing fullback but maybe use it as a guide on how to rate them.

i wouldnt rate faster qbs higher in other categories just because they are fast enough to play those other positions

[DCEBB2001]

Quote: Originally Posted by cheapargentina

i agree i didnt mean to say that strength isn't required to shed a block using ur example of a qb that played S in college couldnt u use the data for players u have and find similarities between those that did play other positions and those that didnt and maybe rate them similarly? this could maybe apply for all positions say a DT around the same size as eric foster or terrance cody could maybe get similar ratings to them in run blocking since they play a little fullback. i wouldn't rate them as high because there is a reason they arent playing fullback but maybe use it as a guide on how to rate them. i wouldnt rate faster qbs higher in other categories just because they are fast enough to play those other positions

Well what often happens these days is a player will be scouted as something he DIDN'T play in college. For example Eric Crouch (QB) as a S and Korey Hall (LB) as a FB. We don't have any data on them to show how good they are as a FB when they come out into the NFL as rookies.

As for QBs who only played QB in their past, how would you rate their catching? How about their TRK? Hit power? MCV? ZCV? These are the questions that I am asking to get people to think about possible solutions...unless you prefer the "random attribute number generator" which is possible as well LOL!