Combine Scores

[Warhammer]

Do we have a repository of combine score affiliations?

I know for OL that their 40 score corresponds to their run blocking, their agility corresponds to their pass blocking, and that bench goes to their blocking strength. But what do these mean for the rest of the positions?

[Kobeck]

I never knew that. And it makes sense.

Must scour this site for more such nuggets of wisdom :D

[sovereignstar]

I never knew that. And it makes sense.

Must scour this site for more such nuggets of wisdom :D

You could use a whole mine as a matter of fact. :)

[Kobeck]

No doubt

[MIJB#19]

I figured some of this each, sort of. I'll post what I think I figured out. All are based on correlations between the blue bars and combine scores, so it's not 100% water proof. Still, I only listed correlation of .666 and above. There were it's close to the .666 barier, I listed it as moderate correlation.

QB - solecismic - avoid interceptions
QB - 40 time - scramble frequency
QB - agility - sense rush / screen passes (both have a moderate correlation)

RB - 40 time - Breakaway speed
RB/FB - strenght - power inside

TE/WR - 40 time - Big Play
TE - strenght - blocking strenght
TE/WR - agility - getting downfield
WR - solecismic - route running (moderate correlation)
WR - strenght - courage over the middle

C/G/T - 40 time - run blocking
C/G/T - strenght - blocking strenght
C/G/T - agility - pass blocking

K - 40 time - kickoff distance
K - strenght - kickoff hang time (moderate correlation)

P - 40 time - punting power
P - strenght - hang time

DE/DT/LB - 40 time - pass rush technique
DE/DT/LB - strenght - pass rush strenght
DE/DT/LB - agility - run defense

CB/S - agility - interceptions


The above listed all seem to be sure things, as far as correlation of randomnly modified numbers can generate sure things. There are also a bunch of two-skills-make-one-combine-score sets or near .500 correlations. I'll list those below.

near .500
QB - strenght - deep passes
RB - solecismic - hole recognition
FB - agility - third down running
K - solecismic - field goal accuracy
DE/DT/LB - solecismic - play diagnosis
DE/DT - strenght - punishing hitter

combo sets
RB - agility - third down running & elusiveness & getting downfield
P - solecismic - punting power & directional punting
CB/S - strenght - bump and run & punishing hitter

[Warhammer]

I'll just confirm based upon some of my own observations...

I figured some of this each, sort of. I'll post what I think I figured out. All are based on correlations between the blue bars and combine scores, so it's not 100% water proof. Still, I only listed correlation of .666 and above. There were it's close to the .666 barier, I listed it as moderate correlation.

QB - solecismic - avoid interceptions - confirm - also known formations
QB - 40 time - scramble frequency
QB - agility - sense rush / screen passes (both have a moderate correlation)

I'll also add read defense to solecismic.

RB - 40 time - Breakaway speed - confirm
RB/FB - strenght - power inside - confirm

TE/WR - 40 time - Big Play
TE - strenght - blocking strenght - confirm
TE/WR - agility - getting downfield
WR - solecismic - route running (moderate correlation) - I've noticed more correlation with agility.... I'll have to look at this a bit more.
WR - strenght - courage over the middle - confirm

C/G/T - 40 time - run blocking - confirm
C/G/T - strenght - blocking strenght - confirm
C/G/T - agility - pass blocking - confirm

K - 40 time - kickoff distance
K - strenght - kickoff hang time (moderate correlation)

P - 40 time - punting power
P - strenght - hang time

DE/DT/LB - 40 time - pass rush technique - confirm
DE/DT/LB - strenght - pass rush strenght - confirm
DE/DT/LB - agility - run defense - confirm

CB/S - agility - interceptions - confirm

I'll add strength for CB/S for BnR coverage

The above listed all seem to be sure things, as far as correlation of randomnly modified numbers can generate sure things. There are also a bunch of two-skills-make-one-combine-score sets or near .500 correlations. I'll list those below.

near .500
QB - strenght - deep passes - confirm
RB - solecismic - hole recognition - this is one I would expect, but it does not seem to hold up in my experience. Could this be solecismic and agility together?
FB - agility - third down running
K - solecismic - field goal accuracy
DE/DT/LB - solecismic - play diagnosis - confirm
DE/DT - strenght - punishing hitter - confirm

combo sets
RB - agility - third down running & elusiveness & getting downfield
P - solecismic - punting power & directional punting
CB/S - strenght - bump and run & punishing hitter - confirm

[Celeval]

QB - 40 time - scramble frequency
QB - agility - sense rush / screen passes (both have a moderate correlation)

RB - 40 time - Breakaway speed
RB/FB - strenght - power inside

It's been my experience that QBs have all the running-based ratings of a RB or a FB, only hidden.. I wonder (and may be impossible to tell this without drawing correlations from the file rather than from the game) if the correlations between QB combine #s and ratings are the same as RB combine #s and ratings.

[MrBigglesworth]

All are based on correlations between the blue bars and combine scores, so it's not 100% water proof.
If you can get the info from the blue bars, that makes the combine numbers worthless. Have you looked at how the combine numbers effect the RED bars? In other words, do the combine scores help you to predict current ability? Red bars correlate with the blue bars, but sometimes a player will end up with blue bars at 50 for two different skills, but end up with 5 in red for one and 45 for another (and not just the ones with no green bars). Or do the combine scores just allow you to spot booms?

[Celeval]

If you can get the info from the blue bars, that makes the combine numbers worthless. Have you looked at how the combine numbers effect the RED bars? In other words, do the combine scores help you to predict current ability? Red bars correlate with the blue bars, but sometimes a player will end up with blue bars at 50 for two different skills, but end up with 5 in red for one and 45 for another (and not just the ones with no green bars). Or do the combine scores just allow you to spot booms?
Not the random boom/busts, but the scout errors - the ones where you draft a guy at 70/70 and he's -3/-3 immediately, -6/-6 after training camp, and down to -14/-14 after three years... or vice versa.

[RPI-Fan]

If you can get the info from the blue bars, that makes the combine numbers worthless. Have you looked at how the combine numbers effect the RED bars? In other words, do the combine scores help you to predict current ability? Red bars correlate with the blue bars, but sometimes a player will end up with blue bars at 50 for two different skills, but end up with 5 in red for one and 45 for another (and not just the ones with no green bars). Or do the combine scores just allow you to spot booms?

I think what MIJB was getting at, was that he did a study, and instead of plotting/analyzing "Red Bars vs. Combine Scores", which would have required progressing through training camp, etc., he did "Blue Bars vs. Combine Scores", which, in a large enough study, should theoretically produce statistically speaking the same exact results with regard to correlations.

[MrBigglesworth]

I think what MIJB was getting at, was that he did a study, and instead of plotting/analyzing "Red Bars vs. Combine Scores", which would have required progressing through training camp, etc., he did "Blue Bars vs. Combine Scores", which, in a large enough study, should theoretically produce statistically speaking the same exact results with regard to correlations.
Yes, I know that he looked at blue bars. I was wondering if he took it further. In other words, blue bars could only have a correlation with combine scores because the blue bars are also correlated with red bars. The combine scores may come from the red bars (or vice versa) though, and if that were the case you would find a higher correlation between red/combine than between blue/combine.

[Celeval]

Yes, I know that he looked at blue bars. I was wondering if he took it further. In other words, blue bars could only have a correlation with combine scores because the blue bars are also correlated with red bars. The combine scores may come from the red bars (or vice versa) though, and if that were the case you would find a higher correlation between red/combine than between blue/combine.
I doubt that - for no other reason then they would have to have some sort of correlation as well with the percentage developed - you wouldn't see any of the 5%, 10% developed guys anywhere near the combine scores of the 50%, 60% guys; and that hasn't been my experience. Anecdotally, of course, but I know I've chosen guys based on quality combine scores who have been not very well developed.

[Warhammer]

I doubt that this is possible simply because you don't know how the red bars are going to be allocated to each. However, you DO know how the blue bars (future potential) are roughly allocated.

[MIJB#19]

I took the blue bar numbers because I had them at hand and figured they could be useful. The above posted numbers are based on just one draft class. I'll do a similar examination of five other draft classes being included. Also, I think this one was done with a class pre the last patch, so things *could* be different for some of the other draft class numbers I've got.

[MIJB#19]

It's been my experience that QBs have all the running-based ratings of a RB or a FB, only hidden.. I wonder (and may be impossible to tell this without drawing correlations from the file rather than from the game) if the correlations between QB combine #s and ratings are the same as RB combine #s and ratings.I doubt that.

For example, the FB bench press seems to be correlated with power inside and not blocking strenght, while the TE and OL positions have a correlation between the blocking strenght and the bench press.