Understanding initial development

[QuikSand]

I confess this is an area that I happen to know rather little about - so I'm hoping to learn here more than teach. I saw an interesting (but maybe not very unusual) guy in a MP league, and thought he might make a good case study. He's a RB -- and that might be the weirdest position to figure out in the game, to me -- and here's how he looks after his rookie training camp.



So, according to my scout's "bars" he looks like a guy with pretty decent skill in several areas -- and in particular his combination of skills in blitz pickup, getting downfield, and third down catching seem to make him a pretty good candidate to be used in passing situations. Seems to me (in football terms) he looks useful as a blocker, won't draw too many passes, but has a chance to do something with the few that do come his was as a safety valve.

What I really want to try to better understand here is this. Looking at his various red/green bars, it sure *seems* to me that he is pretty nearly fully developed to his potential. Sure, there are a few places where he has a bit more green to develop -- but in most cases, that's just another 10-20 points, mostly in pretty secondary stuff. With that in mind -- what on earth shoudl I make of the scout's overall rating of 23/40? I mean -- just looking at what the same scout is telling me in the reds and greens, I would have guessed this guy should have a composite rating more like 30/35 or so (just taking a stab there). It just seems like the 23 is low, and the 40 is high, and that the gap between them is totally absurd. I'll toss in the curiosity that the guy rates a 90 in RB development (or a 9 or a B whatever the second-highest rating is in the presentation you use) so he appears there to be nearly fully developed -- more in keeping with the bars that the scout shows us.

So, that's the challenge to me. Bars seem to say "useful now, but don't expect much more." Scout's overall rating seems to say he's not that good now, but might get better later. Position experience is nearly maxed out already.

It's a computer game. I understand that. I'm not trying to come up with a paragraph explanation of how his college coach used him as a change-of-pace back but how he practiced with the scout team and learned all these certain skills, blah blah blah. Save it. I just want to know if there's something in a guy like this that I should be able to "read" to have a better handle on what he really is in this game.

[DeltaWhiskey]

I might be able to answer the questions your asking with some of the data analysis techniques I've been using in my exploration (by exploration I mean mental masturbation as I seem to be the about the only one getting pleasure out of the efforts). If you're asking "What bars (and possibly other stuff) goes into the overall scores, then linear regression models seem to be the logical starting point to answer these questions.

I've considered doing this for a while, but just haven't decided when to dive in.

[Ben E Lou]

What were his precamp numbers?

[QuikSand]

Originally Posted by DeltaWhiskey

If you're asking "What bars (and possibly other stuff) goes into the overall scores

That's really not what I'm asking, though it might be worth doing regardless. (By the way, I'd hope that even before you started some absurdly long regression analysis, you could join me here in saying that the visible reds and greens obviously do not add up to a 23/40 guy in any reasonable weighting of what we see -- so this guy is either an aberration or is very highly weighted in something we don't see like avoiding fumbles)


What I'm really trying to ask here: Is there a "tell" here in where this player is heading (and maybe what he's worth right now), just given these three data points:

-his relative red/green bars
-his overall scout rating
-his positional experience

[QuikSand]

Originally Posted by Ben E Lou

What were his precamp numbers?

I want to hold off on that for now, just because so many of us are deeply programmed to focus so intently on that as the defining issue in player development. So, in time we can get to that and see how it fits, but for the moment I want to focus just on what we see above.

[DeltaWhiskey]

Originally Posted by QuikSand

That's really not what I'm asking, though it might be worth doing regardless.


What I'm really trying to ask here: Is there a "tell" here in where this player is heading (and maybe what he's worth right now), just given these three data points:

-his relative red/green bars
-his overall scout rating
-his positional experience

Okay, I'll keep masturbating...mentally

[wade moore]

I understand the question you're asking Quik, but must admit I naturally go to the same question that Ben does. I would think that to get any value out of this you would need to find some segment of guys that look like this as well as a segment of guys that seem to look "normal" and see how they develop over a couple of TC's.

Maybe that sounds obvious? I'm not sure if you're looking for someone that already has a feel for this, but I think you may be onto something - just not sure how you get any value without tracking progress of a guy like this.

[QuikSand]

Originally Posted by DeltaWhiskey

Okay, I'll keep masturbating...mentally

I edited in something above, but after you saw it, I trust...

(By the way, I'd hope that even before you started some absurdly long regression analysis, you could join me here in saying that the visible reds and greens obviously do not add up to a 23/40 guy in any reasonable weighting of what we see -- so this guy is either an aberration or is very highly weighted in something we don't see like avoiding fumbles)

(adding on)

Just looking at his reds, there is easily 3 or 4 times as much red in this guy as green, maybe more. Simply no way that you can tell me that in any simple weighted average of the skills, that tiny amount of remaining green adds up to 17 points on the same scale where all that red adds up to only 23. No way, no how. That's my entire point of the post, something seems "wrong" with this guy. What can we learn from that?

[wade moore]

I think the end of this is that one of two things is likely true:

A) His red is wrong
B) His green is wrong

The question is, how do we decide outside of a coin toss? In a guy like this - is it possible that either one is true, or is it always going to be the same answer?

[Ben E Lou]

Originally Posted by QuikSand

I want to hold off on that for now, just because so many of us are deeply programmed to focus so intently on that as the defining issue in player development. So, in time we can get to that and see how it fits, but for the moment I want to focus just on what we see above.

OK. Fair point (and I agree that RBs are weird anyway). Here's my take on it then...

• I *strongly suspect* that positional experience is 100% accurate.
• I *know* that green bars lie.
• I *suspect* that red bars lie, too.
• I * strongly suspect* that fumble rating, despite being hidden, is taken into the current rating.

So, all that being said, without seeing the firstcamp future potential change, I'd suspect that he ends up at around 32/32 or so--if he ever gets there, keeping in mind that I also *strongly suspect* that RBs can peak in season 1.

[QuikSand]

Originally Posted by wade moore

I understand the question you're asking Quik, but must admit I naturally go to the same question that Ben does. I would think that to get any value out of this you would need to find some segment of guys that look like this as well as a segment of guys that seem to look "normal" and see how they develop over a couple of TC's.

Maybe that sounds obvious? I'm not sure if you're looking for someone that already has a feel for this, but I think you may be onto something - just not sure how you get any value without tracking progress of a guy like this.

That's perfectly fine as far as it goes. We learn a lot about players from their ratings movement. But perhaps by focusing so intently on that, we are missing something else? Maybe there are any number of rookie players in this game who don't show big changes in ratings at any point, but who perform dramatically different than their ratings would suggest? Maybe this guy will end up being scouted at 25/25 for his whole career, but end up with 5 yards a carry, great blocking results, and very productive results on his targeted passes? If he's a -2 in camp, you and I totally miss him, right? Maybe we are completely missing something special here.

(Or maybe this guy is just a pile of shit, and he will never be even as good as his current red bars suggest, and we could pre-confirm this by looking at this sort of information other than ratings movement)

[QuikSand]

Originally Posted by Ben E Lou

OK. Fair point (and I agree that RBs are weird anyway). Here's my take on it then...

• I *strongly suspect* that positional experience is 100% accurate.
• I *know* that green bars lie.
• I *suspect* that red bars lie, too.
• I * strongly suspect* that fumble rating, despite being hidden, is taken into the current rating.

So, all that being said, without seeing the firstcamp future potential change, I'd suspect that he ends up at around 32/32 or so--if he ever gets there, keeping in mind that I also *strongly suspect* that RBs can peak in season 1.

I think I agree with all of this, on the surface. And perhaps picking a RB here was a bad decision.

What would his Avoid Fumbles rating have to be for the 23/40 to make any sense? 0/100 and worth 25% of his value or something like that? Just silly, really.

Honestly, my main thesis here is that this guy would be completely unexceptional if my scout was saying he was 30/35 overall, and had these bars. I'd feel like that was a consistent message. Instead, he is giving me these two totally conflicting bits of feedback... bars say 30/35, scout adds it up and comes up with 23/40. Is this just the way scout error works now? Is this some sort of tell that these bars are not for real (some insight into the times when red bars lie, as you put it)? That's where I'm coming from.

[DeltaWhiskey]

Originally Posted by QuikSand

That's perfectly fine as far as it goes. We learn a lot about players from their ratings movement. But perhaps by focusing so intently on that, we are missing something else? Maybe there are any number of rookie players in this game who don't show big changes in ratings at any point, but who perform dramatically different than their ratings would suggest? Maybe this guy will end up being scouted at 25/25 for his whole career, but end up with 5 yards a carry, great blocking results, and very productive results on his targeted passes? If he's a -2 in camp, you and I totally miss him, right? Maybe we are completely missing something special here.

(Or maybe this guy is just a pile of shit, and he will never be even as good as his current red bars suggest, and we could pre-confirm this by looking at this sort of information other than ratings movement)

I've had guys like that, mostly RB, but I've also recently had a WR, who for the life of me, I can't figure out why he's the leading receiver on the team.

Originally Posted by QuikSand

By the way, I'd hope that even before you started some absurdly long regression analysis, you could join me here in saying that the visible reds and greens obviously do not add up to a 23/40 guy in any reasonable weighting of what we see -- so this guy is either an aberration or is very highly weighted in something we don't see like avoiding fumbles)

That's what regression analysis is for, finding the weightings that aren't obvious.

[QuikSand]

Originally Posted by QuikSand

Is this some sort of tell that these bars are not for real (some insight into the times when red bars lie, as you put it)? That's where I'm coming from.

By the way, as I am typing, this is the idea that is taking center stage in my mind. Red bars are lying here, and the scout's 23 is our clue to that.

[QuikSand]

Originally Posted by DeltaWhiskey

That's what regression analysis is for, finding the weightings that aren't obvious.

I'm just getting the feeling that you and I aren't meant to play in the sandbox together.

[DeltaWhiskey]

Originally Posted by QuikSand

I'm just getting the feeling that you and I aren't meant to play in the sandbox together.

How so?

[wade moore]

Originally Posted by QuikSand

By the way, as I am typing, this is the idea that is taking center stage in my mind. Red bars are lying here, and the scout's 23 is our clue to that.

So your thought is that potentially we should trust the CUR score more than the bars? Do you think this may be the case in all situations? I have to confess I haven't paid enough attention to know how CUR lines up with red bars in "bonafide creepers" - is the thought that CUR matches up with red bars for the most part, just that they go up together?

[QuikSand]

I don't think there are all that many people here who don't understand the value of a regression analysis, and the value of finding correlations between properly selected variables. We get it. I happen to agree that doing such work on every position in this game and seeing how things are weighted would have value. (Especially if it gave us some real insight into things like AvInt/AvFum) It was done and discussed here for a previous version of the game, but it's laborious.

Looking as intently as possible, the only place I can see this guy making sense is if the weight for "hole recognition" is pretty high, and that his 3/28 there is casting a pretty big shadow on the composite. I suppose that's possible -- and maybe that's a meaningful part of the explanation for this guy.

Your giant regression project would shed some light. It isn't the endgame of every question here (or elsewhere).

[QuikSand]

Originally Posted by wade moore

So your thought is that potentially we should trust the CUR score more than the bars? Do you think this may be the case in all situations?

I don't know, it's just a thought in progress.

Originally Posted by wade moore

I have to confess I haven't paid enough attention to know how CUR lines up with red bars in "bonafide creepers" - is the thought that CUR matches up with red bars for the most part, just that they go up together?

I'm basically a fellow blindee here, I tend to make nearly all my judgments on future ratings movement, too. I'm just trying to spur a discussion on what else there might be. Perhaps I didn't pick a good case study, after all.

[3ric]

We do know that he still has some potential left to reach, right? I say this with the assumption that when he reaches his full (real) potential all "fake" green bars disappear.

[jdavidbakr]

If you simply take the average of the green/red bars you get a score of 28/41. To me that's close enough to 23/40 to assume that some bars are weighted slightly more than others in the calculation of the overall score, but doesn't seem to be off enough to seem odd, or am I missing something?

[wade moore]

Originally Posted by QuikSand

I don't know, it's just a thought in progress.



I'm basically a fellow blindee here, I tend to make nearly all my judgments on future ratings movement, too. I'm just trying to spur a discussion on what else there might be. Perhaps I didn't pick a good case study, after all.

I think your goal here is a worthwhile one - I just wanted to make sure I understand what you were hypothesizing.

I'm trying to come up with a good way to test it without too many unnecessary headaches.

[QuikSand]

Originally Posted by jdavidbakr

If you simply take the average of the green/red bars you get a score of 28/41. To me that's close enough to 23/40 to assume that some bars are weighted slightly more than others in the calculation of the overall score, but doesn't seem to be off enough to seem odd, or am I missing something?

Yeah, I think I might be falling for some of the really low reds like hole rec and route running and dismissing the value of that remaining potential. Perhaps it's "nothing to see here" with this guy after all.

[QuikSand]

Even if it's true that this guy doesn't tell us much... I'd like to think that there is still some value in trying to better understand the three data sources available for a young player -- component ratings, overall rating, and position experience.

For example -- we have seen a number of cases documented where guys who seemed to be nearly fully developed in bars and ratings, but who had a lot more growth yet to come in their position experience. Many of these guys proved to be major creepers, continuing to develop along with their remaining positional development -- beyond what we would have expected from just the red/green presentation.

Maybe that's just getting to the same subset of players that we tend to spot with overall "creeping," again I don't claim to understand all this terribly well and thought a discussion might help. Maybe it was a bad idea to even post a sample player at all.

[remper]

My (possibly uneducated) guess is that considering his low volatility, and well developed experience at his position, he will gain a few current overall points, and lose a whole lot of future overall points. It seems as though the evidence should point to the number rating being BS, not the bars.

Then again, what do I know...

[QuikSand]

Ok, well for the sake of completeness with this guy, he had a -8 move in training camp... so he's not going to stick on my roster long enough for me to find out if he's worth a damn.

[stevew]

agreed about the fumble rating probably playing a bigger role than we make think in determining his future rating.

[TheMeat]

My take on this, and what i thought immediately upon seeing him, is that hole recognition seems to be the most heavily weighted bar for OVR. I based this on observation that I've seen guys with a high hole recognition and little else have decent OVR, and guys with low HR and high numbers in others have very low OVR.

Much like madden certain bars have much more weight in the OVR for each position. A question I had been meaning to bring up here is to ask whether anyone has figured out the exact weightings for each position? I know roughly what they are, but I'll start a new thread for that discussion later tonight.

[Subby]

Good thread Quik. Solid investment.

[RedKingGold]

This thread delivers.