More observations on player development

[QuikSand]

In my most recent FOF 2004 career, I’ve had a chance to watch fairly undeveloped players try to “blossom”—and have had occasion to see a number of players get better over a long period of time. (I have been playing an Island of Misfit Toys (http://dynamic2.gamespy.com/~fof/forums/showthread.php?t=28962) career, in which all my players have come from the undrafted rookie free agent pool)

While my primary purpose of playing has been for enjoyment and challenge, I have also developed a pretty fair amount of information about player development, and have an observation that seems worth noting here. I don’t think it’s a completely original observation, by any means, but having a very concrete example helps make it a little more helpful, I guess.


I have long held the general belief, which I first adopted after studying player development in FOF4, that the initial direction of movement in a player’s future potential skills is a very strong indication of his eventual potential. That is, following a rookie’s training camp, if his future potential (as measured by the scout’s aggregated score) dropped by a few points, that was a pretty good indicator that the player would never actually reach the apparent potential… and that a rookie like this who dropped from 20/50 to 22/45, for example, might never get to current ratings over 25 or 30 or so – but not even close to 45 or 50. This happens to a very wide swath of rookie players – so I felt like this insight was very useful in assessing young players. I still do – the unusual rookie player who exhibits growth in his potential ratings during his rookie training camp is often one with potential well beyond that which is initially visible.

In any event – I think it has always been pretty clear that this “rule” is more of a guide, rather than an absolute. There seem to be multiple moving parts to the player development routine – we know that playing time plays a role, mentors seem to as well, and that there are players who have mid-career adjustments in skills (to the bad or good) which don’t necessarily seem to be connected to anything else (at least I haven’t seen persuasive evidence of it). So, it’s not as simple as the preceding paragraph might have suggested – we mostly have agreed on this.


So, with all this in mind, I present CB Jeremy Abrams. I picked him up as a free agent in the second year of this IMT career, fishing for usable players at a tough position to fill in this kind of challenge. He looked pretty decent on paper, but had a lousy camp – dropping five points in his potential ratings, from 38 to 33. Ordinarily, I would probably chuck this player away, assuming that he had no real chance to develop because of this initial drop. But this team being what it was, I didn’t have any better options, so he became a starter right away. The team has matured since then – and here is how he has progressed:


Pos Exp CE FE CE FE dCE dFE GP GS
2031 Abrams, Jeremy LCB 1 21 38 23 33 2 -5 16 16
2032 Abrams, Jeremy RCB 2 30 33 32 33 2 0 16 16
2033 Abrams, Jeremy RCB 3 34 34 34 34 0 0 16 16
2034 Abrams, Jeremy RCB 4 34 34 34 34 0 0 16 16
2035 Abrams, Jeremy RCB 5 34 34 34 34 0 0 16 16
2036 Abrams, Jeremy RCB 6 49 49 49 49 0 0 16 16
2037 Abrams, Jeremy RCB 7 50 50 50 50 0 0 14 12
2038 Abrams, Jeremy RCB 8 50 50 50 50 0 0 7 7
2039 Abrams, Jeremy RCB 9 49 49 49 49 0 0 1 1
2040* Abrams, Jeremy RCB 10 59 59 59 59 0 0 15 15
2041 Abrams, Jeremy RCB 11 76 76 76 76 0 0 16 16
2042 Abrams, Jeremy RCB 12 70 70 70 70 0 0 16 16

* - In 2040, the team changes its lead scout


After a two-year injury scare, Abrams has returned to starting duty, and has played through some minor injuries but has yet to miss significant time. He also, incidentally, has been starting mostly at strong safety, as he has better run defense skills than my other two starting-caliber cornerbacks on this team.

After a few seasons of just being the best option on a lousy roster, Abrams has now developed into the highest-rated player on my team. His ratings in nearly every category relevant to a DB are right around 80, with the familiar “void” (a zero) in play diagnosis (and no return skills). This is a high-quality player, who seems to have had two significant development booms, happening in his 6th year and again in his 11th year. (The 10th year jump may have only been a function of a new team scout)

Regardless, he has defied my initial expectations dramatically – he has gotten loads of playing time, quickly reached his initial potential completely, and then has seen growth in his overall skill set – neither of which I ever would have expected to see.


Any more information needed? Volatility, of course. His rating? A 96 – very, very high. While it did not translate immediately to a rookie camp breakout (which I usually look for right away), it’s pretty clear that he has had some “breakouts” down the road… and that high volatility rating might be connected somehow.


I don’t offer a lot more than speculation here… but I thought a concrete example from a public career might be a useful contribution to our collective understanding about player development. Your thoughts and comments are, as always, welcome.

[albionmoonlight]

Do we know if this (gains and losses in green bars) is something that is hardwired into the player from the beginning, or if it is a function of a randomesque process that occurs to all players thoughout the game.

In other words, did the collection of ones and zeros that make up Abrams always have the potential to get to the 80s in it (which potential may or may not have ever been realized depending on mentors, playing time, etc.), or is there some algorithim that takes every player in the league and, for a select few of them, changes their ratings for the better or worse?

I suppose we could answer this question if you have the saved game from his rookie season. Run 10 years from there treating him about the same way and see what happens to him.

[QuikSand]

Alas - I don't have an old saved game to use for that noble purpose.

[MIJB#19]

Just a theory, but worth mentioning I think.

The player file creation help speaks of an overall rating and seperate skills (the red/green bars). Perhaps Abrams' overall skill is coded/randomized to be 80, but his skill bars are coded/randomized to all be in the 30s. Maybe things caught up to eachother and one of the two catches up over time?

To put into perspective, I created an OPU career with a 0 (zero) rated QB, but with 100 potential everywhere. the scouts kept giving him a 0 in overall skills, but per seperate skill the bars kept improving over the years. I haven't tested more with it, but I guess it's worth trying to see how the QB would turn out to be in terms of stats and scoutn ing reports over time.

[cthomer5000]

I think there is also a lot to be discovered about in-season player development. In one of my careers, I've had 4 or 5 players who have seen their future potential go up by between 2-7 points while their current rating rose as well.

Often these guys would even jump 2 points in current and future in a single week. I haven't examined the long-term meaning, but I imagine it has to be a very good sign.

[Celeval]

Here's what I've discovered so far (off the top of my head, since my notes are buried someplace in all the boxes we're moving with):

- There are potentials for each rating. These vary throughout the course of a career - I haven't been able to put the research in to determine how much quite yet.

- There is a percent developed rating for each position. This matches the 'Percent Developed' shown in the draft screen; but only for the primary position. A particular player will usually be developed at multiple positions, but at different rates - i.e. a FL may be 80% as a Flanker but 55% as a Split End. As a player continues to play at a position, these percent develops approach 100%. I believe (although have not proven) that playing someone out of position uses the % developed for that position, rather than their natural one... so someone who is 100% at Flanker but 40% at Split End will be a 40/100 rather than a 100/100 player... but someone 80% and 80% will be 80/100 on both. As a player develops at one position, he gains marginal experience at others... and more experienced players may be 100% developed at multiple positions. This also seems to be somewhat taken into account when moving a player between positions, but haven't figured out how quite yet.

- Certain ratings do not have this percentage apply, and are fully developed (Blocking Strength, for example).

- There is an overall accuracy of scouting level for each player. This is what I believe QS sees with the initial move of ratings. This rating starts out all over the board - usually between 60-140, I think - and is applied as a scalar to all potential stats, regardless of the scout. As a player gains experience, this rating approaches 100. RL examples: Ryan Leaf had an accuracy of 200... so his 10/40 rating seemed a 20/80. After two years, he had an accuracy of 120, so his 10/40 rating seemed a 12/48. Terrell Davis was a 60, so his 50/80 when drafted looked like a 30/48.

- There are a handful of scouting combination ratings that combine to place a mask over the ratings. This seems to be the major cause of different developmental percentages. i.e... if a running back has a percent developed of 50%, and potentials of 80/80/80 in Third Down Running. His scout masks might make that 40/80, 40/80, 40/80 look like 50/82, 37/77, and 25/60. His accuracy of 110 would then be applied, so we might see a 55/90, 40/84, and 28/66 set of ratings for this particular running back. This is (I believe) where scout quality comes into play, but again - haven't gotten that much research done.

- There is not a hidden 'Tackles' rating. :-D

- I have seen the potentials of players change, although I haven't been able to figure out why yet. Injuries are one cause, and it's possible that Volatility is another. I haven't been able to find a tie between Volatility and a high (or low) accuracy rating.

Off the top of my head - I'll try and find some of my notes and correct any mistakes I made.

Kevin

[cthomer5000]

- There is not a hidden 'Tackles' rating. :-D
Well, one man alone is staking that claim - despite probably the strongest correlation between one stat and one skill in the entire game.

[nero]

Is there a utility that tracks a player's development?

Or am I just blind and FOF does it on its own...

[Franklinnoble]

Is there a utility that tracks a player's development?

Or am I just blind and FOF does it on its own...
That would be a pretty sweet utility.

I think most people are just doing it manually...

I'm sure there might be a way to inject some automation into the process... like printing your scouting view rosters to a file, then importing them to Excel, then doing a comparison spreadsheet or something... but that's a lot of work.

[Bee]

I believe Primelord did a utility to track player development...

http://dynamic2.gamespy.com/%7Efof/forums/showthread.php?t=21697

[nero]

I just browsed that board...did the GUI version ever come out?

[Franklinnoble]

I believe Primelord did a utility to track player development...

http://dynamic2.gamespy.com/%7Efof/forums/showthread.php?t=21697
Color me stupid.... thanks, Bee..

[Bee]

I just browsed that board...did the GUI version ever come out?

Not sure...I never used it. You could send a PM to primelord and ask though.

[Celeval]

The reason I did a lot of the above research was to try and crack the scouting formula - got a good bit of the way down that line, but ran out of "free time" :)

Knowing the scouting formula would allow us to directly pull ratings out of the save file and have them be the 'game' ratings as opposed to the 'actual' ratings - which I don't want to track, personally.

[Cap Ologist]

I just browsed that board...did the GUI version ever come out?

No it hasn't, but it is very user friendly.

My advice:

1. Put the change tracker program in your C: folder. I believe that this is the default location for the print to file.

2. Open the roster window and then print to file.

3. Create a shortcut for the change tracker program, I keep it in the same folder as FOF2k4.

4. You have to have two files to use it. So, I usually print to file before training camp, after training camp, after preseason, after super bowl and label them 0000.txt, 0001.txt, etc

5. Run the change tracker program and enter the numbers you want to use.

6. The program will create a html file called player report, and then you can view it.

This has made a big difference in my career, and while a GUI would be nice, it's not necessary.


edited for typo

[nero]

thank you thank you thank you...

after fighting with this the better part of the last 6 hours it works...

very much appreciated :)

[QuikSand]

- There is a percent developed rating for each position. This matches the 'Percent Developed' shown in the draft screen; but only for the primary position. A particular player will usually be developed at multiple positions, but at different rates - i.e. a FL may be 80% as a Flanker but 55% as a Split End. As a player continues to play at a position, these percent develops approach 100%. I believe (although have not proven) that playing someone out of position uses the % developed for that position, rather than their natural one... so someone who is 100% at Flanker but 40% at Split End will be a 40/100 rather than a 100/100 player... but someone 80% and 80% will be 80/100 on both. As a player develops at one position, he gains marginal experience at others... and more experienced players may be 100% developed at multiple positions. This also seems to be somewhat taken into account when moving a player between positions, but haven't figured out how quite yet.

I didn't want this topic to just dissolve without some more comments on this in particular. I think I disagree with nearly everything in this paragraph -- I don't claim any particular "insight" but I think there's a much simpler explanation for all this.

As we know (or should), each player has a rating in every category, even for those that are not revealed for his position (e.g. every player has a rating for "sense pass rush" even though he's never going to use it unless he's a QB). Each position, sensibly, includes a summary of the ratings that are relevant for the position -- so we can focus on the things that the player will actually be called on to do in a game (therefore, don't bother to show "sense pass rush" for a WR or a DE).

The overall ratings are, as nearly as we can tell, just a weighted average of the actual ratings in the relevant category. I believe (but am not certain) that a player who somehow had exactly 50/50 ratings in every category relevant to his position would be rated 50/50 by the scout in the overall ratings. But regardless of the 1:1 correlation, it's pretty clear that the overall rating is simply a composite of the individual skills -- with some hard-wired (and invisible) weighting system to assign more weight to some things than others.

So, if you had a young player with lots of potential whose main position was SLB, the game probably looks at things like run defense, pass rushing, and zone coverage skills primarily. Some weighted average of those things along with the other stuff relevant to the mlb position -- endurance, special team skills, punishing hitter, etc -- all get combined into one overall number.

We also know that with playing time, young players develop their current skills -- they tend to increase toward their potential in those areas. Play this guy at MLB and he'll get better in something like "run defense." But if you play him at MLB, he will get better at "run defense" -- and even if you play him at SS, he'll get better at "run defense" along with everything else. I don't think development of skills is related to the position where the player is used on the field -- I think it's purely a function of playing time.

The calculation of the "overall rating," therefore, isn't (in my judgement) a function of time at the current position -- it's just a simple amalgamation of the current and future ratings. If the above player, still technically listed as a SLB, had played the whole year at MLB or WLB, he'd have the same development as if he had played at SLB... and his resulting current/future ratings would yield the same composite, independent of where he actually played.

What makes this a bit more complicated is the fact that the player's ratings are subject to change if you make a permanent position change. This SLB might be small enough to move to SS, for example -- and if you did so, you'd see some change in his ratings. His new composite would then be changes for two reasons -- first, because the game adjusted his individual ratings like "run defense" because of the position switch, but also because the composie rating now comprises different elements and weightings. Now, as a safety, the composite is likely to place more heavy weighting on pass coverage and interceptions than previously (when the player was listed as a linebacker). So, naturally, we'd expect to see a significant change in the player's composite ratings.


In short... I think the overall player ratings are just weighted averages of the individual skills that are important to the player's listed position. Those individual skillstend to develop with playing time, regardless of the position played. Permanent switches of position complicate all this, since everything is subject to adjustment.

[The Shadow]

What makes this a bit more complicated is the fact that the player's ratings are subject to change if you make a permanent position change. This SLB might be small enough to move to SS, for example -- and if you did so, you'd see some change in his ratings. His new composite would then be changes for two reasons -- first, because the game adjusted his individual ratings like "run defense" because of the position switch, but also because the composie rating now comprises different elements and weightings. Now, as a safety, the composite is likely to place more heavy weighting on pass coverage and interceptions than previously (when the player was listed as a linebacker). So, naturally, we'd expect to see a significant change in the player's composite ratings.


In short... I think the overall player ratings are just weighted averages of the individual skills that are important to the player's listed position. Those individual skillstend to develop with playing time, regardless of the position played. Permanent switches of position complicate all this, since everything is subject to adjustment.

Whoa, hold on a minute, here, Quik. Are you saying that if a LB was switched to SS, his "run defense" rating would stay the same, and his overall rating (say 45/60) would change only because the weighting is different?

So then if a RT was converted to a LT, shouldn't his overall rating stay the same (or very similar) since his "run block" and "pass block" ratings stay the same? I've observed significant reductions when making this kind of a switch.

Also, wouldn't it make sense that a MLB's "run defense" rating should be different than a safety's "run defense" rating? The MLB is much more likely to have to fight through a block by a lineman than a safety and the angles/speed is different.

I'm thinking that when a position switch is made, there is a penalty (or a bonus) in all of the ratings to some degree (a few might be no change) that make up the composite.

Just my 2 cents !! :cool:

[CraigSca]

I'm thinking that a certain value in a particular attribute can be an entirely different value somewhere else. i.e. a 45 run defense rating at SS may only be worth a 30 at MLB. Then again, I don't have any empirical evidence to show this...

[QuikSand]

I'm thinking that a certain value in a particular attribute can be an entirely different value somewhere else. i.e. a 45 run defense rating at SS may only be worth a 30 at MLB. Then again, I don't have any empirical evidence to show this...

That would make sense to me, but I can't find any evidence in the game to support it. I've look for supporting trends -- like in the example you offer, moving a guy from safety to Lb shoudl result in a reduction in his run defense skills... but (theoretically) an increae in his pass defene skills (since the rating shoudl be relative to the position). However, if you fiddle with moving players from position to position, I think you'll see that the only thing that guides the movement of ratings is the player's weight (and the breadth of the position jump, which only serves as a big negative -- e.g. moving from FB to WR always makes a big loss of ratings, where RB to WR might not be).

As dumb as it sounds, there is evidence to support the thinking that a 290-lb DE with a "run defense" rating of 70 and a 190-lb FS with the same rating would be equally effectuive against the run if used as an OLB.

[MIJB#19]

Whoa, hold on a minute, here, Quik. Are you saying that if a LB was switched to SS, his "run defense" rating would stay the same, and his overall rating (say 45/60) would change only because the weighting is different?

So then if a RT was converted to a LT, shouldn't his overall rating stay the same (or very similar) since his "run block" and "pass block" ratings stay the same? I've observed significant reductions when making this kind of a switch.Jumpingon the QS bandwagon here, which isn't something I often do ;), but if you change a RT into a LT and back to a RT, his ratings won't be back to his old RT ratings, your player is wasted forever, at least at that moment, maybe the mysterious ratings bumps can save the guy.

[QuikSand]

Sorry I missed this before... MIJB has more or less given my answer.

Whoa, hold on a minute, here, Quik. Are you saying that if a LB was switched to SS, his "run defense" rating would stay the same, and his overall rating (say 45/60) would change only because the weighting is different?

So then if a RT was converted to a LT, shouldn't his overall rating stay the same (or very similar) since his "run block" and "pass block" ratings stay the same? I've observed significant reductions when making this kind of a switch.

Also, wouldn't it make sense that a MLB's "run defense" rating should be different than a safety's "run defense" rating? The MLB is much more likely to have to fight through a block by a lineman than a safety and the angles/speed is different.

I'm thinking that when a position switch is made, there is a penalty (or a bonus) in all of the ratings to some degree (a few might be no change) that make up the composite.

But hey -- why let someone say it in twenty words, when a thousand will do?


There are two things that happen in this game. They are, I believe, separate.

One is the fact that the game has "penalties" for switching a player's permanent position. We basically know who this works - it's a "backdoor," in my opinion, to keep ruthless gamers from ekeing out silly advantages by shifting every single player's position whenever there is some modest advantage. Wholesale position-shifting doesn't really go on in pro football, and neither should it in FOF. I think we generally understand how this works, and why it's in there.

But I think we need to keep that apart from your issue (and Celeval's) about how player development works.


The second thing that happens in this game is that players' ratings develop over time. They develop much faster with active on-field playing time - we all can see that is true.

Your argument that there is some lenghty series of hidden "developed" ratings that connects to each different position, and varies based on where the player actually got his playing time, fails the simplicity test.

For every player, he has a series of current and potential ratings. Your scout gives you a pretty good estimate of the relevant ones for his position. Based on a hidden formula, your scout also combines these into an overall pair of ratings - current and future. And based on those two numbers, for rookie players, we also see a "percent developed" figure -- which simply equals the current rating divided by the future rating. Simple stuff, actually.

Set aside the confusion over how the game treats players being switched permently from one position to another, and this is really simple. A guy with a "run defense" rating of 20/80 is going to get better at that skill, whether he's lining up at DE, WLB, or SS. And that skill is going to get incorporated when the scout uses his hidden formula to determine the player's overall current rating, which then feeds into the "percent developed" figure.


I'm not trying to argue anything at all about "real football" and whether real world safties can step up and play linebacker with ease (and that sort of thing). I'm just trying to describe how player development seems to work in this game. There's also a separate discussion about how much of a "penalty" players incur when they are asked to play a position other than their formally listed role. Again - a separate discussion, I think.

So, I stick with the wrap-up from above:

I think the overall player ratings are just weighted averages of the individual skills that are important to the player's listed position. Those individual skills tend to develop with playing time, regardless of the position played. Permanent switches of position complicate all this, since everything is then subject to adjustment.

[Pyser]

then the question is, what kind of a hit does a player take for playing out of position. the discussion pretty clearly shows that a permanent position switch can ruin a players ratings. say, a 60/90 right tackle becomes a 47/77 left tackle. however, what happens if you just play him at left tackle? obviously, the player will improve his current value, no matter what...we've discussed that. but at what level does he play out of position?

and then the question on top of that is, is it ever really worth it to switch players positions permanently?

[Celeval]

Your argument that there is some lenghty series of hidden "developed" ratings that connects to each different position, and varies based on where the player actually got his playing time, fails the simplicity test.

I think the overall player ratings are just weighted averages of the individual skills that are important to the player's listed position. Those individual skills tend to develop with playing time, regardless of the position played. Permanent switches of position complicate all this, since everything is then subject to adjustment.
I agree with the second part here - the actual, displayed overall rating is a weighted average based on position.

But while the first one fails the simplicity test... I think it passes looking at the file in specific. :-D The original reason I was going through all this was trying to build a utility for player reporting, that would need to calculate the scouted ratings, and I found the player record to look like this (somewhat condensed for brevity - the ones important to this discussion are bolded):
0-1 Player ID
2 Position
3 Experience
4 Player Number
5 Position Group (determines which filter to display of stats)
6 Team
7 Unknown (0-49, steadily increases throughout career)
8 Injury Type, if any
9 Unknown (6-26, could be propensity for injury)
10 Percent Hurt
11 Injury Length
12 Unused? (Always 0)
13 Percent Developed
14 Unused? (Always 0)
15-23 Individual Ratings, 1-100 (Loyalty, Wants Winner, Volatility, etc)
24-27 Team Status (Franchise, Dates of joining team, etc)
28 Popularity
29-30 Scouting Variances
31 Observation/Accuracy of Ratings
32 ?? Statistic Type ??
33 ?? Statistic Type ??
34 HOF Points
35 # of Honors in career
36 Unused? (Always 0)
37-72 Contract Info
73-76 Unused? (Always 0)
77-80 Physical/Draft Info (Sol. Test, 40 Time, etc)
81 Unknown (0-24)
82-83 Unused? (Always 100)
84 Unused? (Always 0)
85-111 Percent Developed By Position (? 1-100)
112-150 Ratings (375-625)
151-209 Stats
209-251 Other Stuff, mostly unknown in the 0-4 range

The ratings are pretty easy to figure out. They stand out, since they are in the range of 375-625, well outside of the range of all other types. This has been standard practice for the FOF/TCY series - 375 equates with a 0 rating, 625 equates with a 100 rating. These are potential ratings - and as they change, the actual rating changes as well, proportionally.

The initial thought was that 85-111 would match up with these ratings as a 1-on-1 rating to actual percent developed; but there were a couple of problems with this. The biggest is that there are 27 1-100 numbers compared to 39 ratings. What they do match up with are the number of positions (27, considering LT and RT as different, etc). When comparing the positions of players with values in these fields, it becomes obvious - 85 only has quarterbacks with values above 0, 86 only has Running Backs, Fullbacks, Flankers, and Split Ends; 87 has RBs, FBs, TEs; etc.

The next giveaway is looking at these in relation to draftees - once positions are mapped against these values; the value for the primary position matches exactly with #13 (Percent Developed), which is listed in the draftee screen.

The final one is in modification - changing the *primary position* percent developed value will change the actual ratings nearly across the board. I say nearly since it seems some ratings (Blocking Strength, for example) are always at 100% developed.

So what about the secondary positions? Looking at players over the course of a career, not just the primary position changes. There seem to be secondary positions to each position that develop at a lesser rate. So if your C develops by playing C, he also develops (slower) at RG, LG, RT, and LT. From there, it was a straightforward test of playing someone out of position at a secondary position for a season, and watching the development take place - it indeed happened more along the secondary position than the primary.

I have only anecdotal evidence about changing positions permanently - changing a position of someone who is 100% developed at C and at RG will result in little to no ratings change. For someone who is 100% and 0%, there is typically a huge change - there is some randomness to this; really this seems to be what generates the initial message about "large experience reduction at the new position" or whathaveyou. There are tests I'd like to run on this to see how the actual potential ratings change, but I haven't had the time, honestly.

For the sake of saying so here - I have already said this to Jim and others with whom I have shared this in private - I have not and will not delve into any multiplayer files, and am sticking only to single player saves.

As anyone who has seen my Georgia Gridlock probably knows, cuz I suck.

Being able to mess around with individual ratings does confirm the second part of Quik's statement above - the overall rating is definately a weighted average by position. No idea what those weights are yet - especially since once scouting is placed over the basic (developed*potential) / potential formula, things get dicey quick.

[Celeval]

Corrollary to the above, re: player switching positions

Seems the best times to change a player's position is either very early or very late in a career. Early, before development has happened; or late, so that there is no experience hit. Which seems to match with the NFL - most changes seem to be either rookies coming into the league (Jordan Gross was RT in year 1, LT in year 2); or veterans falling into a different role (CBs becoming Ss late in the career).

[gstelmack]

Another comment on player development: one bit that makes it difficult in the game to predict if a guy is breaking out or not is the way scouting is applied to their ratings. There is a very clear difference between how your scout rates a guy not on your roster vs. how he is rated when on your roster. It's evident whenever you sign a free agent, but it also messes with rookies you are trying to draft. The instant you sign a draft pick, his ratings change. Rarely will you see insta-boom here, but I've definitely seen insta-bust. More typical are a few points up or (mostly) down.

But then the ratings change again as soon as late FA arrives. And then they change once training camp is run. I've now hit a point where I consider any movement within +/- 4 to be just perturbations of the scouting algorithm; there's not necessarily any actual movement, it's just a change in the accuracy of your scout's view of the player. Especially with draftees where there is such a range that the player could actually be in. Movement much over that is a good indicator of a boom/bust player.

Otherwise you just have to pay attention to their development, especially guys that increase/decrease during the season.

[QuikSand]

Celeval - I am officially fascinated.

Before I make a buffoon of myself -- I certainly agree that the rating 85-111 must correlate to the 27 different positions. That much seems obvious. And since you say that the entry corresponding to the player's "percent developed" for the player's current position, I'll buy that, too, of course.

But I think I fall short of agreeing with your original assertion from farther above (the segments I'd contest are underscored):

- There is a percent developed rating for each position. This matches the 'Percent Developed' shown in the draft screen; but only for the primary position. A particular player will usually be developed at multiple positions, but at different rates - i.e. a FL may be 80% as a Flanker but 55% as a Split End. As a player continues to play at a position, these percent develops approach 100%. I believe (although have not proven) that playing someone out of position uses the % developed for that position, rather than their natural one... so someone who is 100% at Flanker but 40% at Split End will be a 40/100 rather than a 100/100 player... but someone 80% and 80% will be 80/100 on both. As a player develops at one position, he gains marginal experience at others... and more experienced players may be 100% developed at multiple positions. This also seems to be somewhat taken into account when moving a player between positions, but haven't figured out how quite yet.

So, while I'm certainly buying your evidence, I guess I reach a slightly different conclusion.

I'd say that the hidden "percent developed" rating at a position like DE for a player who has been listed as an OLB is not used if that player is simply slotted into that position in the lineup. Rather, I think that rating exists only as a factor used if the player actually undergoes a permanent switch to that new position.

I'm thinking it basically works like this:

-Temporarily slot a SLB at DE, and the game uses some sort of hidden "penalty" rating to depress his stats (which had been indicated to be minor in nearly all cases)

-Permanently move a SLB to DE, and the game makes two adjustments:

#1 - the player's potential ratings are adjusted based on his weight

#2 - the player's current ratings are a function of the potentials, and his hidden "percent developed" at the new position



Anything in the annals of the data files that would contradict this?

[Celeval]

No, but it does make it straightforward to test when I have an hour.

1. Set up two SLBs with maxed out potentials
2. Set one with 100% development at SLB and 100% development at DE
3. Set one with 100% development at SLB and 0% development at DE

Run about ten seasons, five with each at RDE/LDE, then swap. Compare results.

Reasonable? Or should there be more put into this?

[QuikSand]

That would make sense to me.

[Franklinnoble]

Check this shit out...

http://dynamic2.gamespy.com/~fof/forums/showthread.php?p=564971#post564971

[QuikSand]

I think the post-draft immediate jump has been noted before... though that sounds like a pretty extreme case, I'll admit.

But gstelmack's observations seem pretty on target to me -- once you see an arrow pointing in one direction, that's a pretty strong indicator of where he's going.

That said... look where this thread started. A guy whose ratings jumped downward in his first camp, but ended up booming later in his career, after lots of playing time.

The plot... she thickens.

[Celeval]

Results seem to emphatically favor your assessment, QS:

Some info on the test runs:

- Using the initial roster provided with the game, with a few specific changes, detailed below.
- Cleveland Browns, without draft
- No X-Factor (doesn't matter, since roster editing happens after initialization)
- No personality profiles/team chemistry
- No hiring/firing
- Wall Street level
- No injuries

The following players have been edited to 100-ratings across the board in -all skills- (this includes non-defensive skills)
SLB Warren Mohon
SLB Bryant Kayer

Mohon and Kayer have also been edited to the following combinations:
Loyalty: 50
Wants Winner: 50
Leadership: 50
Intelligence: 50
Personality: 50
Popularity: 50
Volatility: 50

Mohon has been edited to 100% developed in both SLB, RDE, and LDE; 0% everywhere else.
Kayer has been edited to 100% developed in SLB; 0% everywhere else.

The scout view of these players:
Overall: 90/90
Run Defense: 90
Pass Rush Defense: 90/100
Man-to-Man Technique: 90
Zone Defense: 90
Bump and Run Defense: 84
Pass Rush Strength: 99
Play Diagnosis: 90
Punishing Hitter: 100
Endurance: 84
Special Teams: 84

Seasons 1-5: Kayer at LDE, Mohon at RDE, all at '9' playing time
Season 1
*Mohon: 64 tackles, 26 assists, 13.0 sacks, 3 blocks, 21 hurries (485 RP, 544 PP)
Kayer: 46 tackles, 29 assists, 11.5 sacks, 4 blocks, 28 hurries, 1 int, 1 defn (486 RP, 548 PP)

Season 2
Mohon: 79 tackles, 31 assists, 8.5 sacks, 6 blocks, 17 hurries, 4 dfn (498 RP, 550 PP)
*Kayer: 63 tackles, 28 assists, 15.5 sacks, 2 blocks, 24 hurries, 2 dfn (498 RP, 550 PP)

Season 3
Mohon: 63 tackles, 30 assists, 6.5 sacks, 5 blocks, 29 hurries, 2 int (462 RP, 549 PP)
Kayer: 59 tackles, 15 assists, 10.0 sacks, 4 blocks, 35 hurries, 2 dfn (464 RP, 549 PP)

Season 4
*Mohon: 81 tackles, 24 assists, 10.0 sacks, 7 blocks, 22 hurries, 2 dfn (498 RP, 538 PP)
Kayer: 50 tackles, 22 assists, 7.5 sacks, 6 blocks, 17 hurries (498 RP, 538 PP)

Season 5
Mohon: 69 tackles, 31 assists, 8.5 sacks, 5 blocks, 13 hurries, 3 dfn (523 RP, 504 PP)
*Kayer: 60 tackles, 28 assists, 14.5 sacks, 4 blocks, 25 hurries (524 RP, 503 PP)

Seasons 6-10: Kayer at RDE, Mohon at LDE

Season 6
Mohon: 45 tackles, 17 assists, 10.0 sacks, 3 blocks, 42 hurries, 1 int, 2 dfn (455 RP, 585 PP)
*Kayer: 78 tackles, 25 assists, 16.5 sacks, 5 blocks, 26 hurries, 2 int, TD, 1 dfn (455 RP, 584 PP)

Season 7
Mohon: 49 tackles, 17 assists, 10.5 sacks, 6 blocks, 22 hurries, int (474 RP, 487 PP)
Kayer: 61 tackles, 25 assists, 12.0 sacks, 10 hurries, 1 dfn (485 RP, 488 PP)

Season 8
Mohon: 61 tackles, 17 assists, 13.0 sacks, 3 blocks, 28 hurries (473 RP, 549 PP)
*Kayer: 66 tackles, 34 assists, 12.5 sacks, 3 blocks, 26 hurries, int, TD (475 RP, 545 PP)

Season 9
Mohon: 49 tackles, 25 assists, 7.5 sacks, 6 blocks, 26 hurries, 1 dfn (498 RP, 533 PP)
Kayer: 67 tackles, 25 assists, 9.0 sacks, 10 blocks, 19 hurries, 1 int, 2 dfn (497 RP, 531 PP)

Season 10
Mohon: 66 tackles, 23 assists, 14.0 sacks, 5 blocks, 23 hurries, 1 dfn (492 RP, 517 PP)
Kayer: 66 tackles, 19 assists, 5.5 sacks, 7 blocks, 20 hurries, 2 dfn (494 RP, 518 PP)

Totals over 10 Seasons:
Mohon: 626 tackles, 241 assists, 101.5 sacks, 49 blocks, 243 hurries, 4 int, 13 dfn (4858 RP, 5356 PP)
Kayer: 616 tackles, 250 assists, 114.5 sacks, 45 blocks, 230 hurries, 5 int, 11 dfn (4876 RP, 5354 PP)



Some interesting things out of this... first, the differences between the seasons and the players is small enough that I think we can rule out differences in the development per position having any effect while playing out of position. I suppose the next test (not happening tonight) is to compare a maxed-out SLB and a maxed-out DE.

Secondly, it seems that the RDE will consistantly get more tackles, while the LDE will have more success rushing the passer. Good to know.

Kevin

[QuikSand]

Good stuff... thanks for taking the time.

[Vince]

Excellent work, Kevin.

[Alf]

Secondly, it seems that the RDE will consistantly get more tackles, while the LDE will have more success rushing the passer. Good to know.
Thanks Kevin ! I already noticed this latest info thanks to my IHOF Vineyards as I have 2 very similar DEs. That's a useful confirmation.

[QuikSand]

Secondly, it seems that the RDE will consistantly get more tackles, while the LDE will have more success rushing the passer. Good to know.

Wonder if this is built into the game directly, or if it's perhaps just a by-product of a left tackle bias. Since most teams try to place their best OL at left tackle (or at least the better of their starting tackles there) there would be a tendency for (1) that guy to give up fewer sacks to the RDE thean his lesser counterpart would to the LDE, and (2) the team to probably want to run more behind their standout tackle than the lesser guy on the other side, creating more tackle opportunities on that side.

I would guess that the same biases probably apply to a large extent in the multiplayer environment, too.

[Buzzbee]

Jumpingon the QS bandwagon here, which isn't something I often do ;), but if you change a RT into a LT and back to a RT, his ratings won't be back to his old RT ratings, your player is wasted forever, at least at that moment, maybe the mysterious ratings bumps can save the guy.

Is this true? I haven't done many position changes but I was under the impression that if you changed from RT to LT and back to a RT, his ratings would be the same, or if different, only very slightly.

Based on what Celeval has so awesomely exposed, it seems that if there is a difference in the pre-switch and post-switch (back to original position) it *might* be attributed to scout error. The scout would have to "evaluate" the player during/after each switch to come up with the ratings.

If the rating differences are indeed large in a RT-LT-RT switch, then it would support the idea of some sort of "penalty" for EACH switch. If the differences are repeatedly small, or non-existent, it seems the the "penalty" would simply be that the player is less developed at the new position.

One other point, if the RT-LT-RT switch does result in modest differences (1 or 2 points) then this might give some insight in to scout error and the margin. I would assume an excellent scout would see smaller variation between pre-switch and post-switch than a poorly rated scout.

One last point, if the pre and post RT-LT-RT ratings are always the same, then it might suggest the scout's view of that player AT THAT POINT IN TIME is static and only changes at certain "checkpoints" during the season.

Just tossing out some ideas.

[QuikSand]

Is this true? I haven't done many position changes but I was under the impression that if you changed from RT to LT and back to a RT, his ratings would be the same, or if different, only very slightly.

Based on what Celeval has so awesomely exposed, it seems that if there is a difference in the pre-switch and post-switch (back to original position) it *might* be attributed to scout error. The scout would have to "evaluate" the player during/after each switch to come up with the ratings.

If the rating differences are indeed large in a RT-LT-RT switch, then it would support the idea of some sort of "penalty" for EACH switch. If the differences are repeatedly small, or non-existent, it seems the the "penalty" would simply be that the player is less developed at the new position.

I have seen cases that seem consistent with either theory, but my inclination is that there is a penalty applied for each switch. I have fiddled around quite a lot with "big" shifts in position (not just LT to RT, but RB to WR or ILB to DE) and in those cases, it's very common that the player is just ruined by the switch - even if he can switch back to his original position, he's immediately a shadow of his former self. So, I'm inclined to believe that in the cases where a player switches over and back harmlessly it's because he is just a good fit for the switch, and isn't really suffering a penalty.

Not an especially firm opinion, but I don't see any real insight into scout error from this.

[Celeval]

I don't have ratings in front of me to prove it, but this jives with what I've seen. The actual potential ratings do take hits on position changes.

[QuikSand]

So, here's a developing theory then, based on Celeval's postings here.

Let's say that for a given player, the information about his current ratings is not actually stored, piece by piece, in the game. Only the potential rtaings are stored this way. (That is strongly suggested from the above)

However, I'm pretty darned sure that I have seen players who are much more developed in one skill than another -- suggesting that there might be some kind of overall adjustent for each skill -- something that is a constant for all players, or at least all players at a position.

For instance -- If your player has a potential of 100 in run defense, and he is 20% developed, that does not mean that he will be rated as a 20 in current skill. But wnat does it mean?

We might have enough information to uncover this. I'm not sure exactly what it would give us... but it might be useful. Perhaps you could "back out" the expected current ratings for a rookie in the draft pool?

(I'm also very much open to the idea that there is a big hole in this theory -- I'm just tossing it out)

[Buzzbee]

The following players have been edited to 100-ratings across the board in -all skills- (this includes non-defensive skills)
SLB Warren Mohon
SLB Bryant Kayer

Mohon has been edited to 100% developed in both SLB, RDE, and LDE; 0% everywhere else.
Kayer has been edited to 100% developed in SLB; 0% everywhere else.

The scout view of these players:
Overall: 90/90
Run Defense: 90
Pass Rush Defense: 90/100
Man-to-Man Technique: 90
Zone Defense: 90
Bump and Run Defense: 84
Pass Rush Strength: 99
Play Diagnosis: 90
Punishing Hitter: 100
Endurance: 84
Special Teams: 84


Some additional observations. I re-read through this a little more thoroughly and noticed a few things I didn't before. Kevin said he maxed out the potential ratings in all categories. My assumption (although Kevin didn't specifically state it) is that #13 - percent developed was set to 100% for both players, and that the % developed by position was adjusted as listed.

Therefore my expectation was to see 100/100 for all ratings. We can see that isn't the case. Overall, Run, Pass, M2M, Zone, and Diag are all 90. Even the Overall isn't 90/100, but rather 90/90. That would lead me to believe there is some other modifier afecting what we see. This could be Scout Error, or perhaps some other factor that "boosts" players over the 90 mark.

Pass Rush Strength and Punish Hitter are both basically 100 leading me to believe that these are strictly potential*%developed. In addition, Endurance, Special Teams, and mysteriously Bump N Run are all at 84. Rather odd grouping, so perhaps the scout errs by different amounts in different categories.

One other curiousity. Both players had the exact same ratings as presented by the scout. That would seem to indicate that scout error is very formulaic, and not random at all. Two players with idential ratings and positions are seen by the scout exactly the same. If the scout view was random I would have expected slight variations in a few, if not all, of the categories.

Just some additional thoughts. My apologies if I'm stating the obvious.

[QuikSand]

Maybe things like the combine data have some effect on the player's overall ceiling at various tasks -- perhaps this guy's strength, speed, or intelligence are what's placing limits on his ratings, despite the maxed-out editing going on...

[Celeval]

I'm almost 100% certain it's scout error - there are two numbers (29-30 in my post above) that determine a particular variance in the scout error. If I move those numbers around, the scouted ratings change. This matches with the formula we had for the TCY scouting (ref: TCY Helper) - the "masks" will shift in location as well as degree. These guys both had identical scouting variances (forced - meant to mention that before, but didn't), as I wanted them to appear equal as well as actually be rated equally.

The combine data is a good thought - I've always figured that the combine data was generated from the ratings, and had actually been working on a different post to try and draw parrelels (i.e. - the weight lifting usually has a pretty good correspondance with Blocking Strength for linemen, but what about for RB? QB? etc...), but that data will get more interesting with this as a thought.

[QuikSand]

I found the player record to look like this (somewhat condensed for brevity - the ones important to this discussion are bolded):

112-150 Ratings (375-625)


Okay, I'm willing to start a new tangent here, in what has become, to me at least, a pretty interesting thread (that nobody reads).

If player ratings are stored in the game on a scale from 375 to 625... isn't that a little weird? Why not just 0 to 100?

Here's a speculative guess: The scale from 375 to 625 ends up being the actual measure of "effectiveness" at a given skill

What the hell does that mean? I'll try to elaborate...


Say you have two guys -- let's say they are offensive linemen. One has ratings of 40/40/40/40 and the other has 80/80/80/80. How much better is the 80 guy than the 40 guy? Twice as good? How would that translate in the game? (Another way to ask a related question is -- what does a "zero" rating mean? Surely this NFL-caliber player doesn't have paraplegic skills in some relevant area, right?)

Maybe the formula looks like this:

AR = Apparent rating (1-100 scale)
ER = Effective Rating (375-625 scale)
and therefore ER = AR * 2.5 + 375

Which means that ouy two guys, with flat rtaings of 40 and 80 wouldn't have "effective ratings" that are in a 2:1 ratio, but rather something like this:

AR 40 --> ER of 475
AR 80 --> ER of 575

This would mean that the guy with ratings of 80 should perform (in whatever way the game cheks these ratings for performance -- a slippery measure, we'd all agree) something like 21% better than the guy with 40s across the board. (575 / 475 = about 1.21)

Where does this lead us... tough to say. Maybe it should, however, make us a little less appeciative of the guys who have a rating of 20 in a skill rather than zero.


Something to ponder...

[dixieflatline]

Maybe the formula looks like this:

AR = Apparent rating (1-100 scale)
ER = Effective Rating (375-625 scale)
and therefore ER = AR * 2.5 + 275

Which means that ouy two guys, with flat rtaings of 40 and 80 wouldn't have "effective ratings" that are in a 2:1 ratio, but rather something like this:
AR 40 --> ER of 475
AR 80 --> ER of 575

I think this makes a lot of sense. I haven't done a study but I think it's pretty clear that the player who rates at 80 isn't twice as good as the player who rates at 40. Still, Jim could easily do this math inside FOF and not inside the player file. I would not be surprised if Jim converts back to 0-100 AR scale and then uses some different formula to find a new effective rating. The scale 325 to 625 could be a red herring for people trying to figure out how FOF works.

[Pyser]

thats a very interesting theory. definitely changes the way you look at players rated 0 in some areas. one great example of this, in fact, is a good guard in my MP league rated 0/0 in run blocking. yet last season he was something like 30/80 on key run blocks. so obviously he has SOME skills at run blocking.

the only way this info might not match up is kick and punt return stats.

which leads me to my own theory, actually. again, in my MP league, there are 2 stud WRs. one is topped out in almost every rating, and is rated 84/84. However, another WR with very similar ratings (a bit worse, but very high overall on average), has ZERO return ratings, and is only rated a 57/57.

this makes me believe that when a player is rated overall, EVERY rating is included, not just the ones necassarily relevant to his specific position. as such, they can be misleading.

[Buzzbee]

Just a little additional info that seems to be related. This is from the Help file for the Draft File generator.

Ratings in each category are on a 1-9 basis. These ratings are not exact. When a new career begins, Front Office Football uses the ratings in the player file only as a general guideline. Internal ratings are on a 0-1000 scale, and are partially determined by random chance. This adds to the replay value of the game, as you can not count on a player being a superstar or a dud in each new career.

Ratings are divided into position groups. While every player potentially has a rating in every category, his ratings outside of his position group will usually be somewhere below a "1".

Scramble Frequency - desire to abandon the pass and scramble downfield. Unlike most categories, this rating is not assigned on a logarithmic curve.


And some more nuggets from the actual Draft File Generator itself:

Ratings in each category are on a 375-625 basis. These ratings are not exact. When a new career begins, Front Office Football uses the ratings in the draft file only as a general guideline. Internal ratings are on a 0-1000 scale, and are partially determined by random chance. This adds to the replay value of the game, as you can not count on a player being a superstar or a dud in each new career.

Here's a list of the frequency Overall Player Ratings are assigned in the default data file:

600-625 - Rare Superstar - 0.3%
575-599 - Superstar - 1.4%
550-574 - Near Superstar - 1.8%
525-549 - Elite Starter - 1.7%
500-524 - Excellent Starter - 3.5%
475-499 - Very Good-Plus Starter - 4.3%
450-474 - Very Good Starter - 6.0%
425-449 - Good Starter - 16.2%
400-424 - Replacement-Level Starter (decent backup) - 27.2%
375-399 - Roster Filler - 37.6%


Use 375 for players who really shouldn't have ratings at that position. For instance, defensive tackles should receive 375 for the Avoid Interception rating.

All players are rated for all categories, though these ratings are ignored if not relevant to the player's position.

Hope this helps

[QuikSand]

which leads me to my own theory, actually. again, in my MP league, there are 2 stud WRs. one is topped out in almost every rating, and is rated 84/84. However, another WR with very similar ratings (a bit worse, but very high overall on average), has ZERO return ratings, and is only rated a 57/57.

this makes me believe that when a player is rated overall, EVERY rating is included, not just the ones necassarily relevant to his specific position. as such, they can be misleading.

I'm not sure I follow -- are you suggesting that return ratings are not relevant to the receiver position? (I'd disagree)

[Celeval]

Makes sense all around - Jim's used 375-625 as long as I've been looking at his files, and hasn't hid it at all; BB's post shows that. I wouldn't be surprised if the calculations were actually done with 375-625 in mind.

As for the logarithm - I think that's just a question of distribution. Whereas for most ratings, the higher the better - so there are less players with 100 in Blocking Strength than with 0 - Scramble Frequency is more of a measure of tendency, so it's evenly distributed.

[Pyser]

I'm not sure I follow -- are you suggesting that return ratings are not relevant to the receiver position? (I'd disagree)

im saying return ratings ARE factored in to the overall rating. so a wr who is rated 90 in everywhere but in returns might not grade as high as say, a wr who is rated 60 at everything, including return stats.

which, incidentally, is useless to me. id rather see the players rating be based on his main relevant skills, and find my return men seperately.

my point about the return stats being different is this:

the guard i talked about that is rated 0/0 at run blocking, who had a solid season doing just that, we can assume has skills at run blocking that arent showing up (supporting your "effective rating" theory).

where i think the effective theory MAY be proven wrong is in return stats, where a player rated 0/0 might not actually have any skills there. the only way to find out, i guess, would be to plug in some players with no return skills to speak of, and see what they do.

which, now that i think about it, may take us back to the beginning. this all started, i believe, because you had a rb switch to wr, and suddenly his return skills showed up. so, were those ratings there all along?

i think im taking us off topic, though...i just wanted to throw that theory out.

[Buzzbee]

I'm a little confused, or behind the curve (par for the course) and am hoping someone can clarify a little for me. In my above post I see the following statement:

Ratings in each category are on a 375-625 basis. These ratings are not exact. When a new career begins, Front Office Football uses the ratings in the draft file only as a general guideline. Internal ratings are on a 0-1000 scale, and are partially determined by random chance.

Does this mean that the ratings from 375 to 625 are then converted to a 0-1000 scale? Vice Versa? I'm sorry if I'm impeding progress. Just trying to keep up.

[Subby]

Maybe the formula looks like this:

AR = Apparent rating (1-100 scale)
ER = Effective Rating (375-625 scale)
and therefore ER = AR * 2.5 + 275

AR 40 --> ER of 475
AR 80 --> ER of 575
Small thing, really - but I don't follow the math here. According to your formula, an AR of 80 would give you 475, not 575. I am probably just missing something completely obvious.

[QuikSand]

According to your formula, an AR of 80 would give you 475, not 575. I am probably just missing something completely obvious.

Quite right. The "completely obvious" thing was my math error. (Carry the one, fool) Corrected above.

[Subby]

Quite right. The "completely obvious" thing was my math error. (Carry the one, fool) Corrected above.I think the forrmula has to be changed to ER=AR*2.5+375, however...that would allow for players with a 0 AR in a rating to translate correctly on the 375-625 ER scale.

[Subby]

dola

If every player has a rating in every category, does a 375 ER (0 AR) in run defense for a QB equate to a 375 ER (0 AR) in run defense for an OLB? And does it even matter since it would be impossible for a QB to play OLB or any defensive position...

[CraigSca]

I've found this to be a very interesting read as well, as I'm a big fan of this entire genre. I think the range of 375-625 is very interesting and may have been a natural outcome of Jim's thorough study of the game. As other shave said, just because a guy has a "0" skill in a particular category does not mean that it's like wheeling a paraplegic onto the playing field.

Perhaps in his study of play-by-play he determined the truly elite players were generally 62.5/37.5 ~ 1.67 X better than the worst player. Meaning...the bad player would allow on average approx. 5 yards per carry (just throwing numbers out there) in his area of assignment, while an elite player would allow ~ 3 yards. An elite quarterback would complete 67% of his passes while a poor quarterback would complete 45%.

This range may just be a compromise that Jim found across all the different skills. Surely, 1.67 can't be an overall defining figure (I'm sure there are some skills that fall in or out of this range in RL), however, for a game like this it probably fit very well.

Now - the odd part that probably means absolutely nothing - 625+375=1000. People will probably find some significance there, but I'm thinking any significance found would purely be of an imaginary ilk.

[CraigSca]

dola

If every player has a rating in every category, does a 375 ER (0 AR) in run defense for a QB equate to a 375 ER (0 AR) in run defense for an OLB? And does it even matter since it would be impossible for a QB to play OLB or any defensive position...

I'm guessing it would be the same, but never comes up because switching a skill position to a linebacker would essentially make a "bottom-skilled" player in a number of categories. You'd basically void the player of all his talent.

[Subby]

I'm guessing it would be the same, but never comes up because switching a skill position to a linebacker would essentially make a "bottom-skilled" player in a number of categories. You'd basically void the player of all his talent.
I guess that's where the whole rating system potentially breaks down for me then. According to this rating system, a MLB who is 0 against defending the run is no more talented in that area than any kicker...but maybe Jim's safeguard to this is that those kickers would never be put in a position to have to stop the run, so their defending the run ER is more nil than 375...

[CraigSca]

Well, I'm obviously not an expert at this, it's just a guess on my part. Interestingly, what Jim COULD be doing is not use the 375 to 625 system unconditionally. For instance, anything less than 400 could be considered nil, where 400 is actually a zero skill at that position. In that case, a zero is obviously better than a nil. Just throwing it out there but I really don't believe this to be the case.

Would be interesting to find someone convert a placekicker and start him at MLB. If he gets any kind of tackles at all (beyond the accidental tackle here and there you'd expect from a RL move like this) then probably a 375 rating for a PK at tackling IS the same as a 375 for a real MLB.

[CraigSca]

Dola - ah, I see you can't even convert a PK to a linebacker. Oh well...so much for the experiment.

[CraigSca]

dola again...

On a personal level, I have no problem with a 375 to 625 range that treats a placekicker and a MLB with a 375 tackling as equal. If it's true (like I posted above), that a superior player is 1.67 times better than a poor player (based on Jim's findings) then Jim can only make player assumptions on the bottom level of the scale based on the statistical data he has. This data never had a placekicker play as a linebacker. Rather than "guessing" and increasing the range based on this guess for an extremely rare scenario, it makes more sense to tighten up the code (prevent this from happening) and the range (as it represents empirical data, not theory).

[Buzzbee]

I guess that's where the whole rating system potentially breaks down for me then. According to this rating system, a MLB who is 0 against defending the run is no more talented in that area than any kicker...but maybe Jim's safeguard to this is that those kickers would never be put in a position to have to stop the run, so their defending the run ER is more nil than 375...

Subby - perhaps a different perspective might help you reconcile the ratings system. It seems like Jim said at one point that a 0 rating didn't mean that a player was totally devoid of talent, but just that the player didn't have professional level talent in that area. Think of a college player who isn't good enough to make it on a pro roster. He can still punt, pass, kick, tackle, catch and all that good stuff, just not consistently at a pro level. So, in MY mind, a 0 rating simply means that they don't possess professional level skill in that area, not that they are simply unable to perform that function.

So, a MLB with 0 run defense can still make tackles, but just not as often or as effectively as someone who has higher ratings. Kind of the ol' even a broken clock is right twice a day kind of thing.

When you think of it that way, it allows a MLB with a run defense of 0 to be comparable to a K. They both have some skill at making a tackle, just not at the professional level. Hey, kickers make good tackles every once in a while. Not often, but they do tackle. Likewise with a QB who throws an interception.

[Subby]

Buzzbee...

Yeah - I think that, coupled with the position switching limitations - lead me to believe that that ratings hold up pretty well.

Now we just have to figure out what Punishing Hitter actually means... ;)

[Buzzbee]

Was thinking about this in preparation for the upcoming IHOF season. It seems that Celelval has shown that playing a player out of position will use the ratings for the primary position. Therefore, it seems like the only reasons to actually change a position would be if you think that switch might result in improved ratings, or if you have to make a switch to fill a roster requirement (need 2 ILB but only have 1).

I seem to recall that there was some speculation that playing out of position increased the risk of injuries. I may have even read it in the help file, I can't recall. Does anyone know if this is the case? I'd be curious to see if there is a greater injury risk and if so, how much greater? Significant? Minimal?

Anyone have any thoughts or experience in this area?

[Alf]

With my "poor" results in switching positions, I'm more and more likely to play a guy out of position than switch now...