FOF4 - Projecting player development

[QuikSand]

Projecting Player Development in FOF4

After playing the Front Office Football 4 game for some time (including my latest stint being involved with the GroupThink exercise), I feel increasingly confident that we have found and iosolated a major breakthrough in projecting the future development of rookie players in this game. For those of you who are frustrated by seeing many of your young players never reach their projected potential, this “strategy” will completely revise the way you play the game – and these disappointments will be much less frequent.

WARNING: The approach I suggest has a downside. In the FOF4 game, if you select your players using the methods I describe here, you will consistently do a far better job than your computer team rivals. This will, undoubtedly, add to your advantage over the computer teams, thereby making the game much easier to defeat. If you play by house rules, you will almost certainly need to rethink them, as any balanced set of rules employed without this drafting strategy will not be tough enough after you start picking your players this way. So, if you would rather not fundamentally change the way you play FOF4, consider this a “spoiler” warning, and don’t read anything further in this thread.


Caveat: The analyses here are gained from playing the commercial release of FOF4 – I have no special insight into the game’s coding or assembly. My comments are all based on observation from play of the game as released, both pre-patch and post-patch.



EXECUTIVE SUMMARY

-Rookies in FOF4 are frequently masked to disguise their true future potential
-The nature of the masking can be largely assessed by understanding key ratings
-Revealing the masking will minimize busts and maximize breakouts among your rookies


What is “masking” all about?

In previous versions of the FOF games, we were used to seeing rookie players with a variety of red ratings (reflecting current skills) and green ratings (reflecting future potential). In FOF2, these ratings were essentially reliable, as players practically always developed toward their “billed” potential. In FOF 2001, the game introduced “booms” and “busts” – a fairly small category of players who either suddenly exceeded their projected potential (booms) or suddenly lost much or all of their projected potential (busts). In FOF 2001, these booms and busts were revealed (in full) during the players’ first training camp.

Note: there has been much written on the boom/bust phenomenon in FOF 2001, many has been linked in the FOF Journal, though the veracity of individual links there may now be questionable. I won’t rehash those discussions here.

In FOF4, there is a more subtle and elaborate system of “masking” that is used. I don’t think it’s fundamentally different from that used in FOF 2001, but rather it’s just much more comprehensive – rater than affecting something like 5-10% of players like in FOF 2001, the masking system seems to affect practically all rookie players in FOF4 – a very different environment for those making the shift to the newer game.

In essence, what I call masking is this: players are revealed to us with an extra layer of green ratings. The real player might have “true” ratings of 20/40 in a certain category, but what we see is that set of ratings plus a certain mask that makes him appear to be, for instance, 20/80. Of course, a 20/80 player looks a lot more inviting than a 20/40 player – and is likely to be selected much earlier in a draft. This is how the masking system lures us (and the computer teams, too) into selecting players who turn out to be busts. Over time, that mask will disappear, and you’ll be left with the 20/40 player he really was all along – and wasn’t worth the draft pick that was probably based on his 80-point potential.

In my judgment, nearly every player in FOF4 is applied a mask – it’s just the magnitude that is the variable. Plenty of players have tiny masks – maybe only 2-10 points, which don’t amount to much distortion. Many have much more serious masks, which cause huge mistakes in judging their future, unless you properly reveal the masking. I also believe that some players (a fair number, but a distinct minority) have a negative mask, which works in essentially the opposite fashion. (More on that later)

So, the essential underlying themes of the “masking” concept are:

-For rookies, the red ratings is basically real; but
-The green rating is some combination of real potential and phony masking


How to Unmask the Rookies

What we need, then, is some way to find out how much of the player’s projected ability is real, and how much comes from the masking. Fortunately (for us), there is a fairly simple way to do so. The key insight here is recognizing that some of the ratings have changed in nature – and that change is simply as a result of the masking process in the game.

Here is the content of my previous post regarding “formerly static ratings”:

A while back (between patches 1.0a and 1.0b), there was some discussion about some differences in ratings that seemed to happen after the FOF4 game had been patched. I was just re-installing my FOF4 onto my new CPU, and took a moment to look at the pre-patch game, just to check this.

So, for what it's worth, here is the list of skills which (in the pre-patch FOF4 game) were purely "static" - they always showed as a single number, rather than a pair of red/green ratings. In each case, these ratings are now more or less like everything else - players will (frequently) have both current and potential ratings.

Sense Rush
Scramble Frequency
Breakaway Speed
Power Inside
Speed to Outside
Blocking Strength
Big Play Receiving
Pass Rush Strength
Punishing Hitter

Feel free to use this knowledge as you see fit.

So, the key to all this is that when you see one of these “formerly static ratings” showing up with both a red and green number, all of that green is the result of the player’s masking.

In other words, if you have a QB with ratings as follows:
Sense Rush: 20/40
Scramble Frequency: 50/62

What you are really seeing there is a QB whose true ratings are just 20 and 50. The green ratings – suggesting that he can improve in these areas – are simply revealing to you the player’s mask. In this case, we should be very comfortable in assessing that this player is really rated 20 and 50 in these categories, and that over time, those green ratings will go away – and that’s what we’ll be left with.

Simple enough, but how do we use these few ratings to figure out the rest of the player’s skills?


Applying the Mask Across Ratings

After some study, it finally became clear that the masking system was not haphazard, and not random. No, indeed, the masking system in FOF4 is fairly formulaic – perhaps it is so in order to save space required to keep all this (hidden) information compressed into a single item in the player file. (Those people who were working on draft files and so forth found a number of data points that were “unexplained” – I’m guessing that this is among them, in some fashion)

Without belaboring the matter too much, here is how the masking appears to work. (I’ll confess that the numbers do not always come out perfectly, but when we factor in scout error that we know is in the game, the results are nearly always very close – close enough that this is very usable in the game, with high confidence)

Masking for a given player is a fixed percentage of the difference between the player’s real potential rating and 100, applied (more or less) evenly across all ratings.

Or, in mathematical terms:

Masking = [100 – (real potential rating)] * M% (where M is a constant across categories)


Let’s look again at that QB from above – who has two formerly static ratings:
Sense Rush: 20/40
Scramble Frequency: 50/62

In this case, we know (by the nature of these ratings) that the reality is 20/20 and 50/50.

With a real potential of 20, the masking increases that to 40. That’s 20 out of the 80 points by which he fell short – meaning that his mask percentage is 25%. The number, in this example, also works out for the second rating: his real rating is 50, leaving 50 points to get to 100, so his masking there is 12 (or the same 25%).

If we accept that masking is basically according to the formula above, and that the formerly static ratings reveal it to us as truth, then we now have all the tools we need to assess this player top to bottom.

What if he shows a rating of 40/70 in avoiding interceptions? How good is he really?

Since we have M=25% in this case, we can back this out:

Masking Effect = [100 – Visible Potential] * [M% / (100-M%)]

In this case, that means:

Masking Effect = [100 – 70] * [25% / 75%] = [30] * [1/3] = 10 points

…in other words, 10 points of this guy’s potential is phony, a result of the masking. Therefore, his real ratings, without masking, would be more like 40/60. And over time, that’s what he will show himself to be.

In that case, the effect isn’t so severe – the player has pretty decent potential as is, and only loses a little bit of it. He’s not a complete waste, though if he loses a shade of every potential rating he has, that probably means he’s not quite as good as advertised… a minor bust, perhaps.


Let’s take a more serious case. Here are the ratings for a hypothetical rookie offensive lineman:

Run Blocking: 10/55
Pass Blocking: 30/65
Blocking Strength: 20/60
Endurance: 40/70

Outwardly, this is a guy you might consider with a middle-round pick. He projects to be a pretty solid player, and signing him for a few years might be very worthwhile. A guy rated 55/65/60/70 in FOF4 is a valuable contributor, and that’s what this guy projects to be.

Not so fast…

Look at the “key” rating – from the list of formerly static ratings: Blocking Strength. This is a rating that in reality is a single number, with no capacity for growth. However, here we see his number as 20/60, suggesting that he has a great deal of potential there. Player’s don’t have real potential in the formerly static ratings. Eevery bit of what you see there is phony, it’s the masking.

This player, in fact, has no growth potential at all. He has a masking of 50% in every area, and when that is lifted, we’ll find that this guy is, in fact, already at his zenith: he is a 10/30/20/40 player. That is what we would otherwise call a “stiff.” And cases like these are not fiction – they are very common in the game. There are plenty of players out there who look good, never develop much at all, and get cut after a disappointing year or two. That’s what is going to happen to this guy, on the hands of whoever drafts him in the third round. If you haven’t been watching key ratings and understand masking, that could have been you.


So, that’s the basic package of information. Players are masked with a flat percent of the remaining (unachieved) potential. You can use the key ratings (those previously static) to easily measure the magnitude of the masking in effect, and therefore measure how much of a player’s apparent upside will never materialize. This can go a long way toward avoiding costly draft busts.

Without belaboring the discussion, I will note here that the masking does not reveal itself immediately, like in FOF 2001. Rather, it takes a period of time to dissolve – seems to be about five seasons. Players with substantial masking (like the example OL above) will still show some potential in their second and third years, but the amount will lessen each year, until it is finally gone. Playing time seems to hasten the revelation process, much as it does the development toward real potential. I don’t have anything more empirical about this – it’s fairly sensible.


What About Boom Players?

We’re not done. There’s one more point here. I mentioned earlier that players can sometimes have a negative masking. If so, then the effect is something in the opposite direction – making players with some real potential appear to have less, or none at all.

(I will confess the numbers here are less well-developed, but I think the concept still applies well enough to use with some confidence. It’s less precise, as will become clear shortly.

What happens to a player who has some real potential, but the game’s masking process applies a number like “negative 50%” rather than the positive cases like we saw earlier? Well, let’s look at a defensive lineman this time, here are his actual unmasked ratings:

Run defense: 40/60
Pass rush technique: 40/70
Pass rush strength: 70/70
Play diagnosis: 60/80
Punishing hitter: 50/50
Endurance: 50/60

This should be a very solid player, a real contributor. Perhaps not a superstar, but at DE this range of skills would make for a very effective player – maybe a first or second round pick.

However, what if he has a –50% mask applied? If it works exactly the same as the positive masks (and we know that the green rating doesn’t go below the red) then we see a player who looks like this:

Run defense: 40/40
Pass rush technique: 40/40
Pass rush strength: 70/70
Play diagnosis: 60/70
Punishing hitter: 50/50
Endurance: 50/50


Anybody recognize this guy from FOF 2001? He’s (basically) what we would have called a “redliner” in that game – his reds ratings are substantial, but he shows little or no potential for growth (no green ratings – or here just a sliver in one category). In FOF 2001, a player like this would have a pretty high likelihood of breaking out and exceeding his apparent potential.

Same in FOF4 – this is a player whose real potential is being “masked” in the negative. He’s better than he looks.


This is not nearly as precise as identifying the draft busts. We don’t have everything visible, and so we don’t have any way of really calculating their actual potential. But there’s a certain pattern we can look for, in trying to find breakout players:

-Key ratings (formerly static) must not have any green at all, no exceptions
-Key ratings should (preferably) be pretty substantial – indicating even more possible upside
-Other ratings should have fairly little green visible

It’s true that there are going to be players who just happen to fit this mold without having had a severe negative mask applied. It happens. But, the more your player fits the picture above, the more likely it is that he will exceed expectations.


So, where does all this lead us?

Basically, what you read here is what you really need to know about drafting in FOF4. For maximum results, commit to memory the list of formerly static ratings – there’s at least one for every position group. When scouting rookie players, look at those key ratings first. Don’t be lured away by all the pretty green bars, look at the key ratings, and search for guys with all red, preferably fairly good ratings, but all red.

Every time you select a guy with some green at a key rating, you are accepting a guy who will not live up to expectations. Maybe that’s okay – if you calculate that his projected 90 will turn into an 80, maybe he’s still the best guy out there. But your most successful drafting strategy is to simply lay off just about any player who shows even a small hint of green in the key (formerly static) ratings.

(Also be aware that you can get stung by rounding errors here. If a player has an actual rating of 98 in a key rating, he might have substantial masking (like 30-40%), but it might not show up as a full point of green. Be careful in assuming that these players are completely safe.)

If you stick with drafting players who fit the profiles describe herein, you’ll minimize your draft day disappointments, and you’ll find yourself with more than your fair share of breakout players from the middle and late rounds. Your biggest problem, then, will become finding new house rules to keep the game challenging, since you will now be an FOF4 master drafter.

[Masked]

Wow, fantastic analysis.


I think there is a second class of breakout and boom players which is almost completely random. I have run the same draft (even just the last pick of the seventh round) several times to explore these players. A small set of players (1-5 per draft) will see a huge increase in ratings immediately after the draft. Prior to the end of the draft, they can look like anything (green in normally static ratings etc.) and afterwards they still can have any pattern. I have not rigorously tracked the same player after the draft when he did or did not breakout. In these cases, I think the mask stays the same (the 20 or 50% mask illustrated above), but the underlying, true ratings change. As a result, both the red and green ratings change. At this point the player will follow the development profile outlined above by QuikSand.

Most of my observations are based on a single case study, a 6th round scrub DE I selected who immediately after the draft looked like he should have been a top 5 pick. Since then he is a frequent all-pro. This draft in particular, I re-ran several times, and he never again broke out. As I mentioned before, I never simmed a season to see what he did in the majority of cases in which he didn't change after the draft. Interestingly, in one rerun, another late round pick exploded. He never played near these ratings when I used the original draft.

This seemingly random effect is very rare (0-5 players per draft), that it has little effect. It might explain some failures of the static ratings test that I have also found very reliable.

[Pyser]

i think one thing that should be "cross-referenced" here is the players popularity. i have found, very consistently, that the better the fans like someone, the better the player turns out to be. i'm not saying "fans need scorecard" players wont pan out, but i am saying that i've found it rare that an "idolized" player will bust big.

quick, did any of your reseacrh support this? im nearly positive that there is some correlation somewhere in all this.

[albionmoonlight]

This is one of only two things that I have ever printed from FOFC. (The other being that excellent burbon analysis).

[QuikSand]

Pyser, I recall you (or someone) making this claim before... but I have not explored it at all. There's some sense to it (as there seems to be a generally positive correlation between overal player skill and popularity), I just have not examined it to say there's anything more to it than that.

[Pyser]

i was just wondering if youd saved any of your case studies. then it would be an easy look to see whether or not it holds. in any event, i do believe it to be true, whatever that is worth.

[QuikSand]

Just a minor addition - if you're scouting rookies according to this strategy, I generally recommend sorting them by their current ratings, rather than their projected ratings. This tends to weed out most of the fluff in there - the guys with silly green ratings that will never materialze. Especially in the middle and later rounds, sorting by current ratings will help identify guys who are worth your attention much more easily.

[3ric]

After some study, it finally became clear that the masking system was not haphazard, and not random. No, indeed, the masking system in FOF4 is fairly formulaic perhaps it is so in order to save space required to keep all this (hidden) information compressed into a single item in the player file. (Those people who were working on draft files and so forth found a number of data points that were unexplained. I’m guessing that this is among them, in some fashion)

With Quik's findings fresh in mind, I began to compare some of these unexplained values in the draft files. One of these has a percentage-wise 0-100 range, which would suit the "masking" rating pretty well.

But I also found something interesting while bringing up some famous busts and booms from my "Historical project" dynasty thread. One of the more recent examples was the Montana/Simms controversy in the 1979 draft. I traded up to 4th overall to pick Montana, only to have him go to the Ravens with the 3rd pick.

Two seasons later, I'm looking as the winner as I took Simms, who's still looking good, and the Ravens has a high-priced career backup, as Joe's rating has gone down.

The comparison:

												Pos	%	Pro-rating?
Drafted 3rd overall, busted to be an average player				Montana	Joe	QB	94	6
Drafted 4th overall, still has potential to be All-Pro				Simms	Phil	QB	60	0
Drafted in the 5th round, former defensive player of the year.			Ham	Jack	LB	82	0
Drafted late in 1st round, busted big-time.					Payton	Walter	RB	48	11
Was undrafted, currently in his 10th year in league, starting for Denver.	Profit	Joe	RB	37	0
Undrafted, still in the league as an "almost-starter"				Lambert	Jack	LB	97	5
Undrafted and unsigned.								Webster	Mike	C	74	5
Drafted in the 6th round, 2nd in the league in receiving in 1980.		Swann	Lynn	WR	36	0

The "pro-rating" value seem to work as both a "bust probability" and a measure of how severe a loss of ratings the player suffer when busting. I believe the range of that value is 0-16. The lower the rating the better, of course. Tony Mandarich would be the prime example of a player with an "11" or more.

[Blade6119]

hmmm....interesting...*nods as he is not an fof4 owner*

[Blade6119]

grrr....i almost had it!!!

[nilodor]

Ahh the quest for the wigfecta

[Blade6119]

i got it though....

[sachmo71]

I shouldn't have read this thread...house rules, here I come. Thanks for the work, Quick!

[Airhog]

here we go!

[JeeberD]

Nice work Quik...

[QuikSand]

So, in the end, all I've really done here is create a comfortable platform for clowns who think it's interesting to get their name on all the forums at once. Fabulous. Glad I could be of assistance.

[Bee]

See you contribute more than you realize Quik...

[RonnieDobbs]

Well, if anything, the wigfecters drew my attention to this excellent analysis Quik. Great job.

[condors]

quiksand i think you did a bangup job on this

[Franklinnoble]

Originally posted by 3ric

The "pro-rating" value seem to work as both a "bust probability" and a measure of how severe a loss of ratings the player suffer when busting. I believe the range of that value is 0-16. The lower the rating the better, of course. Tony Mandarich would be the prime example of a player with an "11" or more.

What's a "pro-rating"?

[CamEdwards]

wow. I might actually have to give FOF4 another try after all.

[heybrad]

blocked

[fantastic flying froggies]

I guess it's too late to pull a 'hey brad' by now, isn't it ?

[3ric]

Originally posted by Franklinnoble
What's a "pro-rating"?

It's a hidden rating we found in the TCY draft files. It shows how good a pro the player is likely to become.

[PilotMan]

Sweet Job, Great Work.

[thealmighty]

not

[thealmighty]

dola...

i'm not being rude, it is just that I saw PilotMan with 4 of the spots as if attempting a wigfecta and i thought i might intercept, since all the posts he made were very "wigfecta" like

[sachmo71]

What's a wigfecta?

[QuikSand]

For what it's worth, with the changes to the FOF4 game as patched to level 4.01c, most of the advice in this thread has been obsoleted. Thankfully.

[Alf]

Originally posted by 3ric
The "pro-rating" value seem to work as both a "bust probability" and a measure of how severe a loss of ratings the player suffer when busting. I believe the range of that value is 0-16. The lower the rating the better, of course. Tony Mandarich would be the prime example of a player with an "11" or more.

Walter Payton & Joe Montana busting... That's an insult.

Now with the latest patch out, I am intersted in seeing the qulaity (or non-quality) of TCY draft files players.

[Daimyo]

Originally posted by QuikSand
For what it's worth, with the changes to the FOF4 game as patched to level 4.01c, most of the advice in this thread has been obsoleted. Thankfully.

Very interesting thread. I missed out on FOF4 so going back to re-read some of the strategy I missed and was wondering how this might apply to FOF2k4 (not so much during the draft but for post-draft trades/FA's etc) when I came across this near-final post. Still couldn't let the thread pass without commenting on how impressed I am with it.

[QuikSand]

Like many others, I have had "mixed" results with my drafting in FOF 2004. I found a re-reading of this article (at the top of this thread) to be useful, all the while saying to myself "things haven't changed all that much." Hope others find the same thing.

[OldGiants]

When I get some time to really sit down with Excel and my TCY draft files, I'll crunch some numbers that definitely support what QS is saying about busts and booms. But for now anecdotes will have to do.

What I've noticed by keeping a year-end TCY file open while doing the draft, is that the busts are green or red players in college, while booms are yellows who made it through to the late rounds. [Word to those who don't have TCY--Yellow is a current rating of 80+, red is 60-79, green is 59 and down to blues who you don't really want on the team]

Now that I read this, I see that QS's masking logic explains it all. The yellows who last until the late rounds have had there potential masked off, while the busts are now mind-bogglingly obvious. In other words, if you're using TCY draft files and follow this thread, its all there in front of you.

Never take a red/green player early.
Pounce on a late round yellow.

Too simple, really.

Makes using a TCY draft file much less appealing, now, too.

Sorry for that folks.

[Franklinnoble]

Originally Posted by 3ric

It's a hidden rating we found in the TCY draft files. It shows how good a pro the player is likely to become.

Can this be viewed and modified? I assume it's still relevant to FOF2k4 draftees...

[nero]

Due to the nature of the formula this can only be applied to a rookie...

anyone find anything similar for vets?