Volatility happens. A lot.

[aston217]

At least that is what I am starting to think.

I know most of us, me included, never really consider a player's volatility as a significant risk. But I think it is playing a bigger role - over a player's career - than has been stated before here.

If I am misunderstanding the mechanisms at work, I would like for someone to point them out to me so I can be enlightened!

Else, I should not be able to look at one team and point out all these volatility hits during TC:

1) 11th year QB, not creeping up, +7/+7 95 vol.
2) 5th year RB, had held steady, -10/-10. 18 vol.
3) 3rd year WR, slight creeper (+3 rookie TC), still developing, -1/-5. Cut test still +1. 62 vol.
4) 3rd year TE, +4 rookie TC, +1 cut test, still developing, +0/-4. Takes a 14 point or so drop in his BPR bar. 88 vol.
5) 6th year FS, 0 rookie TC but had some upwards movement over his career, steady and fully developed for a couple years now, goes +9/+9. 92 vol. I think the hitter went up just 5, but play diagnosis underwent a fair transformation, IIRC. This one I am just slightly iffy on, I'd have to check his pre-TC bars more. I guess he could have unmasked all at once and was just destined to be a 50/50 player despite a no-movement, 18/35 start.

Maybe I can go more in-depth later on what I saw, but it just seems like subtle, but difference-making vol swings in either direction - as opposed to the very few huge rookie transformations we see yearly - are happening all the time.

Is it enough to make a difference in how I treat players with high volatility? Probably not, but it is just worth noting. Or asking, rather, because if these changes, which all involve huge shifts in bars, are something other than volatility at work, I'd like to know what it is.

[MIJB#19]

1) 11th year QB, not creeping up, +7/+7 95 vol.Sounds like a volatilty change.

2) 5th year RB, had held steady, -10/-10. 18 vol.Could be volatility, but might as well be a random aging hit. RBs go downhill early and fast.

3) 3rd year WR, slight creeper (+3 rookie TC), still developing, -1/-5. Cut test still +1. 62 vol.This might as well be the opposite of a creeper and has been overrated/masked from the start.

4) 3rd year TE, +4 rookie TC, +1 cut test, still developing, +0/-4. Takes a 14 point or so drop in his BPR bar. 88 vol.Could be another case of unmasking/overrated, but who knows what it could be? Was he injured?

5) 6th year FS, 0 rookie TC but had some upwards movement over his career, steady and fully developed for a couple years now, goes +9/+9. 92 vol. I think the hitter went up just 5, but play diagnosis underwent a fair transformation, IIRC. This one I am just slightly iffy on, I'd have to check his pre-TC bars more. I guess he could have unmasked all at once and was just destined to be a 50/50 player despite a no-movement, 18/35 start.Could be anything from unmasking/underrated and a true volatility gain.


All in all, there are so many reasons why players can move in ratings, but every now or then volatility hits, sometimes in small chunks, rarely in big chunks and every once in a while one team gets an additional share of Yahtzee! rolls. It happens. :)

[aston217]

The TE was actually on IR the previous year, from a 'bad cold', though.

Appreciate the response! I am curious about the 'opposite of a creeper' - players who look like creepers, but are going the opposite way, despite a cut test indicating otherwise?

[MIJB#19]

With opposite of a creeper I just mean that it's actually not a creeper but just your average overrated player that somehow had a lucky break in an early camp and then started getting unmasked.

What do you mean by cuttest?

[aston217]

I mean when you cut a player and his ratings go up. (usually) a good sign for creeper types. In this case, the WR goes 18/22->18/23. He also went +1 in his first offseason (another thing creepers do) and +2 in his second offseason.

There's another RB I had traded away who was 6th year, definitely not creeping, and had fallen in recent camps. He went +4/+4 with 15+ points inexplicably gained in Hole Recog (among maybe other things).

[thenewchuckd]

Let us try again:

#1: vol
#2: probably vol. maybe an age hit. vol can happen to anyone, though, it should just be rare for a guy with 18 vol. I saw a guy in the USFL with 15 vol get like a -45 vsod in his rookie season.
#3: was he getting close to 100% developed? If so, it could be the usual creeper hesitation. Otherwise, vol.
#4: vol, definitely. The BPR drop is the dead giveaway.
#5: almost 100% certain that is vol. This one I have a very small amount of doubt with, because he was already a creeper, but as I am trying to explain to Ben E Lou... You don't see this much gain on creep alone when a guy is already developed.

Yes, vol happens more often than people think.

Over in the USFL we had to run training camp twice because of a mistake. On the first run, one team had two vol bumps on older players. On the second run, a completely different player on the same team had a vol bump. That does not sound like a rare thing to me.

How do I know it was vol? Because the behavior was not the same on the two runs and the guys were already 100% developed. Creeping is not random, if a guy is creeping you can pretty much know his development pattern. The only exception is what I call development creepers but even them, once a guy is fully developed you are not going to see them move too much in a short time period.

[thenewchuckd]

Yeah, so I decided to test this more because of a disagreement I had with Ben. Ben is usually right and I hate to disagree with him but I was pretty sure of my opinion, too, so something had to give.

This goes back to an apparent unlock of FOF 2k4 that showed creepers can have big reveals in their true rating, even after they are 100% developed. It was just random if your guy, fully developed, unmasked +18 or not.

However, it is clear that a lot has changed since FOF 2k4. Creeper development seems to have changed with the times, too. I have been convinced of this with the testing that I did recently.

So let me run through the myths and facts of volatility in the modern version of FOF and how it differs from unmasking.

[thenewchuckd]

Myth #1: You have to give your creeper playing time for him to unmask.

This can be true or false. What people may not realize is there are actually different types of creeping now (at least two kinds from my research). And these types of creeping can actually work in different directions. One type, that I call development creep, clearly needs playing time for a player to get better and unmask (the opposite of this seems to be when players hit 100% developed and their rating falls, sometimes dramatically, and never comes back).

The other type, what I call time stamp creep, does not necessarily need playing time. Actually, some preliminary research I have done shows that playing time can sometimes keep this kind of creep masked longer. You can see this easily with major creepers who do not get drafted.

In my tests, there was one QB with a huge mask. Started 40 rated before TC, went +8 after TC and another +13 to 61 future after staff hiring. No playing time... In fact he was never drafted or signed until year 2. I created him using the draft creator - I created him as a perfect 100/100 QB and 0 vol so those are not vol gains. And anyhow, as I will talk about later, vol hits and bumps only happen at training camp.

[thenewchuckd]

Myth #2: Once a player is fully developed, he can still show significant umasking. It is just a matter of dice but if you get lucky, and if you play him a lot, he could unmask like +- 10-18 points.

I am pretty sure this is false. For sure, time stamp creepers follow a pattern. Once the player hits 100% he will drop in rating but gain all of it back (and sometimes more) just after the next staff hiring. But after that, the player holds pretty much constant in rating until he starts having age drops.

In my testing, I followed one QB who had a huge mask, again I created him perfect. He started 15/31, went to 20/42 in his 1st TC, eventually making it to 61 rated before his 5th training camp. I should also mention that the guy had 0 volatility. Anyhow, I ran training camp on him 100 times. He came out on the other side of TC rated 61 or 62. The only exception was one time when he dropped to 56/56 (yes, that was a vsod, believe it or not).

In another sim I followed a FB. I made him perfect again and again he was a huge creeper, 0 volatility. He started 21/27 and went to 26/35 at his first TC. By year 5 he was 52 rated. I ran him through his year 5 TC twenty times, he came out either 51 or 52 rated. I ran him through his year 6 TC 40 times, he either dropped (age related probably) or came out between 53-55 rated. No huge +18 jumps here.

Yes, there is some random in unmasking but it is not huge random. If a guy goes +1 on one run, he's not going to go +10 or more on another run, unless he gets a vsol. Although it is interesting to see that you can see vsod on a guy with 0 vol... More on that later.

[thenewchuckd]

Myth #3: Vsod and vsol are rare.

This is clearly false. It is true that the life or death drops that you see in a guy's first training camp are fairly rare. You will maybe see maybe two to four death drops per class and another two to four gains (although the gains seem to be less dramatic).

But... what maybe is less well known is that vsol and vsod can happen at any time during a player's career, not just in his rookie season. These gains/drops are less dramatic, usually in the plus or minus 8-10 range (although I did have a -30 drop in the CyFL this past season). They are common enough, sometimes you even see two on a team in a given training camp.

How do you know if it is vsod/vsol or unmasking? You can often see it in the static bars. You don't see plus or minus 4 or more in a static bar, in one training camp, just from unmasking. Although, I have noticed that vsod/vsol does not always affect all bars equally, so this may not be a standard approach. But for sure, if your RB gains +10 in breakaway speed, that was vol not unmasking.

[thenewchuckd]

Myth #4: If I draft a player with low volatility, I am safe from vsod.

False. Vsod/vsol can happen, even on a player with 0 volatility. I talked about the QB with 0 volatility above who had the vsod drop in year 5. I also have examples of players with volatility as low as 15 who literally died in his first training camp.

It should be more rare (maybe much more rare) for a player to have a volatility influence if he has lower volatility. But it does not guarantee anything.

[Subby]

Myth #4: If I draft a player with low volatility, I am safe from vsod.
Pretty sure no one actually thinks this.

[thenewchuckd]

Fact #1: time stamp creepers follow a clear development pattern. They will gain in future rating at their first camp and follow that with future gains at staff hiring until they are close to 100% developed. Sometimes, during this time, they will have gains at training camp but not always.

Once time stamp creepers get close to 100% developed, they will drop in rating. This is followed by a nice gain in rating, usually close to their pre-drop rating or higher, just after the next staff hiring period. After this, they will continue to gain at time stamps but not nearly as dramatically.

By the way, I think if you reverse this you will get the same behavior for time stamp droppers but I have not tested with them enough to know.

I am pretty sure time stamp creepers never fully unmask, especially the ones who have big gains in their first training camp (think +5 or more). This is especially true for the ones who start out highly developed, since most of the reveal on this type of player happens before a player hits 100% developed.

[thenewchuckd]

Pretty sure no one actually thinks this.

Maybe many experienced people know this but I know the guy who had his 15 vol player die was pretty shocked. :eek:

[thenewchuckd]

Fact #2: There is another type of creeping and it can work in the opposite direction of time stamp creeping.

I will call this second kind of creeping development creep. You see this when a guy catches up to his future rating before he is 100% developed. So, for example, a guy has yellow experience but is already rated 50/50. As he develops, he will gain in current and future rating.

The opposite of this seems to be when a guy never catches up to his future rating and then hits 100% developed. So he is like 38/50 and shows red development then drops to 39/39. Be careful not to mistake this with a time stamp creeper, because they also drop when they hit 100% developed. The difference is that the time stamp creeper gains it back after staff hiring, the dropper doesn't.

But like I said, time stamp and development creep can work in opposite directions. A player could be dropping at the time stamps but then also have a development creep up. It is interesting to watch.

[MIJB#19]

I like that you tested this, but without letting us know how you tested this, I'm having trouble in believing your claims. Nothing personal or anything.

How did you differentiate between aging drops and volatility drops? How did you know the difference between a development ('creep'?) and a random volatility gain in ratings?

[thenewchuckd]

I'm having trouble in believing your claims. Nothing personal or anything.


No problem. If you read, I sort of outlined what I did in the posts above.

I actually used some ideas from Ben, where he was trying to prove that some creepers never fully unmask. I used the program that allows you to create a draft. I was able to create a whole bunch of perfect players and from there find the ones with a very big mask, like the 15/31 QB who jumped to 42 future in his first camp.

I was then able to follow these players and do tests. As for age drops, are you saying that you think a year 5 QB can have an age drop? If so, it is kind of a red herring, it was an aside to what I was trying to say anyways. The only point there was that players with low volatility can get a vsod or vsol... But according to Subby, everyone knew that already. :)

[thenewchuckd]

Oh yeah, I forgot that age also is a factor in drops so an old year 5 QB might drop... But nothing like that in my test. The guy in particular was actually very young. Going into his 5th season, he was only 25!

I was thinking, maybe you meant my ancedotal evidence where I said you can see two players on a team, in one season, get hit with vol. Yeah, for drops there may sometimes be no way to tell if it was age related or volatility. But mainly my information about this comes from vsol increases.

Personally speaking, I had two of them two seasons ago in FFL. That TC, my LT bumped from 46 to 54 future and he was never a creeper. He is now down to 51 rated as his regular dropping continued (still only in his 4th season so no age drop yet). That same season, my dropper QB went from 47 future to 52.

I have seen a couple other multi-player teams clearly get two vol bumps in one season. I have seen other clear vol drops, too (clearly vol since the players were not old enough to be age drops). Most notably my creeper C in CyFL who crept from 86 to 93 then dropped to 63 in his 3rd season.

[thenewchuckd]

Myth #5: If a player drops current rating in his rookie training camp, he had a vsod hit.

I saw Ben put this out there so I wanted to debunk it. I don't have hard evidence here just experience watching players over the last year or so. Players can, on occasion, drop current ratings in their first training camp and it is just their regular development path.

From what I have seen, it seems to only happen on players who start very highly developed. Think punters and kickers. It seems to just be a case of unmasking happening faster than a player's development. Nothing more than that.

[aston217]

Myth #2: Once a player is fully developed, he can still show significant umasking. It is just a matter of dice but if you get lucky, and if you play him a lot, he could unmask like +- 10-18 points.

I am inclined to think this is untrue, I just don't have any examples handy. In particular, I'm thinking of players who reach full development early in their careers and gradually pop +2, +3, +4 in various FA stages and some TC stages over the rest of their careers.

And actually, with "development creep", there is your OG Whalen. I don't know when he reached full development, but it couldn't have been over 47, and he finishes the season 64.

Something like what happened to Foster (#6 in the OP) or Grier from the CyFL (creeper RB going +13 in year 5) - I'm still just not sure if that is vol or unmasking. It looks just like vol to me, but it would be prettty interesting if it were unmasking.


Myth #3: Vsod and vsol are rare.


To summarize this perhaps in a more palatable way, the "OD" and "OL" parts are rare. The small bumps are what we think are more common. How much more common in 100 vol vs 0 vol, I don't think we've quantified.

Based on that 0 vol FB or whatever you ran 100 TCs for, with one vol swing, perhaps it is a reasonably significant difference.

--

Agree with Subby that people knew #4 already. It's part of why most people don't look at vol. It's really rare and can still happen to guys with really low vol.


I actually used some ideas from Ben, where he was trying to prove that some creepers never fully unmask.


Wait a minute, I think you and I both agree that some creepers never fully unmask.


Myth #5: If a player drops current rating in his rookie training camp, he had a vsod hit.


This is actually the one Ben E Lou said was previous knowledge and one he debunked himself by suggesting that this was unmasking (outside of the obvious, big death sticks).

I have seen that it really only happens to players that start very highly developed. Not always P/K, I've seen it on a RB before -- I think it's just because these positions tend to be ones where players come out very highly developed out of the draft.

[MIJB#19]

No problem. If you read, I sort of outlined what I did in the posts above.

I actually used some ideas from Ben, where he was trying to prove that some creepers never fully unmask. I used the program that allows you to create a draft. I was able to create a whole bunch of perfect players and from there find the ones with a very big mask, like the 15/31 QB who jumped to 42 future in his first camp.

I was then able to follow these players and do tests. As for age drops, are you saying that you think a year 5 QB can have an age drop? If so, it is kind of a red herring, it was an aside to what I was trying to say anyways. The only point there was that players with low volatility can get a vsod or vsol... But according to Subby, everyone knew that already. :)Makes sense. Thanks for explaining!

[thenewchuckd]

Grier from the CyFL (creeper RB going +13 in year 5) - I'm still just not sure if that is vol or unmasking.



I do not know about Foster but to me it is clear Grier was vsol. The big increase in static bars is the key. In my tests, I watched some players unmask huge over the course of 5 seasons. But over 5 seasons, you are very lucky if you see a static bar go +10 much less +3, even with a guy who climbs 40-50 points in overall rating.

No, huge jumps in static bars at one training camp are vsol.

[thenewchuckd]

Wait a minute, I think you and I both agree that some creepers never fully unmask.


Yes, we do.

By the way, we had a discussion about myth #1. I am not sure if many time stamp creepers unmask without playing time but it definitely happens. And to be frank, I think many people have known this already. A couple of things I have talked to vets about before:

-if you're not sure about your late round pick, don't sign him and see what he does in training camp. Whether signed or not, they will do their usual creep or drop, they just don't gain any current.
-creepers who come out of the draft with a high % developed just do not seem to unmask, but play above their ratings. Think about your CB Dominguez, for example.

[Pyser]

I also have examples of players with volatility as low as 15 who literally died in his first training camp.

thats quite the volatility stick of death right there, for him to literally die.

[Sef0r]

thats quite the volatility stick of death right there, for him to literally die.

omg I was listening to some dubstep of adele while I was reading this (don't ask) and I cracked up laughing...lol

[Firefly]

I released a player last season who had been a creeper but then maxed much lower and I immediately gave up on him. Aston picked him up. We're now in FA1 and I checked on him after reading this, and sure enough, he won 6 points back.

You got me on this one, aston.

[thenewchuckd]

You got me on this one, aston.

It is funny, Aston talked to me about Sosinski when you released him. We both thought it was pretty clear that he was going to gain the value back and we wondered why you were releasing him.

The interesting part is: I didn't see that you released Sosinski. I guess Aston trusted me enough not to bid on him? I seriously considered it after he told me...

Anyhow, I have seen that development pattern a few times recently. I think it is a mixture of a bit of development drop combined with time stamp creep. Usually a time stamp creeper does not drop like -6 or -7 when he gets close to 100%... It is usually more in the range of -3. But I guess there could be a bit of random in there that occasionally pushes it higher. On the other hand, time stamp creepers usually gain all of their drop back plus more. Sosinski only gained most of it back, so I am pretty sure there was a bit of a development drop.

[Firefly]

He paid the man about 30m, so no, he didn't trust you not to bid on him at all