Player weights & position switches (INPUT NEEDED)

[gstelmack]

I don't have froggie's data yet, but I wanted to confirm some of the assumptions I gave above, so I looked at all my players that can convert to RB. For those following along, I've updated the spreadsheet at http://home.nc.rr.com/gstelmack/PositionWeights.xls with my latest data and an attempt to predict RBs.

What I've found is that by using the following steps, I can predict what the game will claim the final attributes are for a player switching to RB to within a point (at least for all my test cases):

• If the player is more than (or equal to?) 265 pounds, no switch.
• If the player is a C, G, T, P, K, DE, DT, ILB, or OLB, no switch.
• Start with 100. NOTE: Spreadsheet uses 88, but that was just helping me normalize the values the game gave.
• If the player is a rookie, add 10.
• If the player is a:

• QB, subtract 37
• FB, subtract 32
• TE, subtract 27
• WR, subtract 25
• CB, subtract 12
• S, subtract 27

• If the player is under 215 pounds, subtract ((214-weight)/2)
• If the player is over 219 pounds, subtract ((weight-220)/2)

With that formula, I'm within +/- 1 on all my players, generally +1 close to the 215-219 ideal weight range, and -1 farther out. This means that the weight adjustment is not linear, but I don't think 1 point will change someone's mind about switching a player (and more importantly would have little bearing in a draft situation), so this works well enough for my desired uses. I welcome attempts by any of the math people out there to make this more exact.

Once I figure out the formulae for the remaining positions, I can update Draft Analyzer to give a reasonably solid guess about how well a player will switch instead of a simple yes/no.

The big surprise in all this for me was how poor FBs will be when switched to RB. I almost wonder if this is a bug. I'd expect FBs to generally make reasonable RBs, but an ideal-weight 217-lb rookie FB in my test case would only retain 78% of his attributes when switched to RB. I've actually signed on FA FB in the past who wasn't much of a blocker but had some running skills, thinking he'd make a reasonable RB. I was wrong, and it looks like that will never work.

[fantastic flying froggies]

Greg, I am finally done and e-mailed you the excel spreadsheet with my 3 teams. Let me know if there are any problems with it.

[azjoe_02]

I can't wait till this is finished. I appreciate all the work you guys are doing on it.

[gstelmack]

I'm running into some issues with figuring out the ILB position. Some of the source doesn't make sense unless I assume there are hidden attributes that apply. Or, it's being complicated by the fact that SLBs and WLBs are being treated differently. The analysis is getting complex...

[gstelmack]

Spreadsheet updated again. It has all of froggie's players (which ended up validating the RB formulae), and I'm close on the ILB data (within +/- 2 for non-ILBs switching to MLB). If anyone wants to take a stab at any of the other positions, please feel free. This is going slow and will take a while.

[QuikSand]

Bump with a few very minor updates on this topic.


I have been fiddling with a small data set so far, looking at linebackers who can move to play defensive end. So far, most of the results are pretty mundane:

* The heavier the better, thus far -- I'm sure there is an ideal weight for th DE position, and it must be at least in the mid 270s, since all the LBs I have switched so far have been below that, and they universally get a higher conversion % from my scout with a higher weight. Fits with gstelmack's findings above -- there does seem to be a "sweet spot" and in this case, it does seem to be fairly tough to locate linebackers that heavy

* Outside linebackers are prone to a better switch % than are inside linebackers of equal weight -- the difference seems to be about 3% worth, which is not inconsiderable, especially when the range of these switches is generally around 88-95%

* So far, I don't see any difference between switching a WLB and a SLB - also consistent with gstelmack's earlier findings, that this is managed by "group"



However -- I am finding a few things that are a bit more interesting. All preliminary, of course - but a teaser or two here:

* I'm finding very little pattern between how well the scout says the switch is going to go, and how well the player actually translates to the new position. With a small set (and scout error that is admittedly meaningful with low caliber players) I am seeing almost no correlation between the success of the switch for my players at 92-95% compared to those at 88-90%.

* I am finding that the position switch can, at leat sometimes, yield a sort of "quirky" result in one specific category. I have two players who have seen their endurance ratings change substantialy in this move from LB to DE -- one guy jumped by about 20 point in one trial (not made no meaningful change in others), another guy had about a 20/35 point drop to a complete void in one trial (but was unharmed in others). I have yet to see this in another category, though it may just be coincidence that this has happened both times in endurance.


A few open thoughts about this:

- We know that in FOF 2004, there are ratings "voids" for certain players -- guy who look generally fine in their respective ratings, but then have an odd zero for something or another. It's an interesting quirk in the game, I think (I kinda wish there were more of these players). But we really don't understand this well at all -- the fact that one of my players seemed to "develop" a void in endurance (not droppping to 3/6 or something, but dropping to an absolute 0/0 rating) makes me think that perhaps this is something special -- maybe the mechanism for the voided ratings is peculiar in some way? Maybe it's strictly a function of the scout? Maybe the player really does have skills/ratings there, and we ajust aren't seeing them? Beats me... just thinking openly.

- In another career, I switched a very solid RB to play WR mid-career, and while he made the switch fairly painlessly and became a productive WR, I was surprised to see that he miraculously developed excellent return skills (both PR and KR) after the position switch -- they, too, just sprang from nowhere. As I recall, he had zero ratings in both PR and KR afterward, and after the position switch, he suddenly had high ratings in both. Maybe a position switch is a possible "cure" for a ratings void? Maybe even a harmless position switch -- like bumping a rookie SLB to WLB or the like might turn that zero into a decent rating, perhaps making a deeply flawed player into a valuable one in the process?


Haven't done much testing yet... I'm better with the "forming hypotheses" segment of the scientific method anyway.

[jamesUMD]

It's funny that you mention this topic. I read through it and I was checking a few of my guys just to see who could switch to what. It's usually 3-4 spots that a player can switch to. I have a underacheiving, and under progressing safety that I took with the 19th pick of the draft in the wigfl. For the record he is 6' 1", 226 lbs. I checked, and I have never seen this before. He can switch to the following positions:
RB, FB, FL, SE, SLB, SILB, MLB, WILB, WLB, or FS.

It's Jim Jensen reincarnated (I'm probably dating myself). Has anyone ever seen so many options for 1 player?

[QuikSand]

...every DB with that weight.

[QuikSand]

One continuing theory about position switches. Unsupported, of course.

I'm guessing that one thing that happens when a player's position is switched is that the scout gets a "new look" at him... along the lines of when a rookie player joins your team for the first time, you get a slightly different look. I'm guessing that, in essence, the dice-rolling for scout error is re-done at this point, and this is what explains most of the uncertainty in the switch process.

My best guess is that when your scout says the player is going to retain 90% of his ratings (for instance) in the switch, that in reality that's what ends up happening, however it's actually calculated. However, I think what's also happening is that one set of scout error is stripped out, and replaced with another (potentially in a new category, with a different magnitude of precision) -- and so we are sometimes seeing guys who are (to continue the example) holding the 90% opf their ratings, but they seem to be changing in potential from, say, 54 to 38. What's really happening in these cases, perhaps, is that the guy is actually dropping from 50 to 45, and really is keeping 90% of his previous rating, but the scout error on both ends of the transaction makes it look like a bigger swing than it really is.

Just idle speculation... I think the next game file I open up to tinker with this will include an outstanding scout, at least in the areas I'm focusing on.

[Crim]

Pardon my rookieness, and not to threadjack (since I think this lies just outside the scope of what's being studied in this thread), but with seemingly routine changes like RDT<>LDT, LG<>RG, etc., are there usually any sort of ratings changes for these players?

And thanks for all the above info! A big help for lazy/stupid people like me!


Crim

[Plundun]

Originally Posted by QuikSand

- In another career, I switched a very solid RB to play WR mid-career, and while he made the switch fairly painlessly and became a productive WR, I was surprised to see that he miraculously developed excellent return skills (both PR and KR) after the position switch -- they, too, just sprang from nowhere. As I recall, he had zero ratings in both PR and KR afterward, and after the position switch, he suddenly had high ratings in both. Maybe a position switch is a possible "cure" for a ratings void? Maybe even a harmless position switch -- like bumping a rookie SLB to WLB or the like might turn that zero into a decent rating, perhaps making a deeply flawed player into a valuable one in the process?

I think that example is position and skill-specific. The same 4 positions have punt and kick return skills in TCY and FOF2004. But whereas in TCY I have had PRs and KRs from every one of these position it always seem to be WRs and CBs who hold up their good return skills in FOF2004. I think RBs and Safeties might get some sort of penalty to these skills that perhaps disappeared when you made the switch.
I play mostly with ported players from TCY and the low KR/PR ranking of RBs and safeties might of course be one of the quirks that still exist (TEs always 0 in Big Play Receiving, UCLA and Southern Cal switched around)
In other words I don't think you can get dramatic bumps changing a WLB to a SLB. Generally however it almost always pays off changing a MLB to a OLB or reverse. As long as he has the physical stats to make the switch you will always get the message that he will be the same or slightly better and it usually follows through.
In fact I think there is a correlation between gains from the position switch and gains during training camp. I only switch rookies and always before the training camp and I believe that the bigger the gain when changing the bigger the gain in training camp.
NB This last example is only when changing MLB and OLB.

[gstelmack]

Originally Posted by Crim

Pardon my rookieness, and not to threadjack (since I think this lies just outside the scope of what's being studied in this thread), but with seemingly routine changes like RDT<>LDT, LG<>RG, etc., are there usually any sort of ratings changes for these players?

In most cases, the answer is no, although there can sometimes be a minor little hit. However, note that draftees (especially from TCY) are not always at the right weight for their position. For example, I drafted an SILB who is now playing SS for me because he was below weight at ILB. He could not switch to MLB or WILB.

[mcsestretch]

Originally Posted by gstelmack

In most cases, the answer is no, although there can sometimes be a minor little hit. However, note that draftees (especially from TCY) are not always at the right weight for their position. For example, I drafted an SILB who is now playing SS for me because he was below weight at ILB. He could not switch to MLB or WILB.

I wish I had seen this thread before I posted a new one...

I am trying to find what ratings indicate how well a player will translate from one position to another. Obviously weight determines what position a player may change to but I'm wondering if it also determines how well he will play the new position.

I've tried to find a direct correlation with loyalty/leadership/intelligence/personality/volatility but I can't find a direct correlation with any of them.

After looking at the data for a couple of days I was ready to start beating my head against a wall.

Anybody have any ideas?

[st.cronin]

Originally Posted by mcsestretch

I wish I had seen this thread before I posted a new one...

I am trying to find what ratings indicate how well a player will translate from one position to another. Obviously weight determines what position a player may change to but I'm wondering if it also determines how well he will play the new position.

I've tried to find a direct correlation with loyalty/leadership/intelligence/personality/volatility but I can't find a direct correlation with any of them.

After looking at the data for a couple of days I was ready to start beating my head against a wall.

Anybody have any ideas?

I think I remember reading that the closer a player is to the 'ideal' weight for the position he is switching to, the better the switch. For example, say the ideal weight for RB is 220 (I have no idea what it *actually* is). Two FBs of equal ability, one weighing 240, one weighing 230. The 230 lber would be the better candidate.

[jbmagic]

is it better not to switch the player position?

Put the player where you want them to play in the depth chart.

and you dont get any reduce penalty by doing that like you do if you change a player position right?

[gstelmack]

Originally Posted by mcsestretch

I am trying to find what ratings indicate how well a player will translate from one position to another. Obviously weight determines what position a player may change to but I'm wondering if it also determines how well he will play the new position.

Player and position switching from appear to be the ONLY factors. The spreadsheet I posted earlier backs this up. There appears to be a random tweak (if it says 98%, you might end up with 95-100% or something similar), although Quik is trying to prove that that may just be differences in how the scout views the two positions.

As for playing out of position vs. switching, there was a separate investigation going on. For small roster holes due to injury, I tend to play out of position if possible, but you can also draft guys who will actually play BETTER at a new position than their old. For longer-term solutions, I directly switch. But I also don't switch unless the guy will be 90% or better without a darn good reason.

[QuikSand]

I ram a few tests incolving running backs converting to wide recever -- it's a pretty frequent place to try a transition, I think -- i frequently see RBs who go undrafted, and have spotty skills that probably mean they won't be successful as RBs, but who have pretty good receving skills, and are worth a look there. In fact, in the random career I used to do my study, I found one undrafted RB who turned into a pretty solid-looking rookie WR -- a guy I'd probably be happy to have grapped in about the third round of the draft.

So...a couple very preliminary observations about shifting fairly untalented RBs to become flankers:

- There are several "core" receiving skills that you can see for a RB -- for the most part, your scout's assessment of how well the player will translate seems to apply pretty well to these ratings, on average. If your scout says a given RB will retain 90% of his ratings, then you can expect things like "route running" and "avoid drops" to drop by around 10% -- in averaging the loss in potential in these ratings, the numbers have a vague correlation to the predicted skill loss. Not perfect, but in most cases probably with in the range that could be attributable to scouting error.

- The ratings that are not visible for a RB but are important for a WR include Big Play Receiving (a single value rating, never any difference between current and future values for any player), Courage, and Adjust to Ball. As nearly as I can tell so far, these rating seem to be treated differently in the game during these switches. It looks to me like these ratings are some combination initially pre-determined, but with some occasional (random?) element of change included in the transition. Most times, a giveb player will escape from the position switch with a certain new rating in these areas -- but occasionally there will be a completely different result. I have seen one player, on consecutive trials, convert to a WR with a "Courage" rating of 2/8, and then with a 43/71 -- this is clearly well beyond scouting error, it's something altogether different. (I think there might be less flexibility in the outcome of the BPR rating -- but don't have enough trials to say conclusively)

- I looked for surface connections between the eventual results in these new attributes and anything else that might be used as a predictor -- and haven't found anything. The combine 40-yard-dash time, for instance, doesn't seem to connect at all to the player's eventual ability in Big Play Receiving, for example -- if anything, it seems like the BPR rating is connected just to the player's overall abilities, i.e the better the player is overall, the better the generated BPR score will be.

- There is definitely something weird going on with return skills, which are visible for both RB and WR positions. I'm seeing running backs with zero or minimal return skills turn into receivers with at least passable return skills (say 20/30) on a regular basis - there's a lot of change here, and nearly all of it for the better (though I confess I'm not using a lot of RBs who actually have great return skills to begin with).


Nothing conclusive... but just thought I'd share some impressions.