A Test: Change position, or just play out of position?

[VarlosZ]

I recently started a thread (http://www.operationsports.com/fofc/showthread.php?t=51301) about how best to go about playing one of my LBs out of position in order to keep him on the field more often. There was some disagreement about whether it was better to switch his position and take a moderate hit to his current ratings, or simply to play him at his new position without changing his position.

To get some idea of the costs of each approach, I simmed a season 10 times in which I played my SLB and my WLB in each other's slots without changing their positions, while doing the same with my LT and RT on offense. Then I simmed 10 seasons under identical conditions, except this time I changed the positions of the players in question before proceeding. There's some chance that the two studies are poisoning each other to some extent, but the effect is probably negligible, and I haven't the patience to do them separately.

Anyway, this is how they did.


WLB Kris Beyer is an 88/92 veteran in his 5th year; SLB Antoine Raymond is 72/77, also in his 5th year. Breyer will start at SLB, while Raymond will start at WLB. Switching their postions at the start of the season reduces their ratings to 76/92 and 51/77, respectively. Obviously, they both get some of that back over the course of the season. These are their average numbers with and without changing their positions. Keep in mind that the WLB is always the first LB taken off the field in Nickel and Dime packages, while the SLB is virtually always on the field. This is the main reason for the huge gap in their respective raw numbers. Also note that Beyer gets many more blitzing opportunities than Raymond.


Beyer
________________Tckl_____Asst____Sack_____Hurr_____PRPct_____TkPct___
Don't Switch . . . . 72.8 . . . 24.0 . . . 12.85 . . . 15.7 . . . 5.25% . . . 11.35%
Switch Position . .. 71.7 . . . 23.9 . . . 11.9 . . . . 14.0 . . . 4.81% . . . 11.07%

Raymond
________________Tckl_____Asst____Sack___Hurr____PRPct_____TkPct___
Don't Switch . . . . 55.3 . . . 17.0 . . . 3.5 . . . 3.5 . . . 2.36% . . . 12.85%
Switch Position. . . 47.4 . . . 18.8 . . . 4.0 . . . 2.3 . . . 2.01% . . . 11.60%


Most notable is that it's clearly better in this case -- at least for this season -- not to switch their positions. Beyer's ratings were dropped just 13.6% by the switch, and he got most of that back by the end of the season, but that was enough to be more costly than the penalty for playing him out of position.

Not surprisingly, Raymond suffered more than Beyer after taking a larger ratings hit (he also regained his value at a slower rate than Beyer over the course of the season).

Finally, my defense allowed an average of 14.7 points per game without switching positions, and an average of 16.5 PPG with. Ok, on to the Offensive Tackles.


LT Randal Maze is entering his 3rd year with ratings of 68/82 -- switching him to RT takes his ratings all the way down to 39/82. RT Kwame Harris is entering his 10th year with ratings of 68/77 -- surprisingly, switching him to LT barely reduces his ratings, to only 66/71. Maze will start at RT and Harris will start at LT. The gameplan is set so that runs to the right and left will be equally frequent in all situations. These are their average numbers with and without changing their positions.


Maze
__________________KRB____KRO______BPct______SPct______OPct___
Don't Switch . . . . . 29.6 . . . 89.9 . . . 32.9% . . . 0.27% . . . 17.6%
Switch Position. . . . 20.6 . . . 65.5 . . . 31.5% . . . 1.22% . . . 12.6%

Harris
__________________KRB____KRO______BPct______SPct______OPct___
Don't Switch . . . . . 23.7 . . . 71.1 . . . 33.3% . . . 0.44% . . . 13.83%
Switch Position. . . . 28.5 . . . 82.3 . . . 34.6% . . . 0.38% . . . 16.39%

My offense averaged 26.1 PPG without a switch, and 25.7 PPG with.


First, in spite of my gameplan, rushes were not distributed evenly between left and right. Instead, my coach directed more rushes to whichever Tackle had a better rating.

Second, even in the case of Maze, who took a huge ratings hit, the run blocking effiency was affected only slightly. I would hypothesize that that has something to do with the players' Blocking Strength being unaffected by the position switch -- Maze's 98 strength rating probably stops his BPct from slipping too far.

On the other hand, Maze's pass protection skills completely collapsed, as he was 4.5 times more likely to allow a sack with a position switch than without.

Finally, note that Harris' rate statistics got better with a switch. Apparently, this is because he suffered such a small ratings hit (about 3%) that avoiding the penalty for being out of position more than made up for it. Perhaps tomorrow I'll run some tests to see if I can quantify just what the penalty is -- based on these numbers, I'd guess it's equivalent to roughly a 6% hit on current ratings.

[MrBigglesworth]

First, in spite of my gameplan, rushes were not distributed evenly between left and right. Instead, my coach directed more rushes to whichever Tackle had a better rating.
Do you have the directional running evidence from the game to back this up, or is this just based off of the number of KRB opportunities? Because I find that krb's frequently don't go to the players on the side of the field that you ran to on the play.

[VarlosZ]

Good catch. I didn't check run direction, just assumed that it would correlate with KRB opportunites. And, sure enough, the first season I sim to check on it, my LT has 46% more KRO's than my RT, even though I ran to the right slightly more often. Apparently, creating KRB opportunities is a skill in FOF; not entirely unrealistic, but certainly not how I expected it to work.

[fantastic flying froggies]

did you have injuries turned off for that study?

[VarlosZ]

Yes.

[Celeval]

I think the part that's missing in this - not a criticism, as this is good work - but the longer-term effects. I think it shows pretty well that playing cross-position gives slightly better numbers than the change; but what happens when the player catches back up? Is it a question of an 80/80 SLB switching to 70/80 at WLB, but playing at 75/75 when crossed? Will a position switch be more beneficial down the road?

[Warhammer]

I agree with Celeval, as I would expect the long term answer to be more along the lines of Harris' production than others, down the road.

[QuikSand]

This is a very useful contribution to the pretty slight knowledge base on this topic. Thanks.

[fantastic flying froggies]

I also think that having injuries turned off screws up the results slightly. It is my belief (key word there, no evidence to back it whatsoever) that players playing out of position have a higher risk of injury.
Of course, I don't know how one could test that.

Having said that, nice work. Like QS said, we know so damn little every little bit of info helps.

[VarlosZ]

Thanks for the kinds words.


Yes, it's hard to say at what point it becomes better in the long run to switch a player, and I think it depends on a number of factors. The most obvious one is how large a ratings hit the player takes.

Next is the player's experience -- after his 5th season or so, the penalty for switching a player's position gradually decreases, until you can do it for free (within the same position category, that is) after his 9th or 10th season (usually). So, in some cases, it will be better to hold off on a switch for a year or two. For example, this turned out to be the case with OLB Beyer; after I completed the season I used for the test, I switched his position prior to the next year's training camp. He went from 88/92 to 82/92 (instead of 76/92 the previous year), and by Week 1 of that season he was back up to 87/92.

Also, one might want to switch positions earlier than he otherwise would if he's about to renegotiate the player's contract -- the lower rating gets him a better deal. Of course, that seems pretty gamey, so I assume most players would eschew this strategy.


Later today, I'll run a test on the season again, but this time leave LT Maze and RT Harris at their natural postions. Comparing their production in this test to how they performed in each other's slots w/o a position switch should give some idea of just how much of a penalty there is for being out of position.

[VarlosZ]

Later today, I'll run a test on the season again, but this time leave LT Maze and RT Harris at their natural postions. Comparing their production in this test to how they performed in each other's slots w/o a position switch should give some idea of just how much of a penalty there is for being out of position.

Actually, it occurs to me that that might not be such a good test, since the players will face different competion than they did when they were on the other side of the line. In fact, a quick survey of the DEs opposite my Tackles shows that the LT had a harder assignment (arguably) 14 times out of 16.

So, scratch that idea. If anyone has an idea about how to test the penalty for being out of position without corrupting the data with strength of opposition inequity, I'm willing to do the leg-work.

[QuikSand]

One things this does shed *some* light on, to me, is the magnitude of the "playing out of position" penalty. We know it exists, but I don't think there has been another published effort that gives us much of any insight into this.

Look at LB Beyer, for example. If we take these ratings at face value... then what do we see? (This is a guy whose overall rating drops from 88/92 to 76/92 if he is permanently moved from WLB to SLB) Over the course of the year, he gains back "most" of that lost current skill -- so let's say that for the course of the season his median overall current rating is something like 80. It appears that he is worse off, in year one, making the switch (playing with skill of 80) than he is just playing out of position with nominal skill of 88. To me, this suggests that the magnitude of the "out of position penalty" is something less than 10%, if everything is essentially linear.

Not hard and firm evidence, but for those of us who are prone to playing guys out of position (and used to seeing good production that way) this seems like a bit of a buttress.

[Vinatieri for Prez]

Hey, as BB says, a football player is a football player; it doesn't matter where he plays.

[Celeval]

I think the location of the out-of-position may have something to do with it as well... for instance, a FS out of position as a LCB may have a different hit than a RCB playing LCB.

[QuikSand]

I think the location of the out-of-position may have something to do with it as well... for instance, a FS out of position as a LCB may have a different hit than a RCB playing LCB.

Agreed -- should have added that to my list of qualifications above. WLB to SLB is probably among the slightest transitions available in the game, I'd guess. The penalty may indeed rise from there, even substantially.

I'm just pleased that we have at least something to say about this.