THE BOOK--Playing The Percentages In Baseball

I like it.  Seems to make sense.

I like it, too. Tango, could you give a couple sentences on how the C value is determined.

David, I’m not sure I’m really on firm ground with the catcher.  It’s alot of gut feel.

Also, once you read my article and someone else’s article on catchers in the THT09 annual, it’ll probably give you enough pause to make you believe that we need to have special considerations for the catcher (unless you already believed it to begin with).

One study I’d like to see done is how “third string catcher” perform, relative to actual catchers. There are a few Phil Nevin sorts - corner OF/1B who can play catcher in a pinch. And sometimes you find a Phil Nevin type whether or not you have one, based upon the needs of the team. That would give us a frame of reference for how catchers perform relative to everyone else on the defensive spectrum.

That is my article in the THT09 Annual…

Fantastic! Looking forward to it.

I remember at some point seeing a version of your original chart where the IF positions got a bonus over the OF positions. Have you abandoned that model?

Also, am I correct in saying that this model is based almost entirely on DER/ZR type “plays per BIP” fielding evaluations? I’m starting to think that if you add the IF/OF gap back in and give RF/2B a bonus for their “non-catch” skills over their LF/3B counterparts (arm for RF, DP pivot for 2B) this would look a lot more like the traditional Jamesian defensive spectrum, and would line up better with the known offensive disparity between positions.

The IF over OF bonus would be in the handedness and nothing else.  Therefore, I think this updated model addresses that:
1.25 C
0.75 SS
0.25 CF/2B/3B
-.75 LF/RF
-1.25 1B

You’ll have to point me to where I may have said or implied something otherwise.

It is based on the fielding value based on UZR.  It is also linked to the Fans Scouting Report as a sanity check.

The bonus for RF/LF for their UZR arms (which was not included) would be slight, one or two runs.  Seeing that everytime I run this (depending on the sample data), sometimes the LF comes out ahead with UZR range+error runs, and sometimes RF comes out ahead, I’m not going to be a stickler for 1 or 2 runs, since that’s around the uncertainty level.

As for the DP pivot, if someone wants to offer their translations on this, then please do so. It is possible that if you include the pivot value, that this may swing 3 or 4 runs, enough to make it that the 2B is closer to SS than to 3B.

However, if this is based strictly on the familiarity factor, then I am less inclined to do so.

The only position in which the above spectrum doesn’t correspond to the offensive spectrum is 2B.  If I set 2B to .50, it would match the offensive spectrum.  Then again, 2B are paid far less than 3B, and even if you adjust for the years of service, I don’t think you’ll find that the 2B = 3B in any way shape or form these last 5 years.

For LF/RF, in some years, the better fielders were in LF, not RF.  These days, there are better fielders in RF than LF.  I think there’s been a shift in balls in play toward RF (lots more opposite field hits) in these last 5 or 10 years, which is probably why better fielders are now there.  But if you go back to the 60s, 70s, 80s, it was probably fairly balanced, and maybe even better fielders in LF.

So, until that is resolved, it won’t make much sense to say -.70 RF, -.80 LF.  At that point, I’m implying a precision level that I simply can’t support.  Others like to give out differences to .01, and that’s not me.  I’m giving differences here at the .50 win level.  I don’t want to make it anything less than .25 wins of uncertainty.  You can reasonably argue that it should be -.65 RF, -.85 LF.  But, do you really want to go there?

Maybe we can set it to LF -8 or 8.5 and RF +7 or 6.5

Then for 3B vs 2B on the DP pivot:  I’ve got ratings for 50+ years of infield double play runs, look for my 2nd TotalZone article on Hardball Times for the download.

I was thinking you could take all player who played both 2b and 3b, who were primarily 3rd basemen (more career games at 3rd to determine this?) and see what their combined DP rating is at 2nd per season.

Yup, if you have that, that’d be nice to see.  I’d like to see the DP runs for primarily 3B, primarily 2B, and neutral. 

Also, I have been finding that if you break down the data by over/under 30 years of age, you can handle the aging issue better.  No need to pollute the data with a 34 yr old, if it biases the data one way or the other.

I thought of this when people were saying Hawk was primarily a RF, and Rally (I think) pointing out Hawk and Dale Murphy had the same number of innings in CF.  That Hawk was able to outlast Murphy as a hitter to stay in RF doesn’t now make Hawk a RF!

If you take the position’s innings through age 28 or 30 or something reasonable, I would then make THAT the player’s primary position for research purposes.

(If you make it through age 28, you will get 50% of the player’s career.)