HoF: WAAS may be possible after all

cheekimonk · Post by **cheekimonk** » Sat Jan 28, 2012 11:45 am

I said I couldn't calculate Wins Above All-Star (WAAS) - which I talked about in this post - because we don't have WAR stats back beyond 1995. However, using the OOTP profiles, pitchers actually do have WARs for their entire career. Batters still don't, but they do have wOBA and ZR. From wOBA I can calculate wRAA (weighted Runs Above Average - weighted only because it will be calculated from wOBAs...if you calculate it from standard OBAs you get RAA) and from ZRs I can calculate UZR (Ultimate Zone Rating). WAR just so happens to be calculated from wRAA (10 wRAA = 1 WAR) and UZR (10 UZR = 1 WAR...RAA and ZR are complements in that RAA is the runs generated by a player's offense and ZR is the runs prevented/allowed by a player's defense).

Bingo-Bango, we have WARs for each season in a hitters career and, therefore, we have WAAS. It's unlikely I'll finish it all before this year's voting is complete, but I'll definitely provide the numbers for current and prospective HoF members starting in 2004. Feel free to ignore the stat, of course, but I think Peter Keating's logic is pretty sound and it'll be interesting to see what it says about our HoF (I've already gone through all of the All-Star teams back to 1998 and learned some very interesting things just in that process). Maybe I'll post a separate thread on calculations for those who are interested and so my numbers are transparent.

jumpmancol · Post by **jumpmancol** » Sat Jan 28, 2012 12:37 pm

I'm very interested in this, if for no other reason than I am a stat junkie!

bschr682 · Post by **bschr682** » Sat Jan 28, 2012 6:05 pm

i second that

jcrmoon42 · Post by **jcrmoon42** » Sun Jan 29, 2012 9:21 pm

I think it will be interesting as well. My only concern is that the more arcane stats have not been well calculated by the game in the past. Nobody even knows exactly where OOTP gets stats like ZR and WAR. My trust in them is not high. However, it will still be interesting to see.

cheekimonk · Post by **cheekimonk** » Sun Jan 29, 2012 10:11 pm

jcrmoon42 wrote:I think it will be interesting as well. My only concern is that the more arcane stats have not been well calculated by the game in the past. Nobody even knows exactly where OOTP gets stats like ZR and WAR. My trust in them is not high. However, it will still be interesting to see.

Wait, how would we not know where OOTP gets ZR and WAR? You can find the formulas for those anywhere on the web (and some competing versions, but they're close enough for government work). The only thing missing is the subjective aspect of ZR (in RL, a scorer plots where a ball was hit that the fielder either made a play on or didn't), but I would argue that OOTP is more accurate on that count (you can see the zone a ball was hit to in any game log). It knows what zone balls were hit to, who made a play and who didn't, and from there it's just crunching numbers.

cheekimonk · Post by **cheekimonk** » Mon Jan 30, 2012 7:56 pm

Some very interesting numbers here...I think everyone will be surprised. I've even expanded my original dataset...

jumpmancol · Post by **jumpmancol** » Mon Jan 30, 2012 8:25 pm

Looking forward to it!

---
I am here: http://maps.google.com/maps?ll=39.919650,-105.038002

bschr682 · Post by **bschr682** » Tue Jan 31, 2012 1:52 pm

well hurry up already. we are waiting.

jcrmoon42 · Post by **jcrmoon42** » Tue Jan 31, 2012 9:22 pm

bateague wrote:
jcrmoon42 wrote:I think it will be interesting as well. My only concern is that the more arcane stats have not been well calculated by the game in the past. Nobody even knows exactly where OOTP gets stats like ZR and WAR. My trust in them is not high. However, it will still be interesting to see.
Wait, how would we not know where OOTP gets ZR and WAR? You can find the formulas for those anywhere on the web (and some competing versions, but they're close enough for government work). The only thing missing is the subjective aspect of ZR (in RL, a scorer plots where a ball was hit that the fielder either made a play on or didn't), but I would argue that OOTP is more accurate on that count (you can see the zone a ball was hit to in any game log). It knows what zone balls were hit to, who made a play and who didn't, and from there it's just crunching numbers.

On the big boards, there has been general consensus in the past that the accepted formulae for these stats are not necessarily used completely within the game. They are basically silent on the issue. As such, we are pretty much at Markus's mercy in these matters. I'm particularly dubious about ZR, actually. Not for any particular reason.

cheekimonk · Post by **cheekimonk** » Wed Feb 01, 2012 7:51 pm

I think ZR falls in the same category as Errors - in real life an official scorer (who's presumably watched hundreds of baseball games) makes the judgment on an error, and the same thing for whether a ball fell in one zone or another. That's critical because a ball in one zone could count against or for the player - depending on whether he made the play - while the next zone over is out of reach for any player at the player's position. In a sim, by definition, errors and made/not made plays are generated by probabilities and ratings...fine for errors, but have you ever looked at a ZR chart? I couldn't imagine coding that into a sim.

Still, it's much better than nothing. Ozzie Smith committed 281 errors in 19 seasons while Ryne Sandberg committed 109 in 16 seasons. Who was the better fielder? I can take a stab at that question because I happened to be able to see them both play, but how do I explain to my kids that Ozzie Smith was one of the greatest middle infielders of all time? How do I even know that?

cheekimonk · Post by **cheekimonk** » Wed Feb 01, 2012 8:12 pm

BTW, I hit a snag on the offensive component of WAR for position players in seasons before 1995 (we have WAR going back to then). I was nearing the end with one method until I discovered a hole...now I'm down another rabbit hole (basically went from aiming for wRAA - weighted Runs Above Average - using wOBA - weighted On-Base Average, but had to switch to now trying to mold historical VORPs). I love math!!

blake · Post by **blake** » Wed Feb 01, 2012 8:54 pm

I asked someone at Sabr about using computers to track the ball after it's hit to get definitive range stats. He said right now they're working on that. They're going to get computers to analyze the speed and location of the ball to pump out range numbers. Right now it is still just manual I guess.

cheekimonk · Post by **cheekimonk** » Wed Feb 01, 2012 10:21 pm

blake wrote:I asked someone at Sabr about using computers to track the ball after it's hit to get definitive range stats. He said right now they're working on that. They're going to get computers to analyze the speed and location of the ball to pump out range numbers. Right now it is still just manual I guess.

ESPN had an article about that during last season. Basically, since MLB has more leverage to tell franchises what to do they are going to use that to put sensors and cameras at specific places in every stadium that, with today's processing power in computers, can tell them everything about a ball at any time. MLB is doing it because some teams were planning on doing it and thinking about bringing in stat brokers (like Stats, Inc.) who would have control of the data in exchange for doing the work and MLB was not having that. But, they are keeping that quiet because they're afraid that the public will make the obvious connection between that and letting computers call balls, strikes, foul balls, etc.

The Brewster

The Brewster