Hey guys. I was having a bit of a think about play-by-play data this morning and ended up writing a spreadsheet looking at all of the important (and some unimportant) starters in the big leagues, using run values attatched to said play-by-play data.
Basically, I take Ks, BBs, HBPs, GB%, FB%, LD%, IFFB% and then turn them into percentages of each plate appearance, convert that to runs using the values mentioned in USSM's guide to pitcher evaluation, compare to league average and then multiply by batters faced to get an 'overall value against league', OVAL (note: it should be ovala, but I forgot an A while copy and pasting).
I've also normallised some values like LD% and IFFB%, attempting to take luck out of the equation. This was done by shifting the appropriate value 75% closer to league average than it was. Not ideal, but it was a quick solution and I'm lazy.
Important things to remember:
- OVAL is a counting stat. The more guys you face, the more you are rewarded/penalised
- Data collection was done by hand. Errors are likely
- I couldn't get league average data on balls in play for starters/relievers split, so I ignored it and used MLB averages.
- It completely ignores defense and ballparks.
- Being at '0' is actually pretty good - league average is significantly better than, say, Joel Pineiro.
- I don't really know how to convert this to an ERA scale, and I can't be assed with the effort.
Here are some selected players, will post others on request, or you can just email me to get the whole spreadsheet.
Zambrano, Harang, Ho.
Comments and thoughts welcome.