|
Hi Beatles. Good to hear from you.
I'll know more next week. The batters are nearly done. The fielding is nearly done. The Master db needs work but that's easy enough for me since I correct info in the core OOTP db for accuracy and the conversion of IDs, debut and retirement years is only an hour's worth of work with the process I use. ... So that leaves the pitching. I've spent the past month reverse engineering and following Eric Chalek's steps for calculating the HR, BB, and K rates (WP and HBP too but those matter less). Eric puts a ton of work into the WAA and other value stats (and my use of those stats perfectly derive core traditional stats like R, ER, ERA, etc for the things I need for OOTP). But OOTP relies heavily on HR/BB and K rates and Eric stopped publishing those.
As the day comes to a close and I can turn from day job to hobby time, I think I am close to a model that will replicate (close enough) his process so I can derive those traditional stats through my own model. The iterations I ran last night were the first pushed through it all the way and the results were pretty darn good. Eric and I differ some on individual years but I reckon part of that is how we allocate the Winter Leagues, different guardrail settings, and the fact I think he blends the HR, BB and K's calculated in his year n with those in his year n+1 and year n-1. So his previously published HR, BB and K stat lines might be inherently already three-year recalc'ed and I elected not to do that. My only blending is when IP for a given year is less than 45 IP.
So far as test runs, I've only pushed Joe Williams, Slim Jones, Rats Henderson and Dick Redding through. After I compare some more results to Eric's last published versions of the traditional stats, I think I'll be confident I can replicate it for all the new players (a lot!) that he never published traditional stats for. I'll also press on and redo the original 212 to eliminate the pre-baked in blending.
When will it be done enough to release for folks to test in a public beta (as it were)? ... Tough to say. But I will know more next week once I test the pitching model some more.
|