Hey Guys. I am in the process of trying to put in my due dilligence and hopefully create some type of MLB system over the next year or so. I'm sick of trying to cap games, because the grind is just to difficult on me. I would like to put in my work upfront, and allow a database to tell me the play. Sounds good in theory, but obviously have alot of legwork to do.
My question is what type of data other then scores, openers/closers and starting pitchers will be beneficial to help me along the way.
This is like trying to build a spaceship in your backyard
Start with building something smaller. Try to model 1st 5 innings or props first. Why do you want to compete with guys that are smarter and 100x more experienced than you?
mathy......because I'm a Canadian idiot.......seriously though, that's the reason.
Working on something of this magnitude would keep my mind off gambling day-to-day, which is really starting to get to me. I like to think big...whether that amounts to anything, who knows!
If you're just looking for historical data, retrosheet.org is the way to go - much easier (IMO) dealing with their event datasets from which you can compile any statistic you want, rather than trying to download thousands of individual stats from the web and then tracking, matching, and combining them.
pitching is everything in baseball, all the hitters can hit a pitcher on a off nite, no matter who they are... I just keep track of who is valued high but is the most likely to have a off nite ... like last season, I cleaned up betting against chris carpenter, he struggled like every other game and the odds never reflected it til like mid season, then he improved and I stopped betting against him (first 5 inning bets)
Thanks for the responses.....those sites are loaded.
Is there a good site where I can get historical lines from say 5dimes from the past 5 years to test when I do get to that point. A nice organized layout beats scrapping through individual days, but that might be the only option.
You can get a season's worth of lines team by team at Covers.com -- once you have it programmed, it's simple to sequentially scrape all the teams year by year.