[KingHutch]

I'm looking at putting together an incredibly basic model inspired by a book (Trading Bases) with my own tweaks and then adding to it over time.

Had a quick question regarding calculating a change in WAR from one season to the next with respect to pitchers. This would be done at the start of a new MLB season and so PECOTA predictions of WAR would be used and converted into runs scored. Again, there are lots of assumptions and it's basic - I understand this, just a starting point that I can build on.

Would you add up the starting pitchers' WARs of the previous season (say 2014) and subtract the predicted 2015 WAR of the pitcher who is actually starting on a given day, multiplying this number by 5 (162/32) so you get the WAR of a pitcher, pitching 162 games. This number would change every day and have to be adjusted.

This might sound confusing or stupid but, put simply what I mean is:
Is change of pitchers WAR at the start of a new season (2015), on a certain day = 2014 WAR of all starting pitchers - (2015 WAR of pitcher starting that day x 5)

This is ultimately to get a win percentage through Pythag expectation (RS +RA) and so runs allowed would be adjusted every day depending on who is the starting pitcher.

Hope someone can help and this doesn't sound too ridiculous/stupid, I'm mostly doing this for fun

[evo34]

Do you have WAR for starts only? I.e., what about pitchers who spent half of 2014 starting and the other half relieving? My 2 cents: evaluate the true skill of the current starter vs. the YTD performance of his team's entire pitching staff.