THE BOOK--Playing The Percentages In Baseball

The fangraphs FANS projections for Hughes are pretty wacky: 49 games, 35 starts (!), 150 innings (?). The rate stats matter much more in this case, and are much more reasonable.

The projections for Joba are total out of control. He’s projected for 39 starts (53 games total) and 157 innings.

On a rate basis, Hughes is projected (by essentially everyone) to be better than Joba this season.

The games and starts are extrapolated, and obviously it is not a good algorithm.

I have a good one handy somewhere.  I’ll give that to David at some point.

Can we really get much from this based on 12 non-random fan projection?

Yes, absolutely.  Do you want proof?

http://www.fangraphs.com/fanpdetails.aspx?playerid=4567

That’s FIVE John Johnson team fans and 250 non-team fans.  You can look up for pitchers if you like.  You’re going to get tight ranges.  I’m in heaven at 20 votes, and I’m good with 10 votes, and my minimum level is 5 votes.  You really don’t need alot frankly.

It may not matter in terms of runs in 2010, but if you believe that you need to build a pitcher’s workload up gradually and you think one of them has more upside in the long-term, shouldn’t that inform your decision? 

Suppose you think Hughes is the best starting prospect of the three, but he won’t get there until he builds up to a starter’s workload, shouldn’t he be the one to start?

The best way is what Earl Weaver did: long relief, spot starts, then a binge of starts, back to relief, then the next year, in the rotation.

It’s not an either/or process.  I don’t know why fans think of it like it is.