Pitching Valuation - Moving from Theory into Practice

My valuation discussion earlier this week generated more interest when it came to pitchers than it did for hitters.

And why wouldn't it? My conclusion about hitters was predictable. Adding the pool of undrafted hitters to the draft pool and then recalculating hitter values had a negligible effect. The players at the top of the pool lost $2 or $3, the players in the middle lost $1 or $2, and the players toward the bottom lost $1 or weren't impacted at all.

Pitchers, on the other hand, are vastly impacted by players who aren't purchased in auction. Last year, Jered Weaver earned $26 and wasn't purchased in most leagues, primarily because he wasn't eligible. It's incredibly unlikely that a hitter will do the same thing. So there's an incredible amount of pitching volatility.

sas4 suggests this is why he pays more for a Johan Santana.

This tells me to buy Santana and fill the second roster spot with a scrub -one who will not hurt me with a bad era. If he gets wins, ratio (or 'k's), so much the better.

I have no arguments with paying a few extra dollars for a Johan Santana. Where you need to be careful is with the rest of the pitchers you buy to fill out your staff.

If I want to increase the probability of making money on my pitchers I should overpay for Santana and get a bargain pitcher for 1/3 his price (Byrd, Vasquez, and Penny fell into that category in my league this year) rather than two $23 starters.

Overpaying for Santana is the easy part. The difficult part is finding those "bargain" pitchers.

If you believe what in seasonal and not draft valuation, you should avoid pitchers like Byrd, Vazquez and Penny entirely.

Using draft day rosters to calculate values, 33 American League starting pitchers earned $9 or more last year. Remember, the average pitcher earns $8.67, since pitchers earn $78 per team, not $130.

If you choose to sally forth with "seasonal" values instead, there are now only 17 A.L. starters who earn $9 or more.

You want Paul Byrd? Really? He only lost $2 using draft valuation, but he lost a whopping $14 once you factor in the entire A.L. pitcher pool.

Vazquez earned $8 using draft values from last year, making him acceptable as staff filler. If you're using draft values. But matching him up against the best 108 pitchers from 2006 puts him on the red side of the ledger as well, and he'll lose your team $3.

This is yet another reason why Alex doesn't use full season values to calculate player values, and why I don't either. Regardless of whether or not we're using draft values, season values, or some other method, we know that we need a minimum number of innings to not finish with zero points in ERA/WHIP.

Using "season" values, the best 108 American League pitchers last year complied a total of 9,288 innings. That leaves them 2,712 innings short, or 226 innings per team.

Using "draft" values, the best 108 American League pitchers threw up 11,352 innings. This is still short by 648 innings, or 54 per team. But it's a lot closer.

This is the conundrum in evaluating pitchers. We know that a healthy percentage of pitchers that we buy are going to finish in the red, but we buy them anyway. We might very well cut bait in mid-April or early-May, but by then the damage is typically done.

The other problem with the best 108 pitchers in the seasonal valuation model (I need a better name for this) is that middle relievers are just as unpredictable as starting pitchers. In some ways, middle relievers are more unpredictable, due to limited sample size. A starting pitcher who allows 10 ER in 4 IP will have plenty of opportunities to make up that bad outing due to the quantity of innings he'll throw up. A reliever who goes out and gives up 8 ER in an inning is going to need to throw up a month of zeros just to get back to a respectable level.

So another reason draft valuation works when it comes to pitching is that it reflects the practical nature of our auctions. We don't like the fact that we have to buy 1,000 innings, and three or four starters at a minimum in our auctions. But we do it because we know from experience that the following is true:

1. We need to make innings, and the quantity and quality of free agent starters in April and May is poor. The odds of getting lucky with someone like Odalis Perez, Brian Bannister or Miguel Batista are slim.
2. Middle relievers are unpredictable as well. Dennys Reyes earned $16, Scott Proctor $12 and Wes Littleton $9, but did anyone seriously see that coming?

Because so many more middle relievers go undrafted, this makes them even more fungible. The top 10 undrafted A.L. pitchers last year were Je. Weaver ($26), Reyes ($16), Hector Carrasco ($13), Proctor ($12), Jeremy Sowers ($12), Pat Neshek ($12), Wes Littleton ($9), Matt Thornton ($9), Boof Bonser ($8) and Brandon League ($8). Seven of the top 10 free agent pitchers were relievers. More importantly, most of the relievers were pitchers who went for minimal FAAB bids and were available from Day One. Weaver was a farm player, Sowers was FAABed for $25, and Bonser for $17.

Seasonal valuation is entirely theoretical. But it doesn't take the fact that we need to make innings into account. Building a pitching staff of Johan Santana and eight middle relievers with 30-40 innings each won't work in a league with an innings requirement. The draft values reflect to some degree the market for pitching. You can try to sneak in a Paul Byrd or a Javier Vazquez in your auction. But they won't go for $1. And, if you wait for the $1 starters to round out your staff, they won't be there. Unless you want Edwin Jackson or Jae Weong Seo.