Sunday, June 26, 2011

Rotisserie Values and Pitcher Metrics: Is There a Corollary?


Back in May, Toz wrote an article discussing pitcher metrics versus Rotisserie earnings. In particular, he discussed K/9, BB/9 and G/F rates because of an article Eno Sarris wrote for Roto Hardball.

Rotoman thought that Toz's post had a couple of flaws:
...comparing pitchers to single season roto earnings won't work because you're looking at these component stats in order to clear up the noise inherent in roto earnings.

What shows up in the earnings leaders are outliers, like Buchholz, who in theory at least won't sustain because their components aren't so good...
Fair enough. Instead of looking at one season's worth of Rotisserie earnings, let's look at two years worth of earnings.

I put together a spreadsheet of the top 100 pitchers in innings pitched in 2009-2010 combined. Then I ranked these pitchers by K/9, BB/9, and G/F. Next, I added up the rankings to give each pitcher an aggregate score. The lower the score, the better the pitcher.

For example, Roy Halladay ranks 28th in K/9 from 2009-2010 among these 100 pitchers. He had the best BB/9 and was tied for 19th in G/F ratio. This gives him an aggregate score of 48.5 (29 plus 1 plus 19.5), making his aggregate ranking among the three metrics the highest.

I then devised a league neutral ERA/WHIP/K formula and applied it to each pitcher for these three categories (wins do not sit for this portrait). Then I figured out how much money each one of these pitchers earned per inning.

So how useful are these benchmarks in telling us how good a pitcher is in Rotisserie baseball?

Earnings by Metric Ranking in Groups of 10
1-10: $407.36
11-20: $258.43
21-30: $178.70
31-40: $151.44
41-50: $96.98
51-60: $101.83
61-70: $163.23
71-80: $128.80
81-90: $79.54
91-100: -$15.94

At the top and bottom ends of the scale, the metrics work extremely well. In the middle, though, everything is screwed up. It would seem that you want to avoid pitchers in the 41-50 group and push harder to get pitchers in the 61-70 group. 

Does one specific component metric work better than another?

Dollar values by metric ranking: 2009-2010
#
K/9
BB/9
G/F
1-10
$260.41
$242.49
$142.73
11-20
$296.38
$209.05
$208.97
21-30
$287.90
$185.55
$182.52
31-40
$207.02
$262.65
$107.77
41-50
$184.63
$158.24
$86.96
51-60
$140.65
$74.91
$141.12
61-70
$65.74
$79.96
$227.37
71-80
$50.80
$138.24
$149.33
81-90
$43.05
$77.29
$167.47
91-100
$30.79
$122.19
$136.13




K/9 by itself works very well, BB/9 is OK but uneven, and G/F by itself is nearly worthless, at least in this context.

Eyeballing some of the pitchers on the chart seems to confirm that there's too much variability in the G/F rankings to count it as any kind of valuable metric. Tim Hudson looks like a poster child for G/F pitchers but Fausto Carmona does not. Joel Pineiro might make you believe that a G/F profile pays off, but Jason Marquis will disabuse you of that notion awfully quickly. It looks like if a pitcher has a low K/9 rate, it doesn't matter that much if his outs are coming in the air or on the ground; he's still susceptible to the big inning.

1 comment:

Rotoman said...

Not that you guys aren't busy, but it seems worthwhile at this point to run the same numbers substituting HR rate for G/FB.

I wonder why Eno was substituting if the replacement category isn't better?

One other question that I don't know the answer for: Is GB rate the same as GB/FB ratio? Do they track together?