Friday, December 26, 2008

Statistical Predictability: Stolen Bases

After looking at last year's best average hitters along with their predicted counterparts, let's move along to stolen bases.

Top 10 A.L. Stolen Bases 2008

RankTop 10 SB
AB
HR
RBI
SB
BA
$SB
$
Sal
AP
SW
2007

1Jacoby Ellsbury
554
94750.280233319181510
2
B.J. Upton
531
96744.273203133343730
3
Ichiro Suzuki
686
64243.310193534353040
4
Brian Roberts
611
95740.296183327202334
5
Grady Sizemore
634
339038.268174035322931
6
Chone Figgins
45312234.276151929
29
28
32
7
Carlos Gomez
57775933.2581522111525
8
Alex Rios
635157932.291143331282929
9
Johnny Damon
555177129.303133319202222
10
Ian Kinsler518187126.319123426182422

average
575126137
.288173126252426

Unlike the batting average list, more than half of these players are pure steals plays. More than half (53%) of their earnings come from the stolen base.

Half of them (Ellsbury, Roberts, Suzuki, Gomez, and Damon) are purchased mostly or purely for their base stealing abilities. What jumps out at me more is that there isn't a surprise on this list.

Top 10 A.L. Stolen Base Predictions (Rotoman): 2008
RankTop 10 SB
AB
HR
RBI
SB
BA
$SB
$
Sal
AP
SW
2007

1Chone Figgins
453
12234.276151929292832
2
Carl Crawford
443
85725.273112141403740
3
Ichiro Suzuki
686
64243.310193534353040
4
Brian Roberts
611
95740.296183327202334
5
Jacoby Ellsbury
554
94750.280233319181510
6
Julio Lugo
26112212.2685821
16
16
18
7
Carlos Gomez
57775933.2581522111525
8
Bobby Abreu
6092010022.29693429272928
9
B.J. Upton
53196744.273203133343730
10
Grady Sizemore
634339038.268174035322931

average
536105634
.282152828272527

Seven of the 10 best base stealers in the A.L. in 2008 were also predicted to be the best by Rotoman. Crawford and Abreu just miss, while Lugo misses by quite a bit.

But you'd expect the 10 best expected of anything to be a great disappointment. These guys aren't. They earn what they were paid in 2008, even though they got $1 raise on average from 2007.

This would seem like good news for a no-power strategy. The six point drop in batting average from the actual Top 10 by the predicted Top 10 certainly isn't small. But this still would have won my American League's batting title, while Suzuki, Roberts, and Gomez would have left their buyer 14 steals short of first place in the category, at a relatively modest $72 salary (It should be noted that these three went for $96 in my A.L.).

The thing that makes me think about a Sweeney Plan, though, isn't just the results.

Most of these guys are pure steals plays. Only Abreu and Sizemore come with power; the rest of these guys are steals all the way. It's true if you bought Figgins and Crawford instead of Roberts and Ichiro that you got burned, but you still would have had money to spare.

In the N.L., the pure steals plays used to be all over the place.

Top 10 N.L. Stolen Bases 2008
RankTop 10 SB
AB
HR
RBI
SB
BA
$SB
$
Sal
AP
SW
2007

1Willy Taveras
479
12668.251282919192022
2
Jose Reyes
688
166856.297234246444244
3
Jimmy Rollins
556
115947.277203137373843
4
Michael Bourn
467
52941.22917161718168
5
Juan Pierre
375
12840.283172218142134
6
Shane Victorino
570145836.293153025
23
20
25
7
Hanley Ramirez
589336735.301153941414151
8
Matt Kemp
606187635.290153327212219
9
Matt Holliday
539258828.321123839383845
10
Carlos Beltran
6062711225.284103531343034
11
Randy Winn
598106425.306102711141123

average
552156140
.286173128282732

These days they're much harder to come by.

Tavares, Bourn, and Pierre are the only ones on this list. Everyone else comes with double-digit homeruns, and a price tag that's $2 higher per player than in the A.L.

The SB $ earnings to overall earnings ratio is the same, though.

The National League definitely wasn't as reliable as the A.L. when it came to predictability.

Top 10 N.L. Stolen Base Predictions (Rotoman): 2008
RankTop 10 SB
AB
HR
RBI
SB
BA
$SB
$
Sal
AP
SW
2007

1Jose Reyes
688
166856.297234246444244
2
Hanley Ramirez
589
336735.301153941414151
3
Juan Pierre
375
12840.283172218142134
4
Willy Taveras
479
12668.251282919192022
5
Jimmy Rollins556115947.277203137373843
6
Rajai Davis
18004.056204
3
1
11
7
Rafael Furcal
1435168.35731123192327
8
Corey Patterson
366103414.20566614722
9
Eric Byrnes
2066234.2092227252739
10
David Roberts
107095.28421811814

average
35383328
.272121823232331

The first five guys on this list were predictable. The last five tanked. Injuries (Byrnes, Furcal, Roberts) had something to do with it, but the presence of Davis and Patterson on this list tells me that the days of most National League teams carrying at least one big-time burner are over.

In the N.L., the market is definitely a little more cautious than it has been on these burners. Throughout most of these lists, it has been beating both Patton and Sports Weekly in groups of 10 time after time. Here, they spend $23 per player, but that's a big drop from the $31 these guys earned the year before. At $23 per player, they're not dumping steals, but they're definitely letting the second-tier players like Roberts and Patterson go if the price isn't right.

1 comment:

Anonymous said...

I went with a Sweeney light plan in one of my 5x5 leagues this year. Most of the big power hitters were kept, so I dealt for Ichiro and tried for Avg, Runs, and SB, while trying to win 4 of the 5 pitching categories.

It worked to a great extent and I would have won had I taken Carlos Quentin (as was my plan) instead of Craig Monroe.

I was hoping to do pretty well in RBI by just pure AB accumulation.

My core offense was Ichiro, Gomez, DeJesus, and Lugo with Youkilis, Cano, and Drew for what I believed batting average and runs. I wound up getting Longoria for $51 to shore up my RBI.

Pitching staff was Halladay, Garza, Slowey, Duchscherer, Ervin Santana, Escobar to trade and middle relievers.

Of course all of the pitchers worked out perfectly, so it's like hitting the lottery. Targeting good K/BB pitchers with fine minor league track records does work. Santana was the only gamble.

I find targeting batting average and ABs is where the market is misallocated. I always have my projected high average hitters at higher prices than my opponents. I turns out that I also undervalue HR, so I rarely draft them. But, the ABs are the key. Getting high AB players with good averages not only produces runs, but usually RBI as well.