you're reading...
2017

Why we chose Sam Rice instead of Vlad Guerrero

So we recently elected Sam Rice instead of Vlad Guerrero. I suspect a number of people may wonder about that. Especially when we revealed that we have reason to believe that Rice has been significantly underrated by various forms of WAR. In a nutshell, many of the things that Rice does very well are not (yet) included in BBREF’s WAR calculations, and we believe that when they finally are, he will receive a considerable boost.

We could be wrong, too, so let me lay out the information, reasoning, and estimates for you today, and you can decide if you agree.

What WAR knows

Right now, BBREF’s WAR includes a very solid estimation for Rice’s batting performance at +184 runs. It includes a defensive estimate based on Sean Smith’s Total Zone of +56. It includes a positional adjustment of -96 runs for playing a high-offense position. None of this is especially controversial and squares with Rice’s reputation, his offensive and defensive statistics, and what we know about estimating value.

BBREF’s WAR also includes +13 runs for baserunning. This is an estimate derived mostly from his stolen base record. BBREF doesn’t have play-by-play baserunning for most of Rice’s career. Currently it has nothing prior to 1930, and until somewhat recently didn’t have that much. Similarly, they have no rating for Rice in Rdp, which is double-play avoidance above or below average.

Additionally, because of the lack of PBP information, BBREF’s fielding measures don’t include specific estimates of Rice’s outfield throwing. Both Miller and I rely more on DRA anyway for defense, and its arm estimates are notoriously off. In fact, I swap in BBREF’s Rof rating for all DRA throwing scores for outfielders in the PBP era. DRA estimates Rice gained 5.4 runs through his throwing.

So basically this all means that Rice’s baserunning, GIDP avoidance, and outfield throwing are either missing from his case or could stand a little more scrutiny. So we scrutinized them.

Instant Rice

Here’s the basic facts about Sam Rice’s baserunning.

  • BBREF credits him as +13 runs for his career
  • BBREF now has a great deal of PBP data for his final five seasons (ages 39–43, 1930–1934)
  • Their estimate shows him as basically a league average baserunner during these late decline years.
  • In those seasons we have caught-stealing data for, Rice stole 285 bases and was caught 143 times.
  • In the seasons where we have CS totals for him, his stolen base percentage was 66%.
  • In those final five years we have a lot of data for, he stole at a 62% clip, while the league was at
  • The AL in those years stole at a 59% success rate.

Recently, Retrosheet released information on numerous 1920s and 1930s seasons, some more complete than others. With the caveat that we may not have all the PBP data on Rice, what we’ve learned is that from 1930–1934, Rice was

  • picked off less than the AL in his opportunities
  • made fewer outs on base than the AL
  • took more bases on non-hit events (passed balls, wild pitches, etc) just a tiny bit more often than the AL.

Remember, BBREF is showing him as an average runner from 1930–1934. This doesn’t look like the record of an average runner, does it? I did some back of the envelope math. I pretended that every base taken, regardless how, was worth the same as a steal—about 0.18 runs—and that every out on base cost the same as caught stealing—about 0.32 runs. Those figures are from Extrapolated Runs. XR only cover the years 1955 onward, but we just need a baseline to work from. Obviously, this isn’t super scientific, but it’s getting us close enough. I came up with +6 runs versus the AL for 1930–1934. Seems pretty reasonable.

Next I comped Rice against runners from the play by play era. I looked for anyone who had 0 to +6 runs on the bases from age 39 onward. The list was too small, so I went back to age 38. I then looked for those who over their careers had a similar rate of stolen base attempts and stolen base success versus their leagues. Rice stole 91% more often than his league and succeed 17% more often than his league. This narrowed things down to the compiest of comps.

                   ATT/G   SB%   ATT V.  SB% V.  CAREER                
NAME               V. LG  V. LG   RICE    RICE   rBASER  PA   PER PA 
=====================================================================
Paul Molitor       + 93%   +17%   + 2%      0       78  12167  .0064
Willie Mays        +111%   +25%   +22%    +47%      77  12496  .0062
Barry Larkin       + 83%   +21%   - 9%    +24%      80   9057  .0088
Derek Jeter        + 71%   +11%   -22%    -37%      56  12602  .0044
Barry Bonds        + 97%   +14%   + 7%    -18%      44  12606  .0035
Craig Bigio        + 76%   +11%   -18%    -35%      54  12504  .0043
Ozzie Smith        +128%   +18%   +41%    + 6%      79  10778  .0073
---------------------------------------------------------------------
AVERAGE            + 94%   +17%   + 3%    - 2%      67         .0058

These players averaged 0.0058 runs above average on the bases for their careers. Applied to Sam Rice’s career PAs, we get +60 runs. Knock out the 13 runs that BBREF has him marked down for, and we get about 47 more runs than they credit him with.

GIDdy uP!

BBREF has very little information on Rice’s GIDPs. So I took a similar approach, using comps. They had to be lefthanded, they had to have at least 40 runs on the bases in their careers, and they had to have at least 5,000 plate appearances.

NAME               rGDP  rBaser   PA
======================================
Ichiro Suzuki       55     63   10466
Johnny Damon        50     77   10917
Carl Crawford       35     48    7178
Juan Pierre         28     56    8280
Willie Davis        27     47    7475
Joe Morgan          24     81   11256
Kenny Lofton        23     78    9235
Chase Utley         23     44    7323
Lou Brock           18     72    9468
Tony Womack         16     52    5389
Larry Walker        10     40    8030
Lenny Dykstra        6     46    5282
Barry Bonds          6     44   12606
Delino DeSheilds     4     45    6652
--------------------------------------
PER PA AVERAGE    .003
PER 600 PA         1.6
IN RICE’S PA      27.8</pre>

So here we get an estimate of about +28 runs for avoiding the deuce.

What’s in an Arm?

DRA’s arm adjustments are problematic, and Michael Humphreys acknowledges this in Wizardry.

We think that Harry Hooper was probably shorted by about 35 runs based on his assists record and how often he led his leagues. Right field is the only outfield position where leading the league and assists totals has a relationship to arm value (specifically Rof on BBREF). Hooper led his leagues three times and placed in the top five 8 times. This mirrored his reputation as the best right field arm of his time. Well, Sam Rice led five times and had several other high finishes. His reputation was not as strong as Hoop’s, so I’ve simply given him about 70% of Hooper’s value. Rice broke into the AL as a pitcher and despite an injury that may have contributed to the end of his mound career, the arm was obviously still healthy enough to rack up assists. Because I use a 2:1 combination of BBREF’s Rfield and DRA, I am replacing DRA’s arm value of 5.4 arm runs with my estimation of 23 arm runs. The net of which is 18 arm runs at 2/3s value, or +12 runs.

Adding It Up

So here’s what I’ve got from all this.

  • Running = +47 (net of BBREF’s estimate)
  • GIDP = +28
  • Arm = +12
  • Total = +87 runs
  • That’s somewhere around 8.5 Wins. Which is a ton to be missing from someone’s record but Rice’s arrows all point in this direction. But it turns out that Rice looks like he’s Ichiro of his time.

I’ve previously ranked Rice around 25th in rightfield among HoME-eligible players. Guerrero ranks 23rd. I Rice at 39 wins over his 7-year peak and 62 wins for his career. In my CHEWS system (similar to JAWS but with more peak emphasis), he rates as 47.8. These additional runs, distributed evenly across his career on a PA basis, would now yield a 39.9 peak and 70.4 career, translating to 53.6 CHEWS. That total is very similar to another comparable player, Tony Gwynn, as well as Andre Dawson, and Dwight Evans. Ichiro is about 1.5 points below Rice.

Even if you don’t take all my estimates at face value, Rice still has a strong case. Slash them in half and you get +44 runs, call it four wins. Now I’d rate him at 50.5 CHEWS. That’s in between Dave Winfield (51.1) and Bobby Bonds (50.3) with Willie Keeler (49.9), Sammy Sosa (49.8), and, yes, Vlad Guerrero (49.7) also nearby.

You might very well find another way to look at the data we have on Rice and extrapolate his value. I hope you do and that you report it in our comments! But this is enough evidence for us to elevate the top also-ran in right field to a HoMEr.

Advertisements

Discussion

2 thoughts on “Why we chose Sam Rice instead of Vlad Guerrero

  1. Great article Eric, I’ve been on the Sam Rice train for a while too…
    Anyone else from this time era that is being overlooked due to missing data like Sam Rice that you have discovered?

    Posted by Ryan | February 10, 2017, 8:37 pm
    • Ryan, thanks so much! I’ve worked up quite a few of these 1930s guys, and I’ll share findings in a subsequent article. I know you’re a HOM voter, so I’ve already worked up candidates still getting votes there. If you have any specific requests for a player active for at least a few years of the 1930s, let me know, and I’ll be sure to include them.

      One thing you notice when you start looking into this kind of stuff is just how low the SB%s were in the 1930s and 1940s. Well below any reasonable break-even point. So a half-decent baserunner can gain a few runs over the league that way. Another thing you notice is that there’s very wide variance with respect to how wild some players or teams are on the bases. Bob Johnson for example stole home nine times and was caught seven times. In 2016, MLB stole home just 14 times and was caught 19 times. Whereas in 1940, an MLB just over half as big as today’s league stole home 14 times and was caught 23 times. In 1930 it’s 38 SBH and 51 CSH. It’s not like the difference between the 1880s and the 1930s but it’s a somewhat wilder game on the bases than we play now.

      I’m excited to see what the deadball era looks like in this regard. It should be very different than the 1930s since run environments are so low.

      Posted by eric | February 11, 2017, 12:51 pm

Tell us what you think!

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

Institutional History

%d bloggers like this: