Doug Melvin on Sabermetrics | Disciples of Uecker

Disciples of Uecker

We'd like to go to the Playoffs, that would be cool.

Doug Melvin on Sabermetrics

By on March 13, 2013

Brewers’ General Manager Doug Melvin joined WSSP’s Tim Allen late last week to talk about the upcoming season and other things Brewers. I highly recommend listening to the interview which starts around the 35 minute mark of the link provided. It has more than a few interesting points. Particularly, Melvin came clean about the fact that despite what had been said previously, giving up a first round pick for Kyle Lohse is essentially a deal breaker.

Perhaps the most interesting thing was the rather round about, winding answer that Melvin gave when asked about the subject of Sabermetrics. He hit on some things that he’s said before about the use of numbers to make decisions, but also broke some new territory, at least from what I’ve heard listening to him over the years. The following is a rough transcription of what was said:

Allen: Final thing for you: there is a big debate on my show, Doug, every year… 

Melvin: Really?

Allen: About how…ok (both laughing)….specifically, though, about analytics and Sabermetrics, and I know that you can be a numbers kinda guy. Where does it fall for a GM when it comes to that sort of thing when you compare analytics, which is very important, and I understand that the game is based on numbers, but when you compare that to the chemistry, or the grit, or the heart, or the situation. I get calls all the time that ask “why did Roenicke make this move, the numbers indicate that it was opposite?”  It’s a good debate, it really is because they both make sense at certain times.

Melvin: Yeah, someone asked me, am I old school or new school and I said I’m in school because you’re always learning in this game. Any time we have an opportunity to learn something, we’re willing to sit down and look at it. I believe in –  

Allen: …but do you fall more…

Melvin: I believe in some of the analytics, but I have to know where the information is coming from. There’s 2 or 3 different formulas for WAR, you know, Wins Against Replacement, and I’m not smart enough to figure all that out.  You know, someone told us Prince Fielder’s WAR was 6, so does that mean we won 96 games and if Prince wasn’t with us we’re going to win 90? That doesn’t work that way. The game, there’s so many, the game is so complex but there’s so many things that can effect one play. It can be umpiring, it can be weather conditions, it can be a bad bounce, it can be the placement of the fielder, it can be the certain pitch and there’s so many variables that can effect a play and a number of plays that it’s very hard to determine all of the analytics. 

Some of them, I do believe in some of them. I’m a big believer in ball park effects, I think ball parks effect performance a lot. Certain players can perform better in certain ball parks and that’s just their skill set. The same guy that might pitch well in San Diego might not pitch well in Cincinnati. So I’m a big believer in that. The other part of that is injuries too. Sometimes players play with a little bit of an injury, so their performance may not be up to par, but he may be playing at 80 or 90%.

I believe in some of them, but I’ve been burned on a lot of them too. We had a player by the name of Doug Jennings back in late 80′s and 90′s who had a career .400 on base percentage in the minor leagues and we gave him a chance in the big leagues and Oakland gave him a chance in the big leagues and he just never could find it in the big leagues. It’s different, the biggest thing is you have to learn to adapt, but the physical skills of the players at the big leagues are better. I like to think that a lineup, that you have to have 4 guys who can hit guys 96 to 97 miles an hour. If you don’t, you’re not going to be able to score runs, because that’s what is pitching out there in the 8th and 9th inning. 

I’m a big believer in getting all 27 at bats. When you hit into double plays or get guys thrown out , you might get 23 at bats that game. I would rather have 27 outs swinging the bat and we try to eliminate the 27 outs of the other team. You try to get the double play, they get 26 at bats, you throw a guy out stealing they’ve only got 25 at bats, you get another double play and they’ve only had 23 at bats. That’s big in that regard, it’s very simple. I think the game is a simple game and I think there’s times when we make it too complicated. 

Allen: You taught me a stat years ago that says a guy on first base with nobody out scores at a higher percentage than a guy on second with one out, but that doesn’t take into account inside the dugout. Who’s pitching…

Melvin: …who’s up to bat, what’s the count, do I have somebody that can pinch hit for him, defensively sometimes, certain guys playing defense. You might get a couple hits because they have their backup shortstop in who doesn’t have the same range as the starter. There’s so many variables in the game, and the game is a great game to watch, that’s why it’s fun to watch it. I try not to get too bogged down by the numbers but they’re out there, they’re part of the game today.

You look at it, Earl Weaver was a big believer in the numbers at certain times. But you talk about giving guys days off. He would only give a guy a day off if he thought it would give them a chance to win the game with the guy he was putting in. There’s things through the history of the game that you learn and that, but the numbers and variable are so, there are so many of them. You go into a science lab and experiment and all the conditions have to be the same. I use this example: if you’re dropping two rocks off a building and which one is going to hit the ground first? The conditions have to all be the same. The rocks have to be the same weight, the wind can’t be coming from one side or the other, the ground has to be level, all the conditions have to be the same. In baseball, all of the conditions are never the same when all these numbers get compiled. The conditions are different, ballparks are different, umpire is different, the pitcher is different. There’s so many different variables that can effect that one event…

Allen closes out the interview at that point talking final win totals. You’ll have to listen to hear that.

Set aside a few little errors that come from speaking off the cuff, and most of it makes pretty solid sense. Perhaps the most curious thing was his emphasis on out making, which he interchangeably called “at bats” and “outs.” It was a bit muddled, but what he was clearly talking about there was the idea that outs are precious and that you need to conserve them on offense and try to collect them as fast as possible on defense. This is perhaps the most basic concept of Sabermetrics, and he really seems to believe strongly in it. How, exactly, this squares with employing a manager who led the National League in giving up outs via the sacrifice bunt would have been nice to know, but alas, the question went unasked.

Anyway, this just seemed like something worth passing on. What do you think it says about how the Brewers are being run from an analytical standpoint? Does it make you feel better? Worse? The same?

Share Our Posts

Share this post through social bookmarks.

  • Delicious
  • Digg
  • Newsvine
  • RSS
  • StumbleUpon
  • Technorati

Comments

Tell us what do you think.

  1. Nicholas Zettel says: March 13, 2013

    Awesome! Thanks for posting this, Ryan.

    Over the last two offseasons, I feel like more and more has been made public about Melvin’s (and the Brewers’ front office’s) approach to analysis. While I don’t think he’ll get the game-changing press of other GMs (like those in Arizona, Tampa, Oakland, etc.), I think Melvin and company run a front office that is aggressively analytic. Specifically, I think we can see that with their approach to fielding shifts (as one example).

    I also think the way he approaches the context of the game is smart, too; for all the analytical things written about Yuni Betancourt, for instance, Melvin always defended him, and sure enough, in the context of the Brewers’ batting order, Betancourt was able to produce an average number of runs for a SS (for a number of reasons). I think it’s decisions like that that show a good balance, meaning that he won’t use statistics alone simply for that sake of it.

    • Nicholas Zettel says: March 13, 2013

      BTW, I think what Melvin said about outs makes sense in the context of the Brewers’ ability to beat other teams in terms of scoring manufactured runs, versus allowing manufactured runs. Following Bill James’ Handbook, in both Roenicke seasons, the Brewers were among the better (or best) NL clubs in scoring manufactured runs, and they scored more manufactured runs than they allowed in both seasons.

      I think this reflects what Melvin is talking about — it’s not simply using each and every PA, but it’s also using your PA in ways that are more efficient than your opponent. I think in the Brewers’ case, they are scoring their runs in as many ways as possible, and alongside their power/speed, they are also limiting their opponents’ ability to come back with manufacturing runs.

  2. Vineet Barot says: March 13, 2013

    I am glad that Doug Melvin is a little more forward thinking than other GM’s. However, I see some basic flaws in his reasoning.

    Prince Fielder’s WAR was 6, so does that mean we won 96 games and if Prince wasn’t with us we’re going to win 90?

    Yes. Specifically, it means that if you replaced Prince Fielder with a Hunter Morris or a Mat Gamel or a similar “replacement level” player you’d win 6 fewer games.

    Most people also miss the point of WAR. It doesn’t matter if Fielder actually provided 6 more wins that particular season. It’s a way of evaluating player values across positions, eras, ballparks and time on field. How valuable is a player that hit .320 in 100 games vs. .290 in 150 games? How about a closer with 38 saves vs a workhorse with 200 innings and 3.7 ERA (about the same in case of Aroldis Chapman vs Bronson Arroyo).

    ———
    You taught me a stat years ago that says a guy on first base with nobody out scores at a higher percentage than a guy on second with one out, but that doesn’t take into account inside the dugout. Who’s pitching…
    ———
    When looking at large amounts of data, one can ignore other variables without compromising the conclusions. The data trend show us that bunting should NOT be the default but rather the exception (pitcher at-bat, 1 run game in late innings with runner on 3rd etc. )

    ———-
    The conditions are different, ballparks are different, umpire is different, the pitcher is different. There’s so many different variables that can effect that one event…
    ———-

    Once again, that’s the whole point of looking at large amounts of data. The probability of rolling a 4 on a die is exactly 1/6 but certain die can be loaded or some nights I keep rolling a frickin’ 1 and my sister beats me at monopoly. That doesn’t discount the law of large numbers.

  3. D-Rock says: March 13, 2013

    No Doug, the rocks don’t have to be the same weight.

    • Nkg says: March 15, 2013

      Thanks for pointing that out, I know physics aren’t crucial to running a ball club, but c’mon. We’re talking about some pretty basic facts here

    • Ryan Topp says: March 16, 2013

      I almost mentioned it in the post….

Trackbacks

Websites mentioned my entry.

  1. Carlos Gomez and Managing Expectations | Disciples of Uecker
  2. The Duality of Rickie Weeks | Disciples of Uecker
  3. Rounding the Bases: Managers and General Managers | Disciples of Uecker
  4. Ron Roenicke, Bunting, and Evolution | Disciples of Uecker

Add a Comment

Fill in the form and submit.