Moneyball and the mechanization and computerization of MLB

[Bellhorn04]

Well, don't forget, there have been other "noticeable" changes that you didn't notice.

 

Like dedicated bullpens, dedicated ninth inning specialists aka "closers", LOOGY's? 5 man rotations? Or even 4 man rotations?

 

How about in the 80's, when every player who got to first just automatically ran tried to steal second? And every team pitched out once per inning? That was the Golden Era of the Stolen Base. Ty Cobb's record of 96 stolen bases lasted from 1915 to 1962 and Maury Wills. At the start of the 1980 season, Cobb had sole possession of third place on the Single Season Stolen Base leaderboard. By the end of 1980, he was tied for fifth.

 

So you don't like everyone hitting home runs? What about when no one did? And players like Franklin Baker lead the league with NINE? And his nickname was "Home Run"!! Then that Ruth guy came along and smashed an unheard of 29 home runs for the 1919 Red Sox (a team that collectively hit 32 home runs).

 

What about stuff like Astroturf? If Ty Cobb got to play on it, think Pete Rose is the all time leader in hits?

 

There have been so may changes. Guys holding notecards doesn't bother me and never will...

 

Yes, there were a lot of changes before analytics. But they came (and went) a little more gradually.

 

The recent ones have come in more of an onslaught...

[Kimmi]

Sure, I realize the probability is calculated after the fact.

 

But it does show that the Dodgers went quickly from a 42% chance to a 73% chance, and how crucial those at-bats were to the outcome of the game.

 

I get that. But we don't know whether the outcome would be any different if Snell had stayed in the game.

 

Therefore, the only fair way to assess whether a move is a good move or not is to see what the difference in win expectancy between the two pitchers is before any events take place. We have to remember that just because a move does not work that doesn't mean that it was the wrong move or a bad move. Even the best move doesn't always end in a good result.

 

I really wish I could find the article that showed a very good example of this. It was a similar situation, I want to say with the Nats, where the manager was being heavily criticized for going with one pitcher over another when the move didn't work out. The article showed that the difference in the decision between one pitcher and another is remarkably small.

[Kimmi]

Then why did you bring up the top of win probability if it was not pertinent to this discussion?

 

Of course we don't know what Snell would have done, but we do know that, right up until the moment Cash jerked him out of the game, he was feeling great with excellent command and stuff and was pitching the game of a lifetime. To me that's sufficient evidence to say Snell--who had already struck Mookie and the next two Dodger hitters twice apiece--was a much better choice than Anderson, who said after the game he was tired and did have his good stuff, which is why he gave up the double, the wild pitch, and the groundout that brought Betts home with the go ahead run.

 

But let's ignore all that and just focus on the simple fact that after the game Kevin Cash told reporters that he decided before the first pitch that Snell would not be allowed to face any Dodgers hitter a third time.

 

I believe it was Bell who brought up the topic of win probability.

 

As I've already stated, I agree with what you are saying regarding Snell being the better decision. I'm just saying that the difference between staying with one pitcher over another is not very significant.

 

FTR, based on what I've read, almost every manager, if not every manager, has already decided before the first pitch how they are going to use their bullpen in different scenarios. Most managers will alter that plan based on the way the game is unfolding, but they do have most game decisions already decided before they start playing.

[Kimmi]

And I disagree with you because at the moment Snell was pulled off the mound and Anderson was brought in, the Rays were leading, 1-0, and Snell was not tired and was pitching brilliantly. Snell was the Rays best hope of winning that game. Bringing Anderson in, as we now know, guaranteed the Rays would lose.

 

And you know that Snell would not have made a couple of bad pitches and given up the lead?

 

No, you don't.

[Kimmi]

But if he was going strictly on most recent games, why on earth would he put in a guy who has had 6 bad games in a row?

 

FTR, analytics say that decisions should not be made based on recent games.

 

They should be made based off of season long data and projections.

[Kimmi]

Here are some lines from a Fangraphs article, written by Jeff Sullivan, that I found on the topic. This is mostly in response to the criticism that Boone received during the 2018 playoffs.

 

I’ve long been someone who didn’t think about managers too much. And part of that is because strategic decisions tend to barely move the needle. Fans frequently get on a guy for how he handles a bullpen, and fans frequently get on a guy for when he calls for a bunt, or for how he builds a lineup, but it all makes such a small difference in the odds. Short of conceding a random forfeit, a manager can’t blow a game. Realistically, a manager can’t move the win expectancy more than a small handful of percentage points. Everything is ultimately up to the players. It’s the players who have to execute. It’s the players who determine the outcomes. Every single player in the majors is good. The per-inning difference between one reliever and a slightly worse one is remarkably small.

 

It’s true that, by and large, managers can’t affect the game all that much. Not if we’re being realistic. There’s no would-be managerial candidate out there who would constantly bunt with Aaron Judge, or bring in a backup shortstop to be the closer. Managers all make pretty similar strategic decisions. There’s not that much of an opportunity to make a difference.

 

https://blogs.fangraphs.com/the-sense-behind-ripping-the-manager/

[Kimmi]

That play a couple of years ago when Mookie made a catch and then pulled out his cheat sheet and waved it because he was perfectly positioned. Some people thought that was great. But it made me kind of sick to my stomach.

 

I loved it. One of my favorite Mookie moments. The guy was almost in centerfield and barely had to move a step to catch a fly ball. How great is that?

 

That said, and this might seem contradictory, I'm not a fan of all of the extreme shifting.

[Kimmi]

I expect NFL football to be a highly technological game. It has been for a long time.

 

I don't like it so much with baseball. And I know a lot of other people feel the same. It's not about being right or wrong, there's no such thing.

 

What ultimately matters is whether it has a major impact on fan interest or not.

 

No robot umps!

[Kimmi]

I'm fine with that stuff. I think we're headed towards an iceberg though. Don't lose sight of what is happening in front of you just because you developed a strategy 3 hours earlier.

 

I agree. Not necessarily that we're headed towards an iceberg, but I agree with your last statement.

[Bellhorn04]

I loved it. One of my favorite Mookie moments. The guy was almost in centerfield and barely had to move a step to catch a fly ball. How great is that?

 

That said, and this might seem contradictory, I'm not a fan of all of the extreme shifting.

 

And you're also not a fan of bullpen games, or starting pitchers getting pulled so early. So you do sort of have some contradictions going on, maybe.

[Kimmi]

The effect on the actual game is much more noticeable now.

 

-Defensive shifts.

-Hitters trying to launch the ball, producing more home runs and strikeouts.

-Bullpen games.

-Starters being pulled early so they don't face hitters a third time.

 

As I say, it'll be the fans who ultimately judge whether the game is better or worse as a result.

 

As much as I'm a fan of analytics, I am not a fan of these things that you have listed.

 

That said, it is not turning me off from the game.

 

What might end up turning me off from the game is the insane amount of money being thrown at players these days. That kind of makes me sick to my stomach.

[Kimmi]

And you're also not a fan of bullpen games, or starting pitchers getting pulled so early. So you do sort of have some contradictions going on, maybe.

 

Yes, I do. It's weird, I know.

[Maxbialystock]

I believe it was Bell who brought up the topic of win probability.

 

As I've already stated, I agree with what you are saying regarding Snell being the better decision. I'm just saying that the difference between staying with one pitcher over another is not very significant.

 

FTR, based on what I've read, almost every manager, if not every manager, has already decided before the first pitch how they are going to use their bullpen in different scenarios. Most managers will alter that plan based on the way the game is unfolding, but they do have most game decisions already decided before they start playing.

 

I see nothing wrong with managers running thru scenarios before a game and tentatively deciding what they might do. Cash did that with Snell.

 

But, as you say, Cash's problem was that he apparently didn't watch the game and especially Snell closely enough. Moreover, I think he was unaware that his bullpen was no longer as reliable as it had been. Maybe he even yearned to do what Dave Roberts was doing, sending out six relievers and holding the Rays to 0 runs.

 

Surely you can't believe that "the difference between staying with one pitcher over another is not very significant." Dave Roberts actually did that, and it worked well, but Cash should have realized his bullpen wasn't nearly as good.

Edited October 31, 2020 by Maxbialystock

[Kimmi]

I see nothing wrong with managers running thru scenarios before a game and tentatively deciding what they might do. Cash did that with Snell.

 

But, as you say, Cash's problem was that he apparently didn't watch the game and especially Snell closely enough. Moreover, I think he was unaware that his bullpen was no longer as reliable as it had been. Maybe he even yearned to do what Dave Roberts was doing, sending out six relievers and holding the Rays to 0 runs.

 

Surely you can't believe that "the difference between staying with one pitcher over another is not very significant." Dave Roberts actually did that, and it worked well, but Cash should have realized his bullpen wasn't nearly as good.

 

I have always said that analytics cannot be the end all be all. There has to be a balance between using analytics and using the human element side of the game. A manager and FO have to be willing to change their game plan mid course if the game calls for it. I believe that even the staunchest analytics supporters all agree with this.

 

That said, if there is a 50-50 toss up between two decisions, one based on analytics and one based on the human element, I'm going with analytics.

 

And yes, I absolutely believe that the difference between staying with one pitcher over another pitcher is incredibly small.

[Bellhorn04]

And yes, I absolutely believe that the difference between staying with one pitcher over another pitcher is incredibly small.

 

And yet analytics seems to dictate frequent pitching changes.

 

It's no wonder this stuff gets confusing.

[Kimmi]

And yet analytics seems to dictate frequent pitching changes.

 

It's no wonder this stuff gets confusing.

 

Fair enough. Analytics dictates a lot of things with the understanding that most changes are insignificant, such as batting order. One reason why managerial decisions have such little impact is because every decision that could affect the outcome of a game also has a countereffect.

[Bellhorn04]

Fair enough. Analytics dictates a lot of things with the understanding that most changes are insignificant, such as batting order. One reason why managerial decisions have such little impact is because every decision that could affect the outcome of a game also has a countereffect.

 

Say it's the ninth inning and you have a one-run lead. You have 5 pitchers available in the bullpen, one of whom is your closer - let's say he's been pitching nearly as well as 2013 Koji - and the other 4 are mediocre. What is the countereffect of using your closer, as logic would dictate?

[Maxbialystock]

I have always said that analytics cannot be the end all be all. There has to be a balance between using analytics and using the human element side of the game. A manager and FO have to be willing to change their game plan mid course if the game calls for it. I believe that even the staunchest analytics supporters all agree with this.

 

That said, if there is a 50-50 toss up between two decisions, one based on analytics and one based on the human element, I'm going with analytics.

 

And yes, I absolutely believe that the difference between staying with one pitcher over another pitcher is incredibly small.

 

You absolutely believe that? If you do, please tell me why MLB teams pay a whole lot more for some pitchers than for others. Because they are stupid and don't realize one pitcher is just as good as another and that the difference between pitchers is "incredibly small?"

[Kimmi]

Say it's the ninth inning and you have a one-run lead. You have 5 pitchers available in the bullpen, one of whom is your closer - let's say he's been pitching nearly as well as 2013 Koji - and the other 4 are mediocre. What is the countereffect of using your closer, as logic would dictate?

 

First let me say that managers are not typically going to make moves that are so egregious, as Sullivan pointed out in his Fangraphs article. The decision would be more along the lines of using a very good set up man to close out a game versus using the regular closer to close out the game.

 

For one, it would save your closer for the next night (or the one after that). So perhaps the manager costs the team 1/10 of a win by going with the set up man in tonight's game, but he'll gain that much of a win back the following night. Or vice versa.

 

There are probably several other 'countereffects' based off of specific numbers or splits of the particular players involved.

[Kimmi]

You absolutely believe that? If you do, please tell me why MLB teams pay a whole lot more for some pitchers than for others. Because they are stupid and don't realize one pitcher is just as good as another and that the difference between pitchers is "incredibly small?"

 

As I just posted, we're not talking about an egregious decision, like batting your pitcher #1 in the line up. Managers are not going to make those types of moves. We are talking about staying with a very good pitcher who is working on his 3rd time through the line up versus a very good reliever who is facing the batters for the first time in the game. The difference in such a decision is incredibly small.

[5GoldGlovesOF,75]

As I just posted, we're not talking about an egregious decision, like batting your pitcher #1 in the line up. Managers are not going to make those types of moves. We are talking about staying with a very good pitcher who is working on his 3rd time through the line up versus a very good reliever who is facing the batters for the first time in the game. The difference in such a decision is incredibly small.

 

One pitching move that still haunts me was in the last game of the '98 ALDS vs. Cleveland in Pedro's first year. Sox were ahead in the elimination Game 4 and Derek Lowe was pitching lights out (5 batters, 5 outs, 13 strikes in 19 pitches). Jimy Williams yanked Lowe for "automatic" Tom Gordon, who didn't have it that night and blew the save. The worst part is that Pedro would've started a winner-take-all Game 5.

 

Doesn't seem like such a big deal now, but at the time the title drought was 80 years...

[illinoisredsox]

One pitching move that still haunts me was in the last game of the '98 ALDS vs. Cleveland in Pedro's first year. Sox were ahead in the elimination Game 4 and Derek Lowe was pitching lights out (5 batters, 5 outs, 13 strikes in 19 pitches). Jimy Williams yanked Lowe for "automatic" Tom Gordon, who didn't have it that night and blew the save. The worst part is that Pedro would've started a winner-take-all Game 5.

 

Doesn't seem like such a big deal now, but at the time the title drought was 80 years...

 

Darrell Johnson never should have pinch hit for Willoughby in Game 7 back in 75. I said it as at the time as a 17 year old and I’ll say now at 62.

[a700hitter]

Darrell Johnson never should have pinch hit for Willoughby in Game 7 back in 75. I said it as at the time as a 17 year old and I’ll say now at 62.

And you are still right about it.

[Bellhorn04]

Darrell Johnson never should have pinch hit for Willoughby in Game 7 back in 75. I said it as at the time as a 17 year old and I’ll say now at 62.

 

We don't forget details like that, us Sox fans.

[Bellhorn04]

As I just posted, we're not talking about an egregious decision, like batting your pitcher #1 in the line up. Managers are not going to make those types of moves. We are talking about staying with a very good pitcher who is working on his 3rd time through the line up versus a very good reliever who is facing the batters for the first time in the game. The difference in such a decision is incredibly small.

 

I certainly understand the principle there. The only issue is with how the analytics can take into account the conditions in play in that specific game. Such as the fact that Anderson had clearly been struggling in recent outings.

 

Alex Speier (who said that pulling Snell was defensible in itself) said Anderson wasn't the same pitcher as he was in the regular season, and therefore was not an upgrade over Snell, so he disagreed with that part.

 

I will also say that I'm no longer discussing this for the purpose of proving how wrong Cash was. It's time to give the man a break. I'm just interested in the analytical part of it.

[mvp 78]

Anderson was not a very good reliever for the previous week or so.

[Bellhorn04]

Anderson was not a very good reliever for the previous week or so.

 

Previous 6 appearances Oct 9-24

 

9.2 IP

7 ER

13 H

4 BB

5 K

3 HR

[mvp 78]

Previous 6 appearances Oct 9-24

 

9.2 IP

7 ER

13 H

4 BB

5 K

3 HR

 

If he was pitching in the playoffs the way he had in the regular season, the move would be defensible.

[Bellhorn04]

If he was pitching in the playoffs the way he had in the regular season, the move would be defensible.

 

That's what it comes down to, I think.

[Maxbialystock]

That's what it comes down to, I think.

 

Agree. Pulling Snell and inserting Anderson was indefensible.

 

And just think. If Alex Cora comes back to manage, he won't have to worry about these difficult decisions because Chaim Bloom will just hand him the famous Rays' Book of Foolproof Plays and In Game Decisions.

 

And, once that becomes a norm for MLB managers, the Commissioner can dump the umps and bring on the robots!!!

 

If we fans can't have mechanically perfect baseball, why bother to watch?