Where Rocco gets it wrong OR What you can't see can kill you

[Aerodeliria]

When the Twins opted to select Rocco as their field boss, I was all in. I thought that someone who used spreadsheets and computers to help in decision-making was going to elevate the Twins to new heights. Certainly talent matters, but it seemed that Twins had a pretty good farm system supporting it, and a few important players added here and there meant that I thought the Twins were going places.

However, the past few years there have been many-a-complaint about Rocco not relying on his gut or not having instincts for the game. These complaints are often countered with just the opposite kind of complaints. He does not use enough analysis in key situations and so makes poor decisions. I contend here that they are simply opposite sides of the same coin and Rocco really does cling to his spreadsheets, but it can look like he is going with his gut when he really isn't.

The problem here IMHO is that Rocco does not understand statistical analysis as well as he thinks he does and so he is blind to things that he should be aware of. Let me explain before firing the poison arrows in my direction. (You can do it later.)

Let's first look at the woeful pinch-hitting of Margot. When Margot was 0-10 as a pinch-hitter, Baldelli was still looking at the percentages...Margot should pinch-hit here for Julien because the A's just brought in a LH pitcher and Margot hits LH pitchers at a much higher rate when compared to Julien. The spreadsheet showed that Margot had to be a much better option, so after Margot grounds out to short to go 0-11 as a pinch-hitter, it doesn't change anything in Rocco's mind. He's looking at it like flipping a penny 100 times...the odds are that 50% will be heads and 50% will be tails, but this is erroneous thinking....The problem for Baldelli is that it is not at all like flipping pennies because when flipping pennies there are no corrupting variables. In the case of Margot's pinch-hitting, one cannot simply say, "Well now he is bound to start collecting hits because the odds are really in his favor." Because we assume that Margot is a human, we can also assume that he has emotions. Baldelli does not consider emotional variables because we have no way to measure them during a game, so they are ignored by him. Because ALL teams now use these spreadsheets, this is where instincts become important. The good managers are still using the spreadsheets but they are also able to sense the ephemeral emotions that are at play. Is 0-10 affecting Margot's hitting? Yes! Why? Because he is probably feeling increasing pressure and is somewhat more anxious than at other times. We could measure his psychological state with a questionnaire and then run the data, but it would be impossible during a game and a player would likely lie anyway, so this is where in-game management is key. Letting Margot pinch-hit will not increase his odds as the data on paper would indicate because variables not on the spreadsheet are taking their toll on Margot's hitting.

 

I'll use one more example. Zebby Matthews was pitching very well against Boston as the Twins clung to a 3-1. In the 5th, he gives up a two-out double bringing up the top of the order. This is the 3rd time through the order and the spreadsheet says that Boston should start ripping Matthews. However, Matthews is probably pitching his best game of the season up to that point. Jarren Duran comes up to the plate and Baldelli jogs out and signals immediately to bring in the LH pitcher Irvin to face the LH hitter Duran. There were many a-groan in Twinsville including my own. The spreadsheet shows that Irvin has a better chance of getting Duran out than Matthews, so for Baldelli, it's a no-brainer. Again, all of the psychological factors that are corrupting his data, he cannot see, such as Matthews' confidence being pretty high and Irvin's performances (likely due to increased anxiety) in such situations being pretty bad (after all that's why Baltimore jettisoned him in the first place). Because in the confines of a live game, these kinds of things cannot be measured, so in Baldelli's mind they simply don't exist. And, we are aware of the very costly meltdown that followed. The effect of being unable to sense emotional factors that are at play can be devastating. They can even have a boomerang effect. For example, the next time Matthews pitches in a tight game will he not have the smallest of nagging doubts that if it is the 3rd time through the order that he will automatically be yanked?

 

Psychological variables can be measured. We do it all the time in assessing students via questionnaires and interviews. Granted, it is much tougher for managers on the field to sense these things, but very good managers are able to make pretty good assessments. I remember when TK tried to take the ball from Jack Morris in the 10th inning against the Braves and Morris said, "I have one more inning in me. Let me pitch the 10th." TK just handed him the ball because he could sense that the confidence was real. Baldelli absolutely takes that ball and turns it over the relievers because that is what the data show.

[The Great Hambino]

Good breakdown of the blind spots Rocco seems to have in implementing analytics.  

I'd argue another way he misuses them is in scope.  This shows itself most glaringly in early pinch hitting deployment.  Going all-in on the platoon advantage when the spot is likely to come up in the order again is so short-sighted.  How many times were ninth-inning rallies killed by Margot flailing at right-handed pitching because he hit for Larnach in the sixth and now you have no bench.  The goal is to win the game, not the sixth inning.   He could use the exact same analytical information - the same spreadsheets, if you will - and map out a plan for the rest of the game instead of just that specific matchup when it comes up in the sixth.  Now, instead of looking at Margot vs Larnach against that one lefty, he could look at Larnach facing a lefty now AND ALSO facing what would likely be a righty later in the game vs Margot facing a lefty now and the likely righty later, plus having less flexibility with a depleted bench.  The spreadsheets aren't the problem, the implementation of them is the problem.

A hammer is a useful tool when hanging a picture frame.  But if instead of finding a stud and using the hammer to pound a nail into it, I started cracking myself repeatedly in the temple with it, I would have an unhung picture frame and a massive headache.  Would you look at me and conclude that a hammer is not useful in hanging a picture frame?  Of course not.  You'd say "hey dummy, that's not how you use a hammer."  Misuse of a tool is the fault of the carpenter, not the tool

[Irishman]

I suppose Rocco will be with us for a minimum of a year.
Will Rocco make any changes to his staff? The following people should be let go: Pete Maki, the pitching coach; Luis Ramirez, the assistant pitching coach; Tommy Watkins, the third base coach; Jayce Tingler, the bench coach; Rudy Hernandez, the hitting coaches; and Derek Shomon, the assistant hitting coach. With other coaches, I believe I get along just fine.

Any thoughts? 

[lecroy24fan]

You didn't mention the one who is gone today, David Popkins.

[Gagnerules]

59 minutes ago, lecroy24fan said:

You didn't mention the one who is gone today, David Popkins.

Rocco should have been shown the exit yesterday.

[Richie the Rally Goat]

18 hours ago, Aerodeliria said:

When the Twins opted to select Rocco as their field boss, I was all in. I thought that someone who used spreadsheets and computers to help in decision-making was going to elevate the Twins to new heights. Certainly talent matters, but it seemed that Twins had a pretty good farm system supporting it, and a few important players added here and there meant that I thought the Twins were going places.

However, the past few years there have been many-a-complaint about Rocco not relying on his gut or not having instincts for the game. These complaints are often countered with just the opposite kind of complaints. He does not use enough analysis in key situations and so makes poor decisions. I contend here that they are simply opposite sides of the same coin and Rocco really does cling to his spreadsheets, but it can look like he is going with his gut when he really isn't.

The problem here IMHO is that Rocco does not understand statistical analysis as well as he thinks he does and so he is blind to things that he should be aware of. Let me explain before firing the poison arrows in my direction. (You can do it later.)

Let's first look at the woeful pinch-hitting of Margot. When Margot was 0-10 as a pinch-hitter, Baldelli was still looking at the percentages...Margot should pinch-hit here for Julien because the A's just brought in a LH pitcher and Margot hits LH pitchers at a much higher rate when compared to Julien. The spreadsheet showed that Margot had to be a much better option, so after Margot grounds out to short to go 0-11 as a pinch-hitter, it doesn't change anything in Rocco's mind. He's looking at it like flipping a penny 100 times...the odds are that 50% will be heads and 50% will be tails, but this is erroneous thinking....The problem for Baldelli is that it is not at all like flipping pennies because when flipping pennies there are no corrupting variables. In the case of Margot's pinch-hitting, one cannot simply say, "Well now he is bound to start collecting hits because the odds are really in his favor." Because we assume that Margot is a human, we can also assume that he has emotions. Baldelli does not consider emotional variables because we have no way to measure them during a game, so they are ignored by him. Because ALL teams now use these spreadsheets, this is where instincts become important. The good managers are still using the spreadsheets but they are also able to sense the ephemeral emotions that are at play. Is 0-10 affecting Margot's hitting? Yes! Why? Because he is probably feeling increasing pressure and is somewhat more anxious than at other times. We could measure his psychological state with a questionnaire and then run the data, but it would be impossible during a game and a player would likely lie anyway, so this is where in-game management is key. Letting Margot pinch-hit will not increase his odds as the data on paper would indicate because variables that not on the spreadsheet are taking their toll on Margot's hitting.

 

I'll use one more example. Zebby Matthews was pitching very well against Boston as the Twins clung to a 3-1. In the 5th, he gives up a two-out double bringing up the top of the order. This is the 3rd time through the order and the spreadsheet says that Boston should start ripping Matthews. However, Matthews is probably pitching his best game of the season up to that point. Jarren Duran comes up to the plate and Baldelli jogs out and signals immediately to bring in the LH pitcher Irvin to face the LH hitter Duran. There were many a-groan in Twinsville including my own. The spreadsheet shows that Irvin has a better chance of getting Duran out than Matthews, so for Baldelli, it's a no-brainer. Again, all of the psychological factors that are corrupting his data, he cannot see, such as Matthews' confidence being pretty high and Irvin's performances (likely due to increased anxiety) in such situations being pretty bad (after all that's why Baltimore jettisoned him in the first place). Because in the confines of a live game, these kinds of things cannot be measured, so in Baldelli's mind they simply don't exist. And, we are aware of the very costly meltdown that followed. The effect of being unable to sense emotional factors that are at play can be devastating. They can even have a boomerang effect. For example, the next time Matthews pitches in a tight game will he not have the smallest of nagging doubts that if it is the 3rd time through the order that he will automatically be yanked?

 

Psychological variables can be measured. We do it all the time in assessing students via questionnaires and interviews. Granted, it is much tougher for managers on the field to sense these things, but very good managers are able to make pretty good assessments. I remember when TK tried to take the ball from Jack Morris in the 10th inning against the Braves and Morris said, "I have one more inning in me. Let me pitch the 10th." TK just handed him the ball because he could sense that the confidence was real. Baldelli absolutely takes that ball and turns it over the relievers because that is what the data show.

Really good take and I very much agree with the gist. 

I would add that the roster wasn’t constructed well to support a pinch-hit heavy approach. The most relied upon hitters, Buxton, Correa, Lewis were the least reliable, and so the reserves spent most of their seasons as starters and then when the pinch hit situations occur, Rocco painted himself into a corner with Margot especially, but Martin and others that spent many innings out of position or just plain playing when they shouldn’t have.

What’s hard about all of this is knowing how the execution decisions are made.

So Baldelli kept going back to the same darn well, over and over, with worse hitters replacing bad hitters and the swirl continued.

How does the review/check happen to evaluate strategy/execution effectiveness during the season? Did no one see these decisions backfire over and over and raise their hand to question effectiveness?

[nicksaviking]

Agree, that Baldelli isn't really using analytics. His decisions may be from a data set, but that data set clearly doesn't ever update and it clearly isn't a deep dive into individual at bats, matchups or situations.

Young lefties bat X against left handed pitchers. Pinch runners are X amount more likely to score than the better hitter they are replacing. Young pitchers perform X against hitters the third time through the order. Manny Margot is a capable pinch hitter.

If he's managing via computer these are all surface level evaluations and they are clearly not accounting for current data (if it did, it would say Manny Margot should never pinch hit some time around June). And he's clearly not analyzing if maybe Matt Wallner SHOULD stay in and hit with X amount of outs, against this particular pitcher, in this particular inning, in this particular stadium, in this particular month, etc, etc.

[NotAboutWinning]

1 hour ago, nicksaviking said:

Agree, that Baldelli isn't really using analytics. His decisions may be from a data set, but that data set clearly doesn't ever update and it clearly isn't a deep dive into individual at bats, matchups or situations.

Young lefties bat X against left handed pitchers. Pinch runners are X amount more likely to score than the better hitter they are replacing. Young pitchers perform X against hitters the third time through the order. Manny Margot is a capable pinch hitter.

If he's managing via computer these are all surface level evaluations and they are clearly not accounting for current data (if it did, it would say Manny Margot should never pinch hit some time around June). And he's clearly not analyzing if maybe Matt Wallner SHOULD stay in and hit with X amount of outs, against this particular pitcher, in this particular inning, in this particular stadium, in this particular month, etc, etc.

That certainly seems to be the case. But then you hear that the Twins have an analytics staff of 100+ putting out daily analysis. Maybe Rocco's not getting the memos?

[nicksaviking]

6 minutes ago, NotAboutWinning said:

That certainly seems to be the case. But then you hear that the Twins have an analytics staff of 100+ putting out daily analysis. Maybe Rocco's not getting the memos?

I haven't heard either of those things. Just that they put together a more functional group of people who can use analytical data than they used to.

And even if they did, Baldelli isn't going to memorize 100 points of data every game. Like I said, most he's going to do is take the surface level info.

[Richie the Rally Goat]

46 minutes ago, NotAboutWinning said:

That certainly seems to be the case. But then you hear that the Twins have an analytics staff of 100+ putting out daily analysis. Maybe Rocco's not getting the memos?

there are 12 people in the baseball operations part of the org chart with “analyst” in their title.

https://www.mlb.com/twins/team/front-office

[Aerodeliria]

6 hours ago, Richie the Rally Goat said:

there are 12 people in the baseball operations part of the org chart with “analyst” in their title.

https://www.mlb.com/twins/team/front-office

That's fairly scary....

[USAFChief]

11 hours ago, Richie the Rally Goat said:

there are 12 people in the baseball operations part of the org chart with “analyst” in their title.

https://www.mlb.com/twins/team/front-office

General statement: Holy **** is that org chart bloated.

 

Also, it's impossible to tell how many actual baseball analysts there are. 

[nicksaviking]

1 hour ago, USAFChief said:

General statement: Holy **** is that org chart bloated.

 

Also, it's impossible to tell how many actual baseball analysts there are. 

I think I can solve their budget problem. I count 41 members of the 'Ticket Sales and Service' team and another seven under 'Ticket Operations'.

Those departments probably only need like one employee next season.

[ewen21]

16 hours ago, Gagnerules said:

Rocco should have been shown the exit yesterday.

I wanted him gone after the one and done in the 2019.  His “I’m not frustrated at all” comment basically sums up the atmosphere that’s been around here FOREVER.  
 

Any remember Michael Cuddyer labeling 2010 a “successful season” after being swept once again the first round?  It’s woven into the fabric.  I’m afraid ditching Rocco is the equivalent of moving the deck chairs on the Titanic.  We need to go wayyyy deeper.

[TheLeviathan]

16 hours ago, nicksaviking said:

Agree, that Baldelli isn't really using analytics. His decisions may be from a data set, but that data set clearly doesn't ever update and it clearly isn't a deep dive into individual at bats, matchups or situations.

Young lefties bat X against left handed pitchers. Pinch runners are X amount more likely to score than the better hitter they are replacing. Young pitchers perform X against hitters the third time through the order. Manny Margot is a capable pinch hitter.

If he's managing via computer these are all surface level evaluations and they are clearly not accounting for current data (if it did, it would say Manny Margot should never pinch hit some time around June). And he's clearly not analyzing if maybe Matt Wallner SHOULD stay in and hit with X amount of outs, against this particular pitcher, in this particular inning, in this particular stadium, in this particular month, etc, etc.

I heard Provus tell Barreiro that he asked Falvey about pinch hitting too early in games.  (His example was about removing Wallner in the 5th)  He stated that Falvey said they are "definitely" looking to change that approach.

[Craig Arko]

31 minutes ago, TheLeviathan said:

I heard Provus tell Barreiro that he asked Falvey about pinch hitting too early in games.  (His example was about removing Wallner in the 5th)  He stated that Falvey said they are "definitely" looking to change that approach.

Blessed be.

[Richie the Rally Goat]

9 hours ago, USAFChief said:

General statement: Holy **** is that org chart bloated.

 

Also, it's impossible to tell how many actual baseball analysts there are. 

342 people in all, but that includes the list of former GMs that I didn’t bother taking out

[USAFChief]

5 minutes ago, Richie the Rally Goat said:

342 people in all, but that includes the list of former GMs that I didn’t bother taking out

May be time for some "right sizing" of the overhead.

[Doctor Wu]

All things considered, I think Baldelli is a minor problem when if comes to what is wrong with the Twins. How many more wins would a "better" manager have given us this past season? A handful perhaps. And sure, those extra few wins might have put us back in the playoffs, but I don't think we would have gone past the first round. Then again, look at what the Tigers are doing. Baseball is a wonderful but weird game. I think most organizations would replace the manager after a season like the Twins just had, but that hasn't happened, so we roll with it. The biggest problem remains the health and productivity of the players on the team. Will a new manager give us those better results? Or do we need to change more pieces on the field and in the bullpen?