Bernstein Column

[IlliniKrush]

I listen to B&B pretty much everyday because I want my sports fix, and they do have good content, but even you have to admit that Bernsteins arrogance and dismissive nature towards anyone with a contradictory opinion has really skyrocketed in the last 5 or so years. You either call in and agree with him, or he screams at you and hangs up, then makes fun of you for the rest of the show for daring to have a different opinion. Even his coworkers have been calling him out on it.

 

The score puts through a lot of dumb callers for sure, but there are people with good things to say and he simply does not let them.

 

What Bernstein needs is to be rid of Boers, because he needs a strong voice on the other side that will keep him in check.

I agree, and I wish they'd filter through to put more "smart" fans on the air. They do get the idiots on there, and most of them deserve to be blasted. When B and B actually talk hard core sports, they talk it at a very high level. They know how to watch a game. Their highest level of sports talk is above anyone else in this town that is on the air. Although some of the best B and B is when they aren't talking sports, it's hilarious radio.

 

You're right in that not having Boers would probably keep him more in check, and he's probably gotten worse in the last 5 years.

 

I get his point on this article, and I think he was also emphasizing Dunn's season so far even though more people are jumping into the Rizzo stuff because there was more ink about it. His last 2 paragraphs are really what you need to see, just take a step back sometimes to evaluate what's going on, the numbers, and not jump to ignorant conclusions about it.

[StrangeSox]

"Producing runs" is a hugely complex thing which a lot of these stats try to tease out, and crude metrics like RBI or runs don't really tell that accurate of a story. Being on base (put another way, not making an out) is a huge component of effectively producing runs.

 

You're right that more advanced metrics were originally crafted by fans, but they were picked up by agents and then used by GMs because they give better and/or additional information. GMs wouldn't care a bit about some fancy nerd math numbers that Boras is using unless the GM (and his staff) felt that they were actually valuable. It'd be more accurate to say that they gained wider acceptance in baseball management because they allow teams to better understand the players and the game.

[Kyyle23]

I agree, and I wish they'd filter through to put more "smart" fans on the air. They do get the idiots on there, and most of them deserve to be blasted. When B and B actually talk hard core sports, they talk it at a very high level. They know how to watch a game. Their highest level of sports talk is above anyone else in this town that is on the air. Although some of the best B and B is when they aren't talking sports, it's hilarious radio.

 

You're right in that not having Boers would probably keep him more in check, and he's probably gotten worse in the last 5 years.

 

I get his point on this article, and I think he was also emphasizing Dunn's season so far even though more people are jumping into the Rizzo stuff because there was more ink about it. His last 2 paragraphs are really what you need to see, just take a step back sometimes to evaluate what's going on, the numbers, and not jump to ignorant conclusions about it.

 

I think the what I disagree with most about the article is that he is evaluating Rizzo and Abreu in the same small sample size, but not bringing Rizzos previous year into the conversation. Rizzo started off pretty hot last season IIRC, and fell off hardcore.

[Harry Chappas]

I agree, and I wish they'd filter through to put more "smart" fans on the air. They do get the idiots on there, and most of them deserve to be blasted. When B and B actually talk hard core sports, they talk it at a very high level. They know how to watch a game.

 

How many really intelligent people are calling into radio talk shows though?

 

Dan Bernstein is obsessed with telling people what they have to do and how they have to watch when it comes to sports.

 

I like watching college basketball more than pro as well as college football over pro. In the world of Dan Bernstein that makes me a moron because pro is inherently better than college.

 

In reality I like watching my kids succeed in sports more than I do any athlete on the plant.

 

 

 

 

[ptatc]

Linear weights-based stats exist for the purpose of trying to compare players with a common denominator. For example, Brandon Phillips had 103 RBI in 2013. Mike Trout had 97. Does that make them the same class of hitter? If not, then how can we tell who is better and, by extension, likely to produce more in the future?

 

Well, Brandon Phillips had a wRC+ of 91 that year, and Trout had an insano-pants 176. So there you go.

 

Anyway, that's the purpose. Obviously things like RBI are critical when the game is being played, it's how you win, after all. I'd never want a coach to tell a player not to worry about driving in runs, for example. They just aren't good for evaluating ability. Or, better put, there are many other numbers that serve as much, much better proxies for evaluating ability than RBI and other context-based metrics.

This is true. However a few points. I'm not sure that saying someone who was better this year will be better next year. You will need a pattern of 5 years or so to determine this and I bet I could look at the "basic" stats and tell this as well. Even with those numbers did you really think that Phillips was better than Trout. I don't need any numbers to tell me that. Also if you use the runs scored +RBI -HR formula Trout winds up with 179 and Phillips 165, so even that shows Trout had a better year.

Also, in the dawning age of non- (or decreased) PED usage, not all players will be able to put up great numbers across the board. So I think players will need to be separated out. It's kind of like comparing a WR to an RB in football. Players are going to sort themselves into the OBP specialists and others in the SLG groups and there will only be the select few that can do both and this will be obvious.

[Buehrle>Wood]

I agree, and I wish they'd filter through to put more "smart" fans on the air. They do get the idiots on there, and most of them deserve to be blasted. When B and B actually talk hard core sports, they talk it at a very high level. They know how to watch a game. Their highest level of sports talk is above anyone else in this town that is on the air.

 

 

 

Terry Boers should never be included on any discussion of high level sports discussion. The guy is a lapdog with no actual opinion of his own. That's from someone who listens everyday. There isn't high level discussion because there is 0 discourse on the show.

[ptatc]

"Producing runs" is a hugely complex thing which a lot of these stats try to tease out, and crude metrics like RBI or runs don't really tell that accurate of a story. Being on base (put another way, not making an out) is a huge component of effectively producing runs.

 

You're right that more advanced metrics were originally crafted by fans, but they were picked up by agents and then used by GMs because they give better and/or additional information. GMs wouldn't care a bit about some fancy nerd math numbers that Boras is using unless the GM (and his staff) felt that they were actually valuable. It'd be more accurate to say that they gained wider acceptance in baseball management because they allow teams to better understand the players and the game.

It doesn't need to be that complex. How many did the player account for. Why use other stats to predict it. Of course things like being on base will influence it but it's not as important as actually doing it. Instead of looking at how factors over 5 years predict it, look at 5 years of actually doing it.

 

I'm not saying that the advanced stats shouldn't be used. I use them and watch them all the time. I was one of the first to promote them. I just think it's gone so far and using them like Bernstein did is an abuse of them to make a point.

[chw42]

Correct. However, that is the game of baseball. You cannot control everything to win the game. So why make up a stat that doesn't use it. It came about so agents can get more money for their client because "he did what he could control" nothing else really matters.

 

That is not why sabermetrics exist today, not even close. Weeding out dependent variables in the equation is behind most statistics in general. Because it allows us to evaluate current production and predict future production. There's nothing wrong with trying to objectively look at a player's performance. That has nothing to do with sports agents.

 

Just because you can't weed out every dependent variable doesn't mean you shouldn't try to. That's such a complacent attitude. Stats like wOBA have a much higher correlation to player value and talent than RBI and Runs. Why? Because it's less dependent on what other players do. Why wouldn't you try to isolate player performance? Baseball is an extremely individual sport.

Edited May 7, 2014 by chw42

[Princess Dye]

Terry Boers should never be included on any discussion of high level sports discussion. The guy is a lapdog with no actual opinion of his own. That's from someone who listens everyday. There isn't high level discussion because there is 0 discourse on the show.

 

He's slowing down from his heyday which is natural. Dude is at the end of his career. When he was younger he was the king fire stoker

[chw42]

This is true. However a few points. I'm not sure that saying someone who was better this year will be better next year. You will need a pattern of 5 years or so to determine this and I bet I could look at the "basic" stats and tell this as well. Even with those numbers did you really think that Phillips was better than Trout. I don't need any numbers to tell me that. Also if you use the runs scored +RBI -HR formula Trout winds up with 179 and Phillips 165, so even that shows Trout had a better year.

Also, in the dawning age of non- (or decreased) PED usage, not all players will be able to put up great numbers across the board. So I think players will need to be separated out. It's kind of like comparing a WR to an RB in football. Players are going to sort themselves into the OBP specialists and others in the SLG groups and there will only be the select few that can do both and this will be obvious.

 

The only problem is that Trout was better by a lot and it's not even close.

[StrangeSox]

It doesn't need to be that complex. How many did the player account for. Why use other stats to predict it. Of course things like being on base will influence it but it's not as important as actually doing it. Instead of looking at how factors over 5 years predict it, look at 5 years of actually doing it.

 

How much run production an individual player is accountable for isn't that straight forward--that's why you need more complex understanding. Relying on RBIs or Runs is going to lead to all sorts of silly conclusions about how two different players compare with each other. An individual batter can't force the guys ahead of him or the guys behind him to get on base or get a hit, so it makes no sense to use metrics that heavily rely on things outside of the individual player's control. If I hit a triple every single time I'm at bat but there's never anybody on ahead of me and everyone strikes out behind me, I'm going to have 0 RBIs and 0 runs. If you're a GM looking to evaluate signing me to a contract extension for the next year, you'd be making a horrible decision by relying on either of those two stats.

 

 

 

[CyAcosta41]

For my money, Dan Bernstein (AND Terry Boers) are Exhibits A and B proving that human beings can have such different tastes.

 

When I'm in my car and I accidentally switch to the Score during their show, I trash what's left of my rotator cuff attempting to change the station (to ANY station) before another word assaults my ears. Count me in with those who think they're horrible. The premise of their show (as best as I can tell) is that it offers a unique combination of sports acumen, entertainment, and humor. I don't see any of those things. Not a one; not at all. Then, go one step more and witness the smarmy and often offensive way they abuse their loyal minions and the occasional jamoke who calls in without being in on the joke ... and wwww. I need a shower. It's like walking into a circus freak show.

 

Clearly they've survived because somehow (albeit inexplicably to me) they bring-in the numbers. As I said, we all have different tastes. My dislike for this gruesome twosome is pretty intense. I've gone so far as to boycott their advertisers just because I feel better knowing I'm not feeding the beast.

Edited May 7, 2014 by CyAcosta41

[ptatc]

How much run production an individual player is accountable for isn't that straight forward--that's why you need more complex understanding. Relying on RBIs or Runs is going to lead to all sorts of silly conclusions about how two different players compare with each other. An individual batter can't force the guys ahead of him or the guys behind him to get on base or get a hit, so it makes no sense to use metrics that heavily rely on things outside of the individual player's control. If I hit a triple every single time I'm at bat but there's never anybody on ahead of me and everyone strikes out behind me, I'm going to have 0 RBIs and 0 runs. If you're a GM looking to evaluate signing me to a contract extension for the next year, you'd be making a horrible decision by relying on either of those two stats.

True as I stated earlier. However, this is the way the game is played. You MUST rely on other players in the lineup. Even the stats can't isolate everything. Your example is an extreme case that will not happen in a game. The more likely is that some of the players will strikeout and sometimes there will be people on base ahead of you. You cannot predict this or isolate it with any of the stats. So both of them have some predictive factor. So the advanced stats really didn't tell me much different.

In the example someone used earlier where Phillips looked better than Trout, a GM would be stupid to offer Phillips more than Trout. However we knew that before the numbers anyway.

[ptatc]

The only problem is that Trout was better by a lot and it's not even close.

Correct. But we knew that by watching each of the players, so the numbers really didn't help either way.

[chitownsportsfan]

True as I stated earlier. However, this is the way the game is played. You MUST rely on other players in the lineup. Even the stats can't isolate everything. Your example is an extreme case that will not happen in a game. The more likely is that some of the players will strikeout and sometimes there will be people on base ahead of you. You cannot predict this or isolate it with any of the stats. So both of them have some predictive factor. So the advanced stats really didn't tell me much different.

In the example someone used earlier where Phillips looked better than Trout, a GM would be stupid to offer Phillips more than Trout. However we knew that before the numbers anyway.

 

I can't remember where I saw it now, but there was a saber study that looked at super high ISOs like, players that had ISO above .280 and I believe the conclusion was that wRC+ underestimates the value of players with extreme power, such as

Abreu.

 

Frankly I don't care if his OBP is around 330 in his time with the Sox, yea it would be great if it was higher, but I will take 250/330/530 every single year he's with the Sox and not think twice.

 

UZR likes his defense so far as well, that's a bonus.

[StrangeSox]

True as I stated earlier. However, this is the way the game is played. You MUST rely on other players in the lineup. Even the stats can't isolate everything. Your example is an extreme case that will not happen in a game. The more likely is that some of the players will strikeout and sometimes there will be people on base ahead of you. You cannot predict this or isolate it with any of the stats. So both of them have some predictive factor. So the advanced stats really didn't tell me much different.

In the example someone used earlier where Phillips looked better than Trout, a GM would be stupid to offer Phillips more than Trout. However we knew that before the numbers anyway.

 

A baseball team must rely on the entire team to win. When evaluating talent at the individual level, it's counter-productive to rely on data with a whole bunch of dependent variables. Yes, my example is an extreme case, but I'm using it to highlight just how useless of a stat RBI and Runs can be in evaluating talent and predicting future performance.

[ptatc]

For my money, Dan Bernstein (AND Terry Boers) are Exhibits A and B proving that human beings can have such different tastes.

 

When I'm in my car and I accidentally switch to the Score during their show, I trash what's left of my rotator cuff attempting to change the station (to ANY station) before another word assaults my ears. Count me in with those who think they're horrible. The premise of their show (as best as I can tell) is that it offers a unique combination of sports acumen, entertainment, and humor. I don't see any of those things. Not a one; not at all. Then, go one step more and witness the smarmy and often offensive way they abuse their loyal minions and the occasional jamoke who calls in without being in on the joke ... and wwww. I need a shower. It's like walking into a circus freak show.

 

Clearly they've survived because somehow (albeit inexplicably to me) they bring-in the numbers. As I said, we all have different tastes. My dislike for this gruesome twosome is pretty intense. I've gone so far as to boycott their advertisers just because I feel better knowing I'm not feeding the beast.

The worst part is, Dan used to be fun. When he was starting out at the score (small s) he had a Saturday morning slot. He had a skit where he tried to build sports teams from the old Saturday morning cartoons. The kangaroo from the Sylvester and tweety show was the kicker. The grown up Bam-Bam was the tight end etc. He actually has a good sense of humor but now he just wants to put everyone down and proclaim his superiority and it got old quick.

[chw42]

Correct. But we knew that by watching each of the players, so the numbers really didn't help either way.

 

Except the advanced metrics also showed that Trout was a lot better than Philips...

 

[ptatc]

A baseball team must rely on the entire team to win. When evaluating talent at the individual level, it's counter-productive to rely on data with a whole bunch of dependent variables. Yes, my example is an extreme case, but I'm using it to highlight just how useless of a stat RBI and Runs can be in evaluating talent and predicting future performance.

I understand, I just disagree that they are so much more accurate as it's the true number of runs scored and given up that creates a win or loss.

 

Again, I'm not saying it isn't useful. It's just that I think they are abused and used to try to make poor comparisons like Bernstein did.

[StrangeSox]

Correct. But we knew that by watching each of the players, so the numbers really didn't help either way.

 

Right, you don't exactly need a bunch of stats to see that Trout is better. It was brought up because it is so obvious that Trout is better and more valuable, but the crude RBI/runs stats don't indicate that at all. They're a very poor measure of talent/ability and a poor predictor of future individual performance. Advanced analysis isn't very important when comparing across big gaps in talent, but it is very important when comparing otherwise similar players or finding players who are undervalued by traditional/crude stats.

[ptatc]

Except the advanced metrics also showed that Trout was a lot better than Philips...

But again, we knew that already. It should be used more for comparing apples to apples such as Trout to Miggy. Then again, GMs will offer each a boatload of money.

[chw42]

I understand, I just disagree that they are so much more accurate as it's the true number of runs scored and given up that creates a win or loss.

 

Again, I'm not saying it isn't useful. It's just that I think they are abused and used to try to make poor comparisons like Bernstein did.

 

I think you're getting scapegoating the wrong thing here.

 

Bernstein's article was flawed because the sample size he used was too small and he was making judgements based on those sample sizes. As Dick Allen mentioned before, Rizzo had a .797 OPS two weeks ago. Now it's up around .900. A lot can change in two weeks this early in the season.

 

The issue with Bernstein's article isn't that he used wOBA or wRC+ to make his case. The problem was that he exploited sample size to make his case.

[Eminor3rd]

This is true. However a few points. I'm not sure that saying someone who was better this year will be better next year. You will need a pattern of 5 years or so to determine this and I bet I could look at the "basic" stats and tell this as well. Even with those numbers did you really think that Phillips was better than Trout. I don't need any numbers to tell me that. Also if you use the runs scored +RBI -HR formula Trout winds up with 179 and Phillips 165, so even that shows Trout had a better year.

Also, in the dawning age of non- (or decreased) PED usage, not all players will be able to put up great numbers across the board. So I think players will need to be separated out. It's kind of like comparing a WR to an RB in football. Players are going to sort themselves into the OBP specialists and others in the SLG groups and there will only be the select few that can do both and this will be obvious.

 

Regarding the first sentence bolded: yes, you're right, past performance only goes so far in predicting future performance. However, the closer said performance can be tied to matters of skill rather than matters of context, the more likely that performance is to be repeatable.

 

Regarding the second sentence bolded: I don't think that's true. In fact, one of the big reasons these stats are chosen for this purpose is because they are more stable and predictive than anything else. For example, someone didn't decide that the elements that go into FIP intuitively make sense there, rather Voros McCracken did a bunch of studies to determine which components of pitching are most consistent year-to-year and are most highly correlated with success.

 

EDIT: I think the central theme is that it doesn't come down to "which stat is the best stat," but rather "which stat best answers the question at hand." When that question comes down to comparing players across context or quantifying degrees of contribution, linear weights are the way to go (fWAR, wOBA, et al.). When it comes down to who made the most important play in a game or game situation, traditional context-dependent stats (typically traditional ones) are the only tool for the job.

Edited May 7, 2014 by Eminor3rd

[shakes]

I can't remember where I saw it now, but there was a saber study that looked at super high ISOs like, players that had ISO above .280 and I believe the conclusion was that wRC+ underestimates the value of players with extreme power, such as

Abreu.

 

Frankly I don't care if his OBP is around 330 in his time with the Sox, yea it would be great if it was higher, but I will take 250/330/530 every single year he's with the Sox and not think twice.

 

UZR likes his defense so far as well, that's a bonus.

 

 

I think that is an important point regarding his ISO. I just looked it up and Abreu has a .350 ISO and only a .253 BAPIP. Rizzo has a .200 ISO and .306 BAPIP. That's 22-9 in extra base hits between the two of them.

 

I think Abreu's average and BB% will both rise, which means the ISO is unlikely to remain that high. I honestly think the early power has made Abreu more aggressive than he normally would be. He has very good command of the strikezone, but you can tell when he's up with RISP he just decides he will swing at anything. All part of the adjustment process.

 

Edit: And to add to ptatc point earlier, I think situational hitting doesn't get accounted for well enough in many advanced metrics. Some hitters have very different approaches with runners on and I would greatly prefer a guy like Abreu being a bit more aggressive in those situations. Continuing with Rizzo as an example, he has a great OBP, but with a lineup as bad as the Cubs, you need a guy like him to be more aggressive in some situations and try and get some runs on the board, when it is very likely the hitters behind him will struggle to do that. And pitchers know that.

 

 

Edited May 7, 2014 by shakes

[Eminor3rd]

Correct. But we knew that by watching each of the players, so the numbers really didn't help either way.

 

Maybe YOU did, but the RBI total lied to you about it, and many mainstream media pundits and fans were calling for a huge contract extension for Brandon Phillips now that he had "evolved his game to become a run producer." Further, Phillips himself famously lambasted fan pressure to improve his game citing that his RBI totals spoke for themselves, as if better productivity wouldn't lead to more RBIs as a by-product.

 

I guess what I'm saying is that if you have to ignore a stat in certain situations because you "just know better," what use does that stat have? If it's right except when it's wrong, and you already know when it's right or wrong, you really don't need the stat. It's not telling you anything in terms of player evaluation.