Fire alex cora

[a700hitter]

Except that there are just certain biases we have inherently. Our brains can just be wired flat out weird.

 

Dave Cameron, formerly of Fangraphs, once wrote that the eye test for defense often just turns into "I saw that guy make an error once." This is just a simple confirmation bias that we are all prone to. We go into watching a game thinking Player X cannot field well. He makes a mistake. And on that sample size of one play, the observer feels assured with his initial hypothesis. He can't help it. None of us can...

And they are wired just as weird when compiling or interpreting data.

[a700hitter]

Here's a personal example: Jacoby Ellsbury.

 

When he first came up, I saw him make a few very nice plays. I knew he has speed, so I naturally thought he was a great defensive player. When he finally became the FT CF'er, a poster on that other site claimed he was not plus and had slow reaction times and took the wrong routes to balls hit to him. I started defending JE.

 

Then, I noticed his UZR/150 was negative, so I started watching every ball hit to him more closely. I did notice bad routes and slower breaks on the plays I was able to see on TV (not all that many). It's hard to know what my bias was before and after I started focusing on him more often.

 

To make the example even more complex, the next few years, I didn't notice the mistakes as much and went to check the numbers- sure enough, he had big plus numbers in 2010 and 2011 and was barely plus afterwards. The numbers seemed to support my observations that yes he was worse than I thought in 2009 but then did get better afterwards.

 

Was this just an example that supports my beliefs, so I use it, or not?

 

Judging just from TV isn't optimal. You see a lot more at the games. I have gone to games and had people who had watched on TV tell me about a great catch, but i knew that the player initially broke the wrong way. I don't think the home viewer was employing bias. He just didn't see on TV what I saw in person.

[notin]

And they are wired just as weird when compiling or interpreting data.

 

Actually, not necessarily.

 

It's times like that where many people actually confront their biases. In fact, it might be the best way to confront them...

[notin]

Judging just from TV isn't optimal. You see a lot more at the games. I have gone to games and had people who had watched on TV tell me about a great catch, but i knew that the player initially broke the wrong way. I don't think the home viewer was employing bias. He just didn't see on TV what I saw in person.

 

But it is funny the weird stereotypes our televisions created. for years, people insisted all left-handed pitchers threw across their bodies and got more break, because that was how it looked when the centerfield camera was positioned just over the pitcher's right shoulder...

[a700hitter]

Actually, not necessarily.

 

It's times like that where many people actually confront their biases. In fact, it might be the best way to confront them...

And there are people who can evaluate without bias without needing to be confronted with stats. And then there are those who use stats in a biased way, including every player’s agent. Moon said it well before, and he will correct me if I paraphrase inaccurately, neither side has a monopoly on bias and neither side is less likely to be biased .

Edited October 21, 2019 by a700hitter

[notin]

And there are people who can evaluate without bias without needing to be confronted with stats. And then there are those who use stats in a biased way, including every player’s agent. Moon said it well before, and he will correct me if I paraphrase inaccurately, neither side has a monopoly on bias and neither side is less likely to be biased .

 

At this point, you're repeating a lot of what I say about stats. The bias in stats typically comes from either error or intent. An agent uses intent.

 

But the biases we all have when we observe things or remember things are part of the human psyche. Stuff like Dunning-Kruger effect, confirmation bias, negativity bias, self-serving bias, optimism/pessimism bias, hindsight bias, decline bias, and of course, the ever popular Backfire Effect. That last one is when you try to prove someone else wrong on a subject, and their reaction is to take the correction as a personal attack and then only cling more deeply to their beliefs. We see that one daily on this board.

 

This is all psychology stuff, and admittedly I am not an expert on it, although the Dunning-Kruger effect wants to me to think I am...

[a700hitter]

At this point, you're repeating a lot of what I say about stats. The bias in stats typically comes from either error or intent. An agent uses intent.

 

But the biases we all have when we observe things or remember things are part of the human psyche. Stuff like Dunning-Kruger effect, confirmation bias, negativity bias, self-serving bias, optimism/pessimism bias, hindsight bias, decline bias, and of course, the ever popular Backfire Effect. That last one is when you try to prove someone else wrong on a subject, and their reaction is to take the correction as a personal attack and then only cling more deeply to their beliefs. We see that one daily on this board.

 

This is all psychology stuff, and admittedly I am not an expert on it, although the Dunning-Kruger effect wants to me to think I am...

I think that 95% of what is discussed on this board is opinion that is full of bias whether stats are relied upon or not. It’s opinion. People make the mistake of believing that there is a right or wrong opinions. That’s just horseshit. I’ll go no further than to say that there are strong arguments and weak arguments and everything in between. Asserting Right and wrong on a message board is folly.

[notin]

I think that 95% of what is discussed on this board is opinion that is full of bias whether stats are relied upon or not. It’s opinion. People make the mistake of believing that there is a right or wrong opinions. That’s just horseshit. I’ll go no further than to say that there are strong arguments and weak arguments and everything in between. Asserting Right and wrong on a message board is folly.

 

It's not abut assessing right or wrong as much as it is about approaching (not reaching, just approaching) objectivity.

 

And when it comes to the "human element" of baseball, all humans have flawed observational skills, pre-conceived notions, faulty memories, and can alter their own memories unconsciously. Data doesn't do that.

 

Stats capture the human element better. But that doesn't mean stat geeks do...

[S5Dewey]

It's not abut assessing right or wrong as much as it is about approaching (not reaching, just approaching) objectivity.

 

And when it comes to the "human element" of baseball, all humans have flawed observational skills, pre-conceived notions, faulty memories, and can alter their own memories unconsciously. Data doesn't do that.

 

Stats capture the human element better. But that doesn't mean stat geeks do...

 

soo.... you're saying that those folks who are rating every defensive play to determine defensive skills to calculate dWAR have all those traits, right?

[Kimmi]

Every single MLB player past and present 100% believe momentum, clutch, and chock exist.

Perception is reality. No matter what the stats say.

 

There is a such thing as momentum. It has no predictive value, however, in the outcome of the next series, the next game, the next inning, or even the next at bat.

[Kimmi]

This isn't a swipe at anyone here, but I find it mind-boggling that anyone who's ever played or coached sports at any level could believe that momentum, clutch, and 'chock' don't exist.

 

FTR, choking does exist. You just won't typically see at the major league level. Players who are chokers have been weeded out by that level, or if they make it, they don't last very long.

 

Clutch performances also exist.

[Kimmi]

We both have seen statheads look for some stat that supports their bias. It happens everyday. Even an introductory course on statistics will show you how to manipulate date. The argument that statheads don't employ bias is completely bogus.

 

^^ False statement.

 

The analytics guys are out looking for answers. They are not looking to prove anyone right or wrong. It does them no good to slant stats one way or another. Also, their work is heavily examined by all of the other stat geeks, and if their is something biased or bogus about it, it will not fly.

[Kimmi]

Absolutely true. And sometimes very easy to accomplish...

 

And at the same time, no one can provide stats against our arguments. So if stats are so easy to find to support our 'bias', where are the countering stats?

[notin]

soo.... you're saying that those folks who are rating every defensive play to determine defensive skills to calculate dWAR have all those traits, right?

 

No, because it’s all done by software now..,

[Kimmi]

Here's a personal example: Jacoby Ellsbury.

 

When he first came up, I saw him make a few very nice plays. I knew he has speed, so I naturally thought he was a great defensive player. When he finally became the FT CF'er, a poster on that other site claimed he was not plus and had slow reaction times and took the wrong routes to balls hit to him. I started defending JE.

 

Then, I noticed his UZR/150 was negative, so I started watching every ball hit to him more closely. I did notice bad routes and slower breaks on the plays I was able to see on TV (not all that many). It's hard to know what my bias was before and after I started focusing on him more often.

 

To make the example even more complex, the next few years, I didn't notice the mistakes as much and went to check the numbers- sure enough, he had big plus numbers in 2010 and 2011 and was barely plus afterwards. The numbers seemed to support my observations that yes he was worse than I thought in 2009 but then did get better afterwards.

 

Was this just an example that supports my beliefs, so I use it, or not?

 

 

Here's my example. I was a competitive athlete in high school and college, in a few different sports. I wholeheartedly believed in clutch, momentum, the hot hand, etc. Why would I purposefully look for stats that contradict everything I believed in?

 

IMO, I was unbiased and open-minded enough to let the stats alter my opinions, and they were strong opinions.

[notin]

And at the same time, no one can provide stats against our arguments. So if stats are so easy to find to support our 'bias', where are the countering stats?

 

 

Raw data is not biased.

 

But data analysis can be subject to a whole different set of cognitive biases, like skewed data bias, sample selection bias, cause-effect bias. And so on...

[notin]

There is a such thing as momentum. It has no predictive value, however, in the outcome of the next series, the next game, the next inning, or even the next at bat.

 

Correct. Momentum describes history. But then again, so does “win”..,

[Kimmi]

Raw data is not biased.

 

But data analysis can be subject to a whole different set of cognitive biases, like skewed data bias, sample selection bias, cause-effect bias. And so on...

 

I get that. But those biases will not fly among the stat geek community.

 

I'm still not seeing those analyses countering my arguments about things like momentum.

[a700hitter]

^^ False statement.

 

The analytics guys are out looking for answers. They are not looking to prove anyone right or wrong. It does them no good to slant stats one way or another. Also, their work is heavily examined by all of the other stat geeks, and if their is something biased or bogus about it, it will not fly.

You are bathing in bias, but can't see it.

[a700hitter]

It's not abut assessing right or wrong as much as it is about approaching (not reaching, just approaching) objectivity.

 

And when it comes to the "human element" of baseball, all humans have flawed observational skills, pre-conceived notions, faulty memories, and can alter their own memories unconsciously. Data doesn't do that.

 

Stats capture the human element better. But that doesn't mean stat geeks do...

I can agree with that.

[moonslav59]

Here's my example. I was a competitive athlete in high school and college, in a few different sports. I wholeheartedly believed in clutch, momentum, the hot hand, etc. Why would I purposefully look for stats that contradict everything I believed in?

 

IMO, I was unbiased and open-minded enough to let the stats alter my opinions, and they were strong opinions.

 

I have countless examples like that, too. Times where I gave my opinion only to find out I was wrong and the data either proved me wrong or all but proved me wrong.

[moonslav59]

Here's a scenario:

 

Poster A: John has been in such a bad slump!

Poster B: But he has an .875 OPS the last week and .85o the last month.

 

Poster A: But he never get hits with men on base.

Poster B: He has a .350 BA with men on base and .380 with RISP.

 

Poster A: But never when it matters.

Poster B: He has a .950 OPS when Late & Close and .925 in high leverage situations.

 

Is Poster B right, or is he a dick for using stats as the final argument ender?

[jacksonianmarch]

It's kinda both. Listen, there is time for allowing catharsis and then a time to rein it in

[jad]

Raw data is not biased.

 

But data analysis can be subject to a whole different set of cognitive biases, like skewed data bias, sample selection bias, cause-effect bias. And so on...

 

I agree to some extent. But I think I would go further and say that in baseball (as in any area of research) there is no such thing as 'raw data', insofar as you need to define it in order to analyze it and that very definition often skews or determines the results.

[Bellhorn04]

I agree to some extent. But I think I would go further and say that in baseball (as in any area of research) there is no such thing as 'raw data', insofar as you need to define it in order to analyze it and that very definition often skews or determines the results.

 

I'm getting a little lost here.

 

BA over the last 50 games .279

ERA in last 10 starts 4.12

 

Are those not examples of raw data?

[Slasher9]

the simple fact of collecting the raw data introduces bias

[Bellhorn04]

the simple fact of collecting the raw data introduces bias

 

But how can there can be bias collecting batting averages and ERA's?

 

Some of this stuff is purely objective.

 

And if it's 'cherry-picked', a reader with any baseball IQ can see that and respond accordingly.

Edited October 22, 2019 by Bellhorn04

[Slasher9]

there will be bias until:

- the 2 leagues play under identical rules (see: DH)

- 100% balanced schedules

- robot strike zone

[notin]

I'm getting a little lost here.

 

BA over the last 50 games .279

ERA in last 10 starts 4.12

 

Are those not examples of raw data?

 

They are raw data, but arguably subject to a sample selection bias, inasmuch as they ignore all data from previous games. But they are done for the purpose of showing a hot streak or short term improvement.

 

The question is, how representative are they?

[notin]

there will be bias until:

- the 2 leagues play under identical rules (see: DH)

- 100% balanced schedules

- robot strike zone

 

No.

 

There is a difference between a cognitive bias related to human processing of information and data simply collected in different sets.

 

But you are correct in tht these are factors that get ignored in comparing stats from different players. However, even with all this, the sets will never be equal because hitters cannot face pitchers from their own team (and vice versa) which gives advantages to the teammates of the best players....