THE BOOK--Playing The Percentages In Baseball

I also hate that phrase “whims of random variation” - as if variation is something positive that doesn’t exist unless it happens.  if true talent/performance is a distribution every pitch makes up the variation.

I agree completely.

FIP explains a great deal of the variance in ERA and I really don’t think there is any argument there. For various reasons, the variance that remains unexplained has been termed ‘luck’ when in actuality it’s simply modeling error. If we know more information, we can reduce the variance even further, which seems to be what the author is talking about. A lot of really smart people are already working on this. However, regardless of that, just knowing the pitcher’s walk, strike out, and home run rate is extremely powerful, which is really the take home message for DIPS theory.

I am not a big fan of that kind of verbose, grandiose, and pedantic writing. I used to try and write that way in high school and maybe in college in order to show everyone how smart I was and showcase my vocabulary. At some point some teacher or professor explained to me (apparently not enough, according to some) that writing, especially non-fiction writing, should be clear and concise.

And why this author feels the need to re-spell sabermetrics, I don’t know. I am pretty sure he knows the standard way to spell it…

At the core of the modern theory of projectiles is that they move around in frictionless worlds. At the core of the modern theory of friction is that it is a scalar function of velocity.

Each of these cores is known with certinaty never to be true. Even in outter space, solar winds are a force (that you almost never, ever need to account for if you work for NASA.)

You need to say a lot more before you have an interesting critique of modern physics.

I agree that this author is arguing a strawman, but I don’t agree that his writing is either verbose, grandiose, or pedantic, and I’m having a hard time seeing how it could be seen as such.

Besides that, concision is not a universal virtue in writing, even in non-fiction. As in all things, it depends on what’s being communicated.

”...but I don’t agree that his writing is either verbose, grandiose, or pedantic, and I’m having a hard time seeing how it could be seen as such.”

We differ on our opinions then…

I understand where the author is coming from as I have had the same struggles with the word “luck”, in that luck often implies no skill.

I think many use the “Rounders” view of luck, meaning that only those improbable (non-average) events are luck. So if something has a 2% chance of hapenning then when it doesn’t happen it wouldn’t be luck, but the probable outcome.

When we’re talking stats, the difficult aspect to accept is that every outcome is luck in some sense, due to a batter not being able to get a .287 hit or a .117 part of a HR in one PA. It’s either a hit or non-hit. When stats guys say luck, they are referring to the timing of the event. When non-stats guys hear luck, they’re thinking there’s no skill involved.

Stats guys would say that Freese’s HR was luck due to the timing. Non stats guys would say how can it be luck? He got a hittable pitch and absolutely crushed it.

Almost everything in baseball is luck due to the 0 or 1 aspect. But that doesn’t fit most people’s idea of luck. It’s a difficult concept to accept.

I like the creative spelling of sabermatrix. It’s creative, but it does paint the author as Neo, the only one that can see the matrix for what it is and destroy it and the mindless control it has over humans. It’s an incorrect situation, but it does serve the authors purpose.

Again if a pitcher throws a shutout with only 2 K’s and 3 hits allowed on 25 BIP and no HRA on 14 FB, stats would say the pitcher had a lucky day. That he should have allowed 8 hits and at least 1 HR on those FB’s. The non-stats guy would have difficulty with that under the idea that the pitcher wasn’t just throwing cockshots and hoping for the best.

This is not an easy idea for many to accept and it’s centered on how we perceive the word luck.

For example, if a batter that strikes a ball at 12-degrees with a BB velocity of 85mph, they may have a 78% chance of getting a hit. The only way many would view the situation as luck is if the batter did not get a hit, since getting a hit is the most probable outcome.

I do see how the criticism is extremelly frustrating to those that have sound understanding of numbers, since the criticism is inaccurate.

I thought I’d already written something here; maybe I didn’t get it submitted.
In short, I disagree that there are any events that have probabilities which aren’t 0 or 1. Certainly just having a binary outcome is not enough. But depending on what you’re talking about, we may never be able to predict things better than these proababilistic models, and certainly there are lots of things now where that’s the best estimate we have. This appears to be one of them.
But my big issue with guys writing the article like this one is that they point out problems without giving alternative predictions nyone can say that the model isn’t absolutely perfect, but that’s not very useful. At least you should give a suggestion of how to make better predicitions. Which is why I actually like the part where he’s analyzing mechanics near the end. But I would like to see some more solid predictions from someone being critical, and I’d like to see somethign that can be applied to a large pool of players, not just one.

Circle/7: right, exactly.

Wandering/8: my assumption is that I reject fate, as that’s the only way to ascertain that the event happening (1) or not happening (0) is not subject to luck.  If you think the event occurring is predetermined, but we don’t have god-like knowledge to know all the collisions that are about to take place that will cause that event to take place at that point in time, then what’s the point of talking about it?  You’ve already decided that we can’t know it without being god.

The assumption of luck is that the event is not predetermined.  That’s the basis of the discussion.  If you can’t accept that as the basis, then we can’t have the discussion.

Why is Luck the preferred word choice vs Randomness? Is there any implied difference? I’m not aware of any discipline that uses the word Luck.

There is no “preference” necessarily.  It’s random variation centered around a true mean.  That’s what it is.

If you think “random chance” conveys it better than “luck”, then fine.  If you want to offer a new word, such that the anti-luckinistas aren’t going to write 1000 words on the subject of “humans” and “computers”, then fine as well.

My experience is that the anti-luckinstas will always argue about something.

Related thread:

http://www.insidethebook.com/ee/index.php/site/comments/god_and_500/

No complaint with random variation around a distribution, just that this word “luck” seems vestigial to baseball (at least on the science side; plenty of luck watching a broadcast). Luck vs. skill is a “debate” of sorts, and I don’t have any help to offer the anti-luckinistas, but I don’t think I’ve ever read anyone on the “randomness vs skill” debate (well, except NN Taleb, who writes well); it seems that all disciplines are ok with the idea that events may distribute randomly around a skillset (or distribution, norm, true talent), but not as many are ok using the word luck in the same scenario. Not sure why, the word seems to have a different feel (at least to me). Not that changing the word will keep the naysayers from saying “nay,” but interesting to me that “luck” is the word in play.

Tango/9: I don’t think anything is predetermined (well, not in the way you mean it, anyway). I also don’t think it’s random. It’s influenced by choices, which aren’t random either, but also not predictable to large extent. Which is why I basically agree with your framework for the discussion.

There’s a true mean (known or estimated) after every choice is made. 

As a new choice is made, then the mean will change.

Even when the LAST choice is made, you STILL won’t get a mean of 0 or 1.  Therefore, since an event MUST result, and if the true mean is a fraction, then what else is going to explain whether it happened or not?

It’s either that the final choice made leads to a mean of 0 or 1, it’s predetermined fate, or we have random variation around that mean.

As an example of the “final choice”, is a free throw.  Once that ball leaves your fingertips, there is nothing left to influence the ball.  It’s a question of physics.  The event occurring or not (basket made or missed) is now a mean of 0 or 1.

But, how is that interesting at all?

The interest is while the ball is still in the player’s hand.  As long as the mean is not 0 or 1, then whether he makes the basket or not (AT THE POINT THAT YOU FREEZE TIME), is subject to randomness.

I must be missing something. Seems to me that your 3rd and 5th paragraphs are somewhat contradictory. Probably I’m just missing some nuance. At any rate, I agree with your 5th onward, with the caveat that things can still be interesting after the last decision, just not important to study for decision-making. Just like you can find a mystery novel interesting even if you’ve already read it, but it isn’t the same really, because there’s nothing to figure out.
As a physicist, I find physics interesting. But again, it’s on a different level altogether.
It’s largely semantic and philosophical. The baseball research is entirely concerned with things that happen while there’s still some way of using what you know. And that must be before the last choice is made.

Where I might disagree with you is on the choices themselves. Let’s say I drink coffee 70% of mornings, the rest, tea. On any given day, I may not be able to tell you why I picked tea that day. But that doesn’t mean it’s random. There is a reason, whether it can be explained by me or not.

If there’s a reason, then it’s been predetermined.  The chance of it happening is 1.

If the choice has been made, and the chance of the event occurring is 1, then it’s predetermined (at that point in time).

If a choice has been made, and the chance of the event occurring is NOT 1 (or 0), then you have random variation around that true mean.

Luck ONLY applies when the chance of something happening is not 0 and not 1.

For your coffee/tea example: when you make your choice, and if the conditions repeat themselves exactly (at the point you make your choice), and if the result is ALWAYS you choose tea, then it’s predetermined (at that point of your decision-making).  There’s no luck there, even if you can’t explain why you keep choosing tea, given identical conditions.

Agree. The last line of your post 18 is the point I’m trying to make, except you should say determined rather than predetermined. My philosophic point is that all things that happen are like this.
Now, this DOESN’T mean that the probability was 1 or 0 say 5 days beforehand, at least for anyone constrained by time. It was .7. Which also doesn’t make my subsequent choices random.