How fast should a fastball be?

[Redsauce]

This is an amazingly awesome study over at Hardball Times about fastball speeds. Got some really amazing stuff on what you could do with the Pitch f/x data.
I still get the feeling that MLB is just providing this data free until we get addicted, then they're going to screw everyone to the wall & charge massive amounts of cash just to look at it.
My venting is over, but go read this study. It's ri-****-donkulous.

[Alloutwar]

good stuff.

[HoustonGM]

I actually like that study other than it uses his own stat inside it (runs).

What's that have to do with anything? Is it bad to use a stat that is expressed in a familiar format - the run? He didn't just randomly select numbers. There's actual logic and fact behind it. It's not an exact number that measures exactly what someone did, like hits or home runs, but it is a very valuable estimator.

[Coach Owens]

I like my good friend Matthew Carruth's stuff better.

[FRENCHREDSOX]

Till fastball speed:accuracy is actually VIABLY used then this is all hypothetical

[HoustonGM]

My reasoning is because though you may feel it's valuable (and it may well be) he's trying to prove an idea. (Or more accurately - that his line of his reasoning is based on solid concepts.) The best way to do this is to use accepted 'foundations' of reasonings. It may not be any less valid the way he did it, just I dunno, people like foundations to be of time-tested and widely accepted tangible stats. It seems it took an otherwise subjective test, and made it a little biased although I don't that was his intention, especially since his conclusion matched conventional wisdom.

I guess it's ok because his intended audience is more sabr based and more willing to accept his run definition with less questions then the general baseball fan might.

Linear weights, or run values, are time-tested and very well accepted by anybody with any knowledge of baseball stats at all beyond batting average. The concept is mind-numbingly simple. Each batting outcome is assigned a number of average runs:

Anyway, batting runs measures a player's run production above that of the average batter, by assigning a run value to the various outcomes of a plate appearance—for example, a single is worth (on average) just under half a run, the value of a walk is around one-third of a run, an out is worth around negative .25 runs, and so on.

The averages come from the actual real-life averages. It's incredibly simple. See Batting Runs.

Of course, it gets more involved as you go deeper into linear weights, but the basic premise is ridiculously simple.

[ohms_law]

Too simple, if you ask me...

[Redsauce]

Ah...great...this has already degenerated into a flame war of "it's simple, if you don't get it, you're stupid"...
Can't we all just enjoy the article?

[HoustonGM]

Ah...great...this has already degenerated into a flame war of "it's simple, if you don't get it, you're stupid"...
Can't we all just enjoy the article?

Nobody's flamed anybody...

[HoustonGM]

Thing is, these types of studies don't really do anything for the average fan, whether or not they're based on more "accepted" methods. Like you said, they're more for the sabermetrically-minded, or... major league teams, where it can actually be put to good use, and major leagues teams should really understand the concepts of linear weights (although, I'm afraid, there's still a portion of teams that don't).

[AndrewOsborn]

Nah, I was just saying that it would be more widely accepted as a viable study if he used more commonly accepted results.

It can't be accepted as viable if it doesn't use already accepted methods? The ones that we "accept" now had to start somewhere, correct?

[HoustonGM]

It can't be accepted as viable if it doesn't use already accepted methods? The ones that we "accept" now had to start somewhere, correct?

It's really irrelevent because the study does use a method that is commonly accepted by anybody with any reasonable amount of baseball statistical knowledge. It's just not widely known by the mainstream.

And after all, statistical studies should use methods that are commonly accepted by statisticians, not by the fans, just like scientific studies should use methods that are commonly accepted by scientists, not by the common person.

[HoustonGM]

Well calling sabr a science is stretching it a little :P It's more of a fuzzy science

I didn't call it a science.

[HoustonGM]

Well you said just like, so I sort of assumed

It was an analogy. Scientific studies should use methods which are accepted by scientists. Likewise, statistical studies should use methods which are accepted by statisticians.

But, technically, sabermetrics is a science. Sabermetrics is statistical analysis, which is a mathmatical science.

[Clay Dreslough]

Ah...great...this has already degenerated into a flame war of "it's simple, if you don't get it, you're stupid"...
Can't we all just enjoy the article?

If you enjoy the article, you're stupid.