SB Nation - Login for mobile commenting

Lookout Landing

Oh My God Oh My God Oh My God Oh My God Oh My God

I can't read this article because it's blocked at work, so thank goodness for Tango:

Cory also explained that we can expect the roll out of Hit-f/x, a system similar to Pitch-f/x that would use the technology already in place to track the initial batted ball data. Trajectory, angel [sic], velocity, etc. measurements would all be recorded but the technology would be limited to just the initial batted ball data. The Hit-f/x system would not be able to track the entire trajectory of batted balls ... I was told that [the Hit-f/x system] is definitely on the radar for the ‘09 season.

We all knew that HITf/x would get here eventually, but I don't think anyone in his wildest dreams ever imagined we'd see it in 2009, which seems to be what the article is suggesting (although if there's a clarification elsewhere in the link that I can't see, feel free to correct me). That just blows my mind in all the right ways.

Now, this isn't ultra-super HITf/x, in that there wouldn't be full trajectory measurements. So it would be difficult (albeit not impossible) to use this data for purposes of defensive evaluation. But for hitters and pitchers, this is exactly what we've always needed in order to complete the picture. Knowledge of a batted ball's characteristics off the bat - primarily velocity and angle - would render subjective classifications like "GB" and "LD" obsolete. It would take away the uncertainty and feed us nothing but cold, hard data. Which pitchers allow the hardest-hit balls in play? Which young hitters are taking a step forward? How much of "true" LD% is skill and how much is luck? What effects do certain park environments have on batted balls? HITf/x has the capability to answer some pressing and hugely important questions that we just haven't been able to tackle real effectively with the tools at our disposal.

These are exciting times we live in. Prepare yourselves for the glory.

Star-divide

A thought:

Presumably, one of the things that will eventually come out of HITf/x is a "theoretical" BA/OBP/SLG slash line, where hitters are rewarded and penalized for batted balls that either normally drop or are normally caught. This line will, I imagine, show a stronger correlation year-to-year than the actual BA/OBP/SLG lines do. I wonder, how will people choose to deal with screaming line drives hit directly at defenders?* Bad luck, or bad hitting? I'm guessing the latter, but is that really fair? I suppose a lot of work will have to be done in identifying how much control a batter has over the horizontal angle at which he puts the ball in play. God this is exciting. No, you're a dork.

* there's also the matter of hitting into a defensive shift. But now I'm getting ahead of myself.

0 recs  |  51 comments

Comments

I'm already working out the algoritm for tRA f/x

I suspect it will involve a pair of convolution integrals.

I thought you guys said that hit/fx would have been impossibly difficult to fund.
I meant full field hit f/x
So they will be tracking the initial seconds of the ascent phase?

And computing the data into some sort of poor mans hit/fx?

They'll have everything except the trajectory and where the ball landed

but then we already have estimates of the latter, which allow us to estimate the former.

So... you're saying this is good news?
HitFx is Wrong

because how can you know anything if you aren’t there to see them play?
Nerds.

The cameras are there.

The cameras know all.

Is that you, big brother?
EXACTLY!!

If a real person were there, they could tell you if the hitter hit with conviction, if he was displaying any fear and could determine how gritty the ball was off the bat, not to mention if the hitter showed leadership that could contribute to team chemistry as he ran the bases.

And whether the pitcher can play 3rd base.
PitchFx is also bullshit.

It doesn’t record how feared a pitcher is when they pitch to different players. How are we supposed to know who’s the most feared of their era? That’s the most important stat.

This is new....

In the past when I’d come to Lookout Landing and see a post title similar to this I’d think about all the negative things that must have just occured. Not anymore…weeeeeeeeeeeee

Agreed.

Before it would have been something horrible like “X traded for crap” or “Horrible signing even my mother knows is horrible”.
Now this… well, I guess it is… good?

AAAAAAAAAHHHHHHHHHHHHH
Word.

someone already posted this link in the Misc. thread, and it’s pretty awesome news, particularly that a beta version may be rolled out as early as this year. Defensive stats are about to get a bit more excellent.

It was Edgar for Prez
And pitching stats are about to be perfected.
How much will this improve pitching stats

Lets say W/L records allow us to estimate pitching performance to a 25% degree of accuracy.
ERA => 75%
FIP => 90%
xFIP => 93%
tRA => 95%
+Hit/Pitch f/x => 96%?

This at all correct? How bit of an improvement are we talking about? Each step we take is harder to improve on. Just wondering what the thinking is.

I imagine defensive stats will take the longest to adjust
Do you think batting stats need adjustment?

You hint at this in your ‘final thought’ and now that hit f/x is (maybe) not just a nerd fantasy the question needs to be seriously considered. Batters aren’t pitchers – they see all kinds of defenses – and luck, bad or good, can’t be sustained for an entire season. It’s not luck if it is. If you use hit f/x to create a t(wOBA) or whatever maybe Vidro’s line from 07 would look a bit more in line with his 08. But what would happen to Ichiro’s line?

I dunno. In terms of hitting I think any improvement on wOBA can only be incremental. However, having precise ball speed and angle off the bat data would make stats like PMR and UZR far more objective and maybe would not take all that long to implement.

It'll help tremendously in hitter projection.

Not sure how much it’ll help in hitter evaluation.

Very true

I hadn’t thought about projection. Anyway, I think the general consensus is yippee and whatever it’s applied to doesn’t matter so much.

Thanks for making this distinction between the two.

I feel stupid for struggling to come up with what seems like such a simple way of differentiating between the two for a really long time, and now I feel like a dolt because it sounds so simple that way. So from now on I will just plagiarize you.

Huh. This is unexpected.

And delectable.

Yay!

This LLemming will presume our fearless leader’s joy should be matched by my own, even though I have yet to delve into statistical analysis with any serious depth.

Oh…goody! Gum drops!

the holy grail

would be something like fieldf/x where the position of the fielders is measured at the time the pitch is thrown and again when the ball is fielded. Then you could really tell who is in good position (has to move less than their peers) who has the best range and so on. I imagine this is at least a few years off.

That would be the goal

When that day comes, defensive metrics will turn into cold hard stats that even Raul Ibanez can’t deflect.

This is great news.

Here’s a question: Is it worth thinking about spin on the ball when translating this data into theoretical slash lines (or tRA-like pitching metrics)? For instance – sometimes a LH batter will smoke a line drive to the opposite field that slices toward the foul line like mad. I doubt that hitting like this is a skill, and ultra-super Hit f/x would obviously solve this problem. But maybe hitters with certain swing types are more likely to put spin on the ball – which seems to be outside the scope of the initial speed/angle/trajectory data. Just wondering if anyone has thought about this…

There is some discussion of spin in the Inside The Book thread linked in the post
Oh yeah

From Mike Fast himself. Thanks.

Yes!

Finally, we will have enough data to irrevocably determine that Adrian Beltre does, indeed, need to aim better.

I hate it when I come home, and Beltre has peed all over the seat. >=(
Yeah, I know. It pisses me off.

He’s always claiming he didn’t do it, or is blaming the other roommate. If only we had TOILETf/x, we could settle this debate once and for all.

Will this eventually give us more pretty heat charts?

I love those.

pretty soon we're going to start animating the games in CG Graphics

just like Toy Story, Ice Age, Finding Nemo, etc. Who needs real baseball when you can use graphics to emulate the real thing?

In all seriousness, this is awesome. I wonder how long it will take until GM’s start utilizing this data.

Good GMs?

Like three weeks.

Dayton Moore? Uh…

This is so so so ridiculous
Once we've perfectly quantified everything

Will there be any reason to pay attention to baseball anymore? Will there be any real decisions to make in player evaluation? I’m all for this since the Mariners apparently aren’t oblvious to its existence, but once we can perfectly quantify everything I feel like player moves will just turn into a baseball mogul type situation

I think we're a long ways away from being anywhere close to having everything perfectly quantified.
This is a huge step forward though
Oh yeah.

But there are always going to be ways to gain a competitive advantage. Just because you have data doesn’t mean you know how to use it.

And humans have a particular way of tending to defy data-driven absolutes.
The humans who play the game I mean.
It will never be possible to 100% accurately project future performance
Once Health f/x comes out in 50 years
Time travel
Then we can stop worrying about all this stats stuff

and get back to watching baseball for the reason why we really love it.

Luck, the thing we try to weed out of stats while evaluating players, will always exist.

And it’s a big reason baseball is so awesome.

That's why nothing will ever replace the games

But on a player acquisition front moves will be pretty cut and dry if a single metric can be developed that accurately captures the entire abilities of a player

You must Login with your SB Nation account and be a member of Lookout Landing to post a comment.