When Progression Strikes

by R.J. Anderson on Aug 17, 2008 11:30 AM EDT

Remember pre-trade deadline when we pondered whether a trade was needed given possible progressions by Cliff Floyd and Willy Aybar? Well, don't look now, but I think that's beginning to happen.

Cliff Floyd is now fourth on the team in OPS at .819. In his last 10 games he's got an on-base percentage over .500 (!) and a slugging of .600. He's walked eight times (although four came in one game) hit a homerun and has four doubles. The best part is his BABIP (.281) does not match the expected average of his 23.4% line drives.

In Willy Aybar's last 10 games he's got a .956 OPS, including a .600 slugging as well. Stop me if you've heard this before but his LD% is 19.1 and his BABIP is in the .230's, I think you can figure out where I'm going with that one.

Don't look now, but the Rays are in the midst of their best offensive month of the season, a direct equalizer to July which should be the worst.

Tweet Comment 49 comments | 0 recs |

Do you like this story?

Comments

Funny how when i used a small sample size with Byrd, you chastised me

sldo. i’m not a fan of BABIP because it skews the high K player (Upton) higher as far as BABIP is concerned

by sternfan1 on Aug 17, 2008 11:47 AM EDT reply actions

BABIP is just one piece of the puzzle

You can claim regular batting overrates low-power, low-walk players. (And it does.) But with additional information, it plays its part. For example, if a guy has a .250 AVG and .400 SLG percentage, but is traditionally a .300 hitter, you wouldn’t also criticize his SLG if his career average was .450 — he’s still showing as much power, and his SLG will rise as he begins to hit more singles.

A player might be getting lucky or unlucky with his BABIP (not that this is as directly measurable as many will claim) and as his BABIP adjusts going forward, his overall stat line will adjust. Nobody should claim that one player is better than another based solely on BABIP. Of course not striking out, taking walks, and hitting homeruns are needed to complete the picture.

my blog // calling BJ Upton lazy is lazy // past results do not guarantee future performance

by Sky Kalkman on Aug 17, 2008 12:09 PM EDT up reply actions

I'm speaking strictly for BABIP

if you take two players who have everything equal except one Ks 50x/yr and the other 100x, the higher K player has a higher BABIP

by sternfan1 on Aug 17, 2008 12:12 PM EDT up reply actions

but BABIP and K's are a CAUSE of larger-scale stats like AVG

If you claim two players have the same AVG but one strikes out more, then yes, that one has a higher BABIP. But you shouldn’t start with AVG. Start with K-rate and BABIP and then see how they combine for AVG. Just because a player has a higher K-rate doesn’t necessarily mean he has a higher BABIP. They often do, because it’s hard to stick in MLB if you strike out AND can’t hit the ball hard, but BABIP still tells you something worthwhile about a player.

my blog // calling BJ Upton lazy is lazy // past results do not guarantee future performance

by Sky Kalkman on Aug 17, 2008 12:17 PM EDT up reply actions

I said all else being equal, and i mean ALL else

the player with the higher K total has the higher BABIP

by sternfan1 on Aug 17, 2008 12:30 PM EDT up reply actions

...I don't think you understand what BABIP is.

"STP is me. He can do everything I can do." - R.J. Anderson

by P Brady on Aug 17, 2008 12:36 PM EDT up reply actions

true

but what’s your point?

I could say all else being equal, the player with more K’s hits more homeruns.

my blog // calling BJ Upton lazy is lazy // past results do not guarantee future performance

by Sky Kalkman on Aug 17, 2008 1:03 PM EDT up reply actions

Is it really true?

Why would it be? Doesn’t BABIP remove Ks from the equation (in principle of course).

The only relationship I could see is if you assume players that strike out more generally hit the ball harder when they make contact. That’s just an assumption and not always true.

by tallyray on Aug 17, 2008 1:05 PM EDT up reply actions

yes K's are removed from BABIP -- HRs too

if a player K’s more, that’s fewer balls in play. if he has the same BABIP as another guy, he has to make up that AVG and SLG difference somewhere, and it could be in HRs (since I’m defining everything but K’s and HR’s to be equal — that’s a dumb “what-if”, but it’s what sternfan was talking about .)

my blog // calling BJ Upton lazy is lazy // past results do not guarantee future performance

by Sky Kalkman on Aug 17, 2008 1:10 PM EDT up reply actions

but yes

even if all else isn’t equal, MLB hitters who K more do tend to hit more HRs.

in fact, there’s a positive correlation between K’s and overall offensive productivity. it’s a great example of why correlation isn’t causation.

my blog // calling BJ Upton lazy is lazy // past results do not guarantee future performance

by Sky Kalkman on Aug 17, 2008 1:11 PM EDT up reply actions

He said that high K = higher BABIP

and that wouldn’t be true. There could be a slight correlation but I doubt it is anything significant.

The major advantage of Ks is usually HRs. Usually. Since HRs are removed from BABIP that doesn’t influence the calculation.

As you said, K’s don’t equal production. I’d say HR hitters generally strike out more. HR’s generally result in higher productivity. That doesn’t mean K’s = productivity.

by tallyray on Aug 17, 2008 2:08 PM EDT up reply actions

yes, higher K = higher BABIP

if all else is equal. but it’s not important, because it doesn’t mean anything.

and, even if all else ISN’T equal, MLB players with high K’s do tend to have high BABIPs.

my blog // calling BJ Upton lazy is lazy // past results do not guarantee future performance

by Sky Kalkman on Aug 17, 2008 2:18 PM EDT up reply actions

Why does higher K = higher BABIP

That just doesn’t make sense to me.

by tallyray on Aug 17, 2008 2:19 PM EDT up reply actions

FWIW

Just took 2008 statistics from THT.

Using the 77 qualifying players from the AL I ran correlation between K/PA and BABIP and there was a coeffecient of .11. Stolen bases correlated at .27.

I know the caveats are the confounding variables and the sample size but I’m not sure how else to approach it. The next step would be to go through the last couple years and search for similar pairs. I’d just like to see why some think that a connection exists before taking it further.

To note, I’m not trying to prove anyone wrong, just wondering the train of thought that goes along with this.

by tallyray on Aug 17, 2008 2:31 PM EDT up reply actions

two cases

1. We assume “all else is equal”, meaning AVG, OBP, SLG, etc. If two players are equally productive but one strikes out more, he has to make up for that with positive contributions when he DOES make contact. Therefore he has to get more hits when he puts the ball in play.

Some numbers: Assume two players have a .300 AVG over 500 ABs (excluding HRs, which are defined to be equal). That means each player has to get 500*.300 = 150 hits. One strikes out 100 times and the other strikes out 200 times. The first has 400 balls in play to get 150 hits for a 150/400 = .375 BABIP. The second has 300 bals in play to get 150 hits for a 150/300 = .500 BABIP. The numbers aren’t reasonable, but make the point.

2. All else isn’t equal (reality): Majors league hitters have to do something productive — hit for average, take walks, hit for power, whatever. The more a player strikes out, the more he has to do when he doesn’t strike out in order to be any good. That means he has to either hit more homeruns or hit for a higher average on balls in play. If a hitter does not strike out much, he doesn’t have to have such a great BABIP or as much power, because every ball in play has a chance at a hit and he puts almost all balls in play.

That help?

my blog // calling BJ Upton lazy is lazy // past results do not guarantee future performance

by Sky Kalkman on Aug 17, 2008 2:41 PM EDT up reply actions

A little

As far as point 1 goes, I see what you mean. I wasn’t thinking of it terms like that. However, to counter, you have

Player A: 150 hits, 500 ABs = .300 Avg, 100 Ks, .375 BABIP
Player B: 150 hits, 500 ABs = .300 Avg, 200 Ks, .500 BABIP

You totally removed HRs from the equation. Player A hits no HRs. Player B only strikes out 150 times and hits 50 HRs. You’re at equal BABIP.

As far as point 2, there are too many assumptions that remove the central point.

The argument is that players with higher K’s have a higher BABIP and I can’t see where that is statically proven. There are too many factors and too many different types of players. A player does not have to make up for K’s with higher BA, not in today’s game. Players can get by with a lot of Ks and low BAs by doing other things right, like taking walks and hitting homeruns. There is little correlation between the two.

by tallyray on Aug 17, 2008 2:50 PM EDT up reply actions

on point one, we assumed that all else was equal, which would include HRs, right?

on point two (reality), i guess you’re right, although i’m confused intuitively. i took fangraphs data with a min of 50 PAs and 150 PAs and there’s little correlation between K-rate and BABIP. there’s a strong correlation between ISO and BABIP, though, which makes sense.

my guess is that there’s a correlation between AVG on balls in play including HRs and K’s.

my blog // calling BJ Upton lazy is lazy // past results do not guarantee future performance

by Sky Kalkman on Aug 17, 2008 3:02 PM EDT up reply actions

Quicky and dirty example.

Two similar players:

Grady Sizemore is hitting .265/.379/.513
Ian Kinsler is hitting .319/.376/.518

Sizemore is more Upton like, 78 walks and 95 Ks
Kinsler is more contact driven, 45 walks and 67 Ks

Sizemore has a BABIP of .280 and Kinsler has a BABIP of .335

by tallyray on Aug 17, 2008 1:04 PM EDT up reply actions

However BABIP shows if a player has been unlucky over that sample

Of course K’s will drive the AVG down, but a lower BABIP means that the rate stats are probably lower than they should be.

I do agree with sternfan that the sample is very small though Floyd’s improvement is pretty large.

by tallyray on Aug 17, 2008 12:23 PM EDT up reply actions

You always find a way to bring Upton into a thread, don't you?

by R.J. Anderson on Aug 17, 2008 1:12 PM EDT up reply actions

Expected BABIP and LD% over a small sample...

Wouldn’t LD% be almost pointless over such a small sample?

Also, where’d you get the split LD%’s?

by tallyray on Aug 17, 2008 12:26 PM EDT reply actions

I'm using the full season LD%/BABIP splits, not the last 10 games.

by R.J. Anderson on Aug 17, 2008 12:47 PM EDT up reply actions

From where?

Just curious since I figured that’s what you were using but HBT had different #’s.

by tallyray on Aug 17, 2008 12:59 PM EDT up reply actions

StatCorner.

by R.J. Anderson on Aug 17, 2008 12:59 PM EDT up reply actions

Odd.

Fangraphs: 19.1/.232 for Aybar 18.8/.305 for Floyd.
THT: 19.1/.232 for Aybar 18/.300 Floyd

by R.J. Anderson on Aug 17, 2008 1:01 PM EDT up reply actions

I changed the line to reflect on Aybar's THT/FG numbers, I'm not sure about Floyd.

by R.J. Anderson on Aug 17, 2008 1:03 PM EDT up reply actions

Floyd was the one I looked at and noticed the difference

Any idea why there is still such a spread between sites? That would significantly impair analysis.

by tallyray on Aug 17, 2008 1:09 PM EDT up reply actions

I believe SC pulls their data straight from the MLB play-by-play data.

I’m not sure if that makes it more accurate or not. I don’t know how Baseball Info Solutions gathers their data.

by R.J. Anderson on Aug 17, 2008 1:10 PM EDT up reply actions

everyone gathers the data by hand

you’ll just find differences between how different companies (and perhaps scorers within those companies) record batted balls. there’s some subjectivity involved differentiating between liners and flyballs. I believe one company even has “fliners”.

that’s why we need HITf/x

my blog // calling BJ Upton lazy is lazy // past results do not guarantee future performance

by Sky Kalkman on Aug 17, 2008 1:12 PM EDT up reply actions

Exactly

saying a ball traveled 38 degrees over the horizon at its highest point would be much more useful than saying it was a liner

by GomesSweetGomes on Aug 17, 2008 1:34 PM EDT up reply actions

Makes classification easier.

by R.J. Anderson on Aug 17, 2008 1:35 PM EDT up reply actions

funny how 1 week

can change a whole outlook. I guess it’s whatever the numbers say. Floyd is a bum one week and the next he is a very good player. amazing

by blazinrayz on Aug 17, 2008 1:39 PM EDT reply actions

I don't recall calling him a bum.

by R.J. Anderson on Aug 17, 2008 1:44 PM EDT up reply actions

if people thought floyd was a bum before this week, that's the problem

he’s a slightly above league-average hitter who hadn’t hit well so far in 2008.

and if people think Floyd’s going to SLG .600 the rest of the way, that’s another problem.

and if people think Gomes shouldn’t be expected to hit around his career average the rest of the way just because he hasn’t hit well so far this year, that’s another problem

my blog // calling BJ Upton lazy is lazy // past results do not guarantee future performance

by Sky Kalkman on Aug 17, 2008 1:45 PM EDT up reply actions

But there is more to it than just numbers, right?

I mean, we don’t know if Gomes has a toe or finger or quad muscle that is ever so slightly altering his swing and, consequently, productivity. Athletes are not robots, games are not played in a vaccum and nature does come into play. I guess if you said, “assuming Gomes is in the exact same physical condition as prior years he should hit around his career average,” that would seem more accurate to me.

These stats do give us some indication as to what “should” happen, do doubt. However, there are SO many variables that we cannot possibly account for that any real statistical prognostication seems iffy to me, at best.

by stpetelawyer on Aug 17, 2008 2:03 PM EDT up reply actions

yes, there's lots of other information you could include

like toe injuries. but do we have any of that information? no, and we often don’t. in those cases, and based on history, a weighted projection of past seasons will give you a better idea of what to expect going forward than just current season numbers. the important part is that this fact is based on what’s actually happened with real baseball players. it’s not a model, it’s empirical. it will miss high sometimes and it will miss low sometimes. but a weighted projection is a better guess than just using 2008 numbers.

the other problem with the whole injury/mechanics explanation is that it’s a circular logic to go like this: poor performance —> injury/mechanical issue —> poor performance.

my blog // calling BJ Upton lazy is lazy // past results do not guarantee future performance

by Sky Kalkman on Aug 17, 2008 2:21 PM EDT up reply actions

I've been calling for Floyd's head for a while

I don’t know if this changes anything.

When the rosters are at full capacity I believe Gross is more valuable than Floyd.

by tallyray on Aug 17, 2008 2:11 PM EDT up reply actions

yes, I agree that while Floyd's bat is still decent, he's an extra guy on the roster

with Crawford healthy, both Gross and Floyd aren’t needed, and Gross can do everything Floyd can do, plus more.

my blog // calling BJ Upton lazy is lazy // past results do not guarantee future performance

by Sky Kalkman on Aug 17, 2008 2:23 PM EDT up reply actions

Gross is > Floyd.

by R.J. Anderson on Aug 17, 2008 2:27 PM EDT up reply actions

Gross is lucky he's been clutch

Because the things he does well are the things that the average fan doesn’t notice.

by tallyray on Aug 17, 2008 2:32 PM EDT up reply actions

I'm a bit shocked nobody has called for Hinske's head.

Especially when his BA was in the .230-.240 range.

by R.J. Anderson on Aug 17, 2008 2:37 PM EDT up reply actions

well, he started off really strong, which tends to stick in people's heads the whole season

but i get your point. people aren’t exactly rational.

my blog // calling BJ Upton lazy is lazy // past results do not guarantee future performance

by Sky Kalkman on Aug 17, 2008 2:41 PM EDT up reply actions

Exactly what I was going to post...

Hinske did what players have to do now a days and that’s hit well in April. Pena is the exact opposite. Yet up until a week ago all you heard was how great Hinske has been and how much Pena was struggling.

Another thing with Hinske vs. Floyd. Hinske has been forced to face lefties, hitting .130/.245/.217 against them in 53 PAs and driving down his rate stats. Floyd has only had to face lefties 8 times so his overall line is basically the same as his righty split.

by tallyray on Aug 17, 2008 2:53 PM EDT up reply actions

Exactly...

People were really down on Floyd just a week or two ago. Every post was about how Floyd had to go or Floyd is worthless.

by save_the_trop on Aug 17, 2008 7:03 PM EDT up reply actions

OT: LOL 6-0 Blue Jays after half an inning...

Certainties in Life: Death, Taxes, and Gabe Gross coming through in the ninth

by JMB on Aug 17, 2008 2:05 PM EDT reply actions

with Beckett starting...

Certainties in Life: Death, Taxes, and Gabe Gross coming through in the ninth

by JMB on Aug 17, 2008 2:07 PM EDT up reply actions

As far as Floyd is concerned

He is what he is. He’s a DH that will run hot and cold, but what is more important to us is the experience he gives a young team. Right now, thankfully, he’s getting hot at a time we desperately need him to (w/o Longo ’til the end of the month and CC for at least the rest of the regular season).

Certainties in Life: Death, Taxes, and Gabe Gross coming through in the ninth

by JMB on Aug 17, 2008 2:15 PM EDT reply actions