Kemp and the Quest for 1 bWAR

[thewupk]

I can't argue that....I want to say the correlation coefficient is like 0.8 and change. It's good but not amazing. And we don't have anything better...at least not publicly available.

And I don't watch a ton of games. I just watch Braves games. And I've seen Gattis and Klesko et al in the Braves LF. To my eye there is no way Kemp's D (which I've said is clearly terrible based on how he moves) come close to impacting how he has hit for us. I also don't think he'll hit this well going forward. I also expect him to only get worse moving around. And I hope we can dump him ASAP.

I think Statcast will push that number higher. I'm a big believer in it so far. And I would agree with your assessment on Kemp so far.

[thewupk]

You need to go back and look at some threads from last year then. In fact, I'm pretty sure you were one of them. I know others remember the "Kemp is a below average hitter" gibberish from last year

That would depend if someone was referring to his time in San Diego only or what he did in Atlanta. There is a pretty significant gap between the two.

[chop2chip]

Not sure. But UZR/WOBA/FIP accurately captured the value of the difference in runs scored and allowed in 2016.

Coincidence isn't science. I'm not denying there is a correlation, maybe even a strong correlation between UZR/DRS/etc. and true defensive value, but I can tell you without hesitation that every single team that accesses the Statcast database has the ability to derive a cleaner metric.

[thewupk]

Coincidence isn't science. I'm not denying there is a correlation, maybe even a strong correlation between UZR/DRS/etc. and true defensive value, but I can tell you without hesitation that every single team that accesses the Statcast database has the ability to derive a cleaner metric.

I don't disagree with that. But I think you can look at the roster moves a team makes and make a decision on whether that teams believes in defensive data by looking at that players DRS/UZR numbers.

[smootness]

I can't argue that....I want to say the correlation coefficient is like 0.8 and change. It's good but not amazing. And we don't have anything better...at least not publicly available.

And I don't watch a ton of games. I just watch Braves games. And I've seen Gattis and Klesko et al in the Braves LF. To my eye there is no way Kemp's D (which I've said is clearly terrible based on how he moves) come close to impacting how he has hit for us. I also don't think he'll hit this well going forward. I also expect him to only get worse moving around. And I hope we can dump him ASAP.

I'm honestly not sure what this means. Are you saying that there's no way Kemp's D comes close to negating his offense? Because that has been true this year. Has it been true consistently every year? No. And if you acknowledge that defense has value, then it has to impact his overall value.

[Southcack77]

Add up the Braves total WAR, then adjust for bad luck in 1 run games. It almost perfectly predicts the Braves record.

If defensive WAR was wrong, how could that be the case? If Kemp's WAR value is too low, which player's WAR value is too high?

So if you add up the WAR and then adjust for the reasons that it doesn't accurately predict the Win total it's perfectly predictive for one team.

Sweet.

[Southcack77]

The best assembled teams do have the most valuable players overall on it though. It's kind of the whole point. And if you are only looking at LF's then why should the positional adjustment even matter since all LF's get hit with that penalty.

we don't seem to be looking at only LFs. I totally agree that defense in LF might be very important in a playoff series and might be determinative. Clearly a team that has good offense and defense is LF is ahead of the team that only has one of those things. By enough to swing a series or alter the record signficiantly? Don't know, but still better.

[smootness]

we don't seem to be looking at only LFs. I totally agree that defense in LF might be very important in a playoff series and might be determinative. Clearly a team that has good offense and defense is LF is ahead of the team that only has one of those things. By enough to swing a series or alter the record signficiantly? Don't know, but still better.

Why would LF potentially determine a playoff series if it doesn't really matter in the regular season?

[Southcack77]

I don't think the reference was that teams necessarily care about it less....that may be true.

I think it's that teams value it less than publicly available war. Or that they have defensive metrics that they feel are more accurate. Or they feel positioning is more important than the player. Basically it's new enough that there isn't an industry consensus and the StatCast data is so new and massive that teams are constantly inovating. I think everyone values D. I think everyone knows it's paid less than offense and therefore there is a potential inefficiency there.

the Braves have pretty traditionally valued defense to some degree back in the days when they had great pitching.

What they have done in the rebuild and what they are doing now does not necessarily equal what their ideal roster would look like.

As far as long term planning, I guess the one example we have is Inciarte, who is primarily a defensive player.

[Southcack77]

I can't argue that....I want to say the correlation coefficient is like 0.8 and change. It's good but not amazing. And we don't have anything better...at least not publicly available.

And I don't watch a ton of games. I just watch Braves games. And I've seen Gattis and Klesko et al in the Braves LF. To my eye there is no way Kemp's D (which I've said is clearly terrible based on how he moves) come close to impacting how he has hit for us. I also don't think he'll hit this well going forward. I also expect him to only get worse moving around. And I hope we can dump him ASAP.

Yeah, that's pretty much what I think as well.

[thewupk]

So if you add up the WAR and then adjust for the reasons that it doesn't accurately predict the Win total it's perfectly predictive for one team.

Sweet.

WAR based standings are going to be very similar to a teams Pythag record which is based on runs scores vs runs allowed. Clearly the teams that score more runs then they allow will generally be better than those that do not. Obviously that will not always line up with actual wins and losses. The 2016 Rangers are the most recent example of this.

[Southcack77]

Why would LF potentially determine a playoff series if it doesn't really matter in the regular season?

Luck.

Theoretically a misplayed ball could cost a game, which could turn a series.

Consider Buckner's flub. Really it was mostly bad luck and over 162 games it would have been mostly irrelevant. But in that series it cost the Sox dearly.

[chop2chip]

I don't disagree with that. But I think you can look at the roster moves a team makes and make a decision on whether that teams believes in defensive data by looking at that players DRS/UZR numbers.

In most cases, yes. But when you have players who play to the extreme (either all bat or all defense), where a team is forced to make a trade-off, I think it muddies the water. Even the Cubs made the decision to stick Schwarber in LF everyday despite him being an obvious liability defensively (maybe even moreso than Kemp FWIW).

Matt Kemp is going to be an interesting test case if the Braves try to trade him this summer and he has an OPS above .900.

[thethe]

So if you add up the WAR and then adjust for the reasons that it doesn't accurately predict the Win total it's perfectly predictive for one team.

Sweet.

Brilliant!

[chop2chip]

What's ironic with this discussion is that the Braves are #6 in overall defensive value per Fangraphs.

[smootness]

Luck.

Theoretically a misplayed ball could cost a game, which could turn a series.

Consider Buckner's flub. Really it was mostly bad luck and over 162 games it would have been mostly irrelevant. But in that series it cost the Sox dearly.

If it can cost you games in the postseason, it can cost you games in the regular season.

[thethe]

If it can cost you games in the postseason, it can cost you games in the regular season.

1 game in the regular season is not as impact full as it is in the postseason. That's why an ace is more valuable than a mvp hitter in the playoffs

[thewupk]

In most cases, yes. But when you have players who play to the extreme (either all bat or all defense), where a team is forced to make a trade-off, I think it muddies the water. Even the Cubs made the decision to stick Schwarber in LF everyday despite him being an obvious liability defensively (maybe even moreso than Kemp FWIW).

Matt Kemp is going to be an interesting test case if the Braves try to trade him this summer and he has an OPS above .900.

In the end it does come out to making the best use of your roster that you can. Schwarber has a possible impact and I'm sure the Cubs would be fine with that defensive liability if he hits like he is projected too. Similar to a prime Adam Dunn type.

I think we already have a good idea for the market of sluggers who are bad on defense. Trumbo got around 12 million a year after smashing 47 bombs with a 123 WRC+ and is about on the same level defensively that Kemp is.

[thewupk]

What's ironic with this discussion is that the Braves are #6 in overall defensive value per Fangraphs.

Yeah. Having one of the the best defensive players in baseball goes a long way towards that.

[smootness]

1 game in the regular season is not as impact full as it is in the postseason. That's why an ace is more valuable than a mvp hitter in the playoffs

Everything is potentially more important in the postseason because there are fewer games, I get that. But if the argument is that it's mostly irrelevant in the regular season, then you're saying that it doesn't really have an impact over 162 games. Well, if something doesn't have an impact over 162 games then it is statistical noise.

And if it is statistical noise, then a bad defensive player, like Kemp, is no more likely to hurt you in one series than a good one. So if you're saying that a bad defensive player is more likely to hurt you in a small sample, then he is equally as likely to hurt you over the course of an entire season.