Matt Kemp

[smootness]

Same line of thinking when we talked about a guy with a reputation for having a good arm vs a guy that actually has a good arm. While both guys may produce equal value with their arms, the guy with the actual good arm as measured by statcast would be the guy to invest in long term.

This is where, IMO, it just depends on how you interpret and use the data. Sure, all things being equal, of course I'd rather have the guy who actually has a good arm rather than the guy getting by simply on the reputation of having a good arm. But if teams are valuing the player's arm as though it is not good, yet base runners are still not running on him, then you may find value in the market there. Of course, in this specific example, if teams now know his arm isn't as good, that will change the way runners react, as we discussed before.

This specific example doesn't really play out, but it's an example of how the data can be used in two different directions depending on the market and what other teams are valuing. If you can find a gap where other teams seem to be missing something in the data, you can use it to your advantage in either direction.

[thethe]

I agree. I prefer to track contributions. Whether by luck or skill it still happened.

But then that catch you referenced in a blowout is meaningless.

[nsacpi]

Why would this not be productive?

number of rainy days is likely to be a wash (pun intended) across players....ditto for the other variables you seem interested in observing

[nsacpi]

But then that catch you referenced in a blowout is meaningless.

It isn't meaningless...it gives some information on the players ability to make similar catches when the game is on the line

[thewupk]

But then that catch you referenced in a blowout is meaningless.

Is it? Should we look at all hitters and take away their hits in blowouts since they don't matter? Those stats do exist but I don't see anyone using those to determine how good a player is.

[thethe]

Is it? Should we look at all hitters and take away their hits in blowouts since they don't matter? Those stats do exist but I don't see anyone using those to determine how good a player is.

But it goes back to ability vs. Value. For ability it matters. Value it doesn't. I love stats. I love math. It's the assignment of value to when you get in sketchy territory when you introduce too many variables.

[nsacpi]

But it goes back to ability vs. Value. For ability it matters. Value it doesn't. I love stats. I love math. It's the assignment of value to when you get in sketchy territory when you introduce too many variables.

a player can't control whether a ball is hit to them in a key situation...but we know the distribution between key and unkey is random and evens out...maybe one or even two seasons is not long enough to even out...but I think with someone like Heyward sample size of defensive opportunities is no longer an issue

[thewupk]

But it goes back to ability vs. Value. For ability it matters. Value it doesn't. I love stats. I love math. It's the assignment of value to when you get in sketchy territory when you introduce too many variables.

You can't control sequencing. A team can have 10 hits and 1 run or 3 hits and 5 runs. Each hit should be valued the same regardless of the events before or after it. WAR is designed to show us how well a player did in a given year and not what their actual ability is.

[thethe]

a player can't control whether a ball is hit to them in a key situation...but we know the distribution between key and unkey is random and evens out...maybe one or even two seasons is not long enough to even out...but I think with someone like Heyward sample size of defensive opportunities is no longer an issue

I've never argued against heywaRd defensively. To me he was just as impact full as druw derensively.

[thethe]

But an outcome in a determined game is meaningless to value.

[thewupk]

But an outcome in a determined game is meaningless to value.

It isn't meaningless when that value is based on runs saved and created.

[thethe]

It isn't meaningless when that value is based on runs saved and created.

A run saved in a 5+ run game in the 9th inning has virtually 0 impact to that team's win loss expectancy. Therefore, it should have no impact when you are valuing how much a player contributes to wins and losses in a season.

Now if you want to get into his range and other defensive metrics that fine and I agree. But when converting those values to a win loss impact then these things should be considered.

[nsacpi]

A run saved in a 5+ run game in the 9th inning has virtually 0 impact to that team's win loss expectancy. Therefore, it should have no impact when you are valuing how much a player contributes to wins and losses in a season.

Now if you want to get into his range and other defensive metrics that fine and I agree. But when converting those values to a win loss impact then these things should be considered.

you can make situational adjustments if you want, but it won't affect estimates of a players value as the sample grows....assuming the situations in which a player has an at bat are random

[thethe]

you can make situational adjustments if you want, but it won't affect estimates of a players value as the sample grows....assuming the situations in which a player has an at bat are random

Agreed but when you are talking about defensive chances it may take 3-5 seasons until that happens. So value over a shorter period of time is skewed.

[thewupk]

Agreed but when you are talking about defensive chances it may take 3-5 seasons until that happens. So value over a shorter period of time is skewed.

It's still skewed towards actual plays on the field. A plays defensive value is usually a 3rd of their offensive counterpart. It's why you need around 3 years of defensive data to get a read on how good a player is. But it still doesn't change what happened in a given year.

[Enscheff]

A run saved in a 5+ run game in the 9th inning has virtually 0 impact to that team's win loss expectancy. Therefore, it should have no impact when you are valuing how much a player contributes to wins and losses in a season.

Now if you want to get into his range and other defensive metrics that fine and I agree. But when converting those values to a win loss impact then these things should be considered.

Time for you to go learn about WPA (win probability added). Then go look up how few innings are actually played with a score difference of 5+ runs, and get back to us.

After all that, go read about how predictive WPA is vs other stats that don't take context into account.

After you have educated yourself, get back to us about how useful such stats are.

[Russ2dollas]

I'm still stuck on the idea that some posters think we'll flip kemp.

I think Coppy is all in on the Kemp Kool-Aid. I think if Kemp is crushing the ball with a .900 OPS at the break we won't trade him.

I'd love the idea of flipping him for anything of value and using DPeterson. B/c I think DPeterson will be close to what real Kemp will give us for WAR for a fraction of the cost.

I just can't see Coppy trading Kemp. I think he loves the guy.

[thethe]

It's still skewed towards actual plays on the field. A plays defensive value is usually a 3rd of their offensive counterpart. It's why you need around 3 years of defensive data to get a read on how good a player is. But it still doesn't change what happened in a given year.

But the noise in a year can be significant for meaningless plays.

Either way. Good talk and I'm glad you we can always have a good natured back and forth.

[thewupk]

But the noise in a year can be significant for meaningless plays.

Either way. Good talk and I'm glad you we can always have a good natured back and forth.

Just in the way that offense can be skewed in a given year with a super high BABIP. Or how Freeman in 2016 had a 1021 OPS with the margin over >4 runs and it was only 845 in late and close situations. I don't think either should diminish or heighten (if reversed) a players season. Those situations are out of the players control.

[thethe]

Just in the way that offense can be skewed in a given year with a super high BABIP. Or how Freeman in 2016 had a 1021 OPS with the margin over >4 runs and it was only 845 in late and close situations. I don't think either should diminish or heighten (if reversed) a players season. Those situations are out of the players control.

Ageeed. The players ability should not take a hit at all.