Nick Markakis' defense rated favorably by Inside Edge fielding metric

[thethe]

I'm not even saying this ends the discussion about Markakis. It's just funny to me everybody was so sure he was horrid because every metric showed this and yet we see one system giving him a solid evaluation. I think Markakis will be good but it won't look that way because Heyward was so fantastic out there. The real question is just how much defense contributes to wins and losses and just how much teams should pay for it.

 

[thethe]

Walks don't matter.

I think the impact of walks and the number of walks will significantly decrease as we move further away from the steroid era. A lot of what is known about baseball from Saber metrics was devised from corrupted data. Walks, bit to be more clear batters who draw pitches will always be important but I think we are going to see a much different brand of baseball in the future.

 

[yeezus]

The real question is just how much defense contributes to wins and losses and just how much teams should pay for it.

Yup, agreed. And right now no one knows, and I'm not sure there will ever be an even close to perfect way to measure it.

 

[chop2chip]

Yup, agreed. And right now no one knows, and I'm not sure there will ever be an even close to perfect way to measure it.

We have a near perfect understanding of the importance of defense.

We just aren't even close to knowing how to evaluate it precisely enough.

 

[chop2chip]

I'm glad that this new defensive statistic exists. I hope it begins to peel the layers back more and more about the imperfections of UZR and DRS.

People in the statistical community rely WAY too much on those two metrics.

 

[zitothebrave]

I'm not even saying this ends the discussion about Markakis. It's just funny to me everybody was so sure he was horrid because every metric showed this and yet we see one system giving him a solid evaluation. I think Markakis will be good but it won't look that way because Heyward was so fantastic out there. The real question is just how much defense contributes to wins and losses and just how much teams should pay for it.

I think Markakis is a bad defender from every piece of evidence I've gathered, including people i know who watch the game and I trust their opinion as they're unbiased. Not like say my sister who said that Victorino was as good of a defender as Andruw Jones and I immediately knew she was being a homer (same sister said Nick Foles was a better QB than Matt Ryan also earned an instant guffaw and Ryan Howard deserved the MVP award he won)

 

[thethe]

We have a near perfect understanding of the importance of defense.

We just aren't even close to knowing how to evaluate it precisely enough.

We do not have a clear understandin how of much it contributes to wins and losses not even close.

 

[zitothebrave]

I'm glad that this new defensive statistic exists. I hope it begins to peel the layers back more and more about the imperfections of UZR and DRS.

People in the statistical community rely WAY too much on those two metrics.

They're relied upon because they're the best that's out there. Inside Edge, and RZR are cool stats, but they have issues too. All stats have issues. If you have 4 metrics, 3 things tell you one thing 1 tells you another, they usually paint a good enoguh composite picture.

 

[chop2chip]

We do not have a clear understandin how of much it contributes to wins and losses not even close.

Of course we do.

We know that a run saved = run created (defense = offense).

We also know that runs differential explains a very very high percentage of wins and losses (I believe regression models have it at about 85% ish). With the remaining unexplained being completely left to luck related factors. This is known as the Pythagorean record.

 

[thethe]

Of course we do.

We know that a run saved = run created (defense = offense).

We also know that runs differential explains a very very high percentage of wins and losses (I believe regression models have it at about 85% ish). With the remaining unexplained being completely left to luck related factors. This is known as the Pythagorean record.

Calculating a run saved is not an exact science. The technology is just starting to get to a point where we can get reliable data.

 

[chop2chip]

They're relied upon because they're the best that's out there. Inside Edge, and RZR are cool stats, but they have issues too. All stats have issues. If you have 4 metrics, 3 things tell you one thing 1 tells you another, they usually paint a good enoguh composite picture.

Just because something is the best out there doesn't mean it requires you to rely on it.

I think in general both UZR and DRS are fine, but using them to evaluate individual is overly simplistic. I don't think that people shouldn't look at them, but I just think they are too noisy to use as anything more than just a broad stroke way of measurement, such as, "Bad, average, good, excellent, Andruw". With the caveat that those you label "good" may actually be "bad" if evaluated in a different context.

Inside Edge also has major issues too because they still rely too heavily on subjective analysis.

 

[chop2chip]

Calculating a run saved is not an exact science. The technology is just starting to get to a point where we can get reliable data.

I'm not referring the stat run saved. I am referring to the value of saving a run in a sort of broad abstract sense.

It's petty on my part because we are essentially arguing the same thing. I'm just trying to convey the message that we know the value of preventing runs we are just very imprecise in our measurement of run prevention out side of the simple box score.

 

[thethe]

Just because something is the best out there doesn't mean it requires you to rely on it.

I think in general both UZR and DRS are fine, but using them to evaluate individual is overly simplistic. I don't think that people shouldn't look at them, but I just think they are too noisy to use as anything more than just a broad stroke way of measurement, such as, "Bad, average, good, excellent, Andruw".

Inside Edge also has major issues too because they still rely too heavily on subjective analysis.

I agree. We would be foolish not use data that is available to us but we must understand the limitations of it in its current form.

Also, I think every defensive metric has a High level of subjective analysis.

 

[zitothebrave]

I agree. We would be foolish not use data that is available to us but we must understand the limitations of it in its current form.

Also, I think every defensive metric has a High level of subjective analysis.

Every stat has a high level of subjective analysis. Who determines that a hit is a better struck ball than an out? Can a guy have a good average and be a bad hitter who was just lucky and gets figured out? Could go on but all stats require input and input can be flawed.

 

[thethe]

Every stat has a high level of subjective analysis. Who determines that a hit is a better struck ball than an out? Can a guy have a good average and be a bad hitter who was just lucky and gets figured out? Could go on but all stats require input and input can be flawed.

Of course that's the case but defensive metrics have the highest level of subjectivity.

 

[chop2chip]

Every stat has a high level of subjective analysis. Who determines that a hit is a better struck ball than an out? Can a guy have a good average and be a bad hitter who was just lucky and gets figured out? Could go on but all stats require input and input can be flawed.

The only level at which batting average is subjective (outside of the defined rules which are extremely arbitrary) is the case of an error. Which mind is you is very subjective but no where near the degree that defensive statistics are. I mean, come on, the statistical community universally agrees that defensive statistics are pretty shaky.

And mind you for every offensive statistic there are 10 more that try to fill in the holes. I think we are at the point where we basically have offense figured out. Defense, not so much.

 

[zitothebrave]

The only level at which batting average is subjective (outside of the defined rules which are extremely arbitrary) is the case of an error. Which mind is you is very subjective but no where near the degree that defensive statistics are. I mean, come on, the statistical community universally agrees that defensive statistics are pretty shaky.

I guess I should clarify my point as I realize I didn't make it clear, when you attach value to a stat, it's highly subjective

You can argue degrees, but if Jason hits a frozen rope that's right at a fielder 3 times and then squibs one off the end of the bat what's to determine that the last result was his best result? Similar to defensive stats, logn term average stats normalize, you sometimes need a few years before that happens though.

 

[zitothebrave]

Of course that's the case but defensive metrics have the highest level of subjectivity.

In your opinion. The second you attach a value to a result you're adding subjectivity. How much from the truth? We don't know because we dont' know the truth, if we did we wouldn't need those stats.

 

[chop2chip]

In your opinion. The second you attach a value to a result you're adding subjectivity. How much from the truth? We don't know because we dont' know the truth, if we did we wouldn't need those stats.

That's not an opinion. That's a fact. We know far more about how to accurately measure offense. The value of either is neglible because we know the theoretical value of a run created vs a run prevented. The only subjectiveness is measuring how runs are scored and how runs are prevented.

How is this a debate?

 

[thethe]

In your opinion. The second you attach a value to a result you're adding subjectivity. How much from the truth? We don't know because we dont' know the truth, if we did we wouldn't need those stats.

There are less variables in offensive stats. Therfore yiu can do deeper analysis and have a higher confidence level with the results. It's not even open for discussion Zito. The whole baseball community agrees with this.