Posted by dahsdebater on 2/14/2016 11:20:00 AM (view original):
It's more of a runaway than I thought.  UZR has Heyward at 89.2 since 2011.  2nd place is Andrelton Simmons at 68.4, and Gordon is 3rd at 67.2.  DRS has Simmons ahead of Heyward.  I personally have always felt that UZR is the most accurate comprehensive defensive stat, but I think a lot of people tend to lean on DRS now.  So I guess you could argue that stats have Heyward as the 2nd-most valuable defensive player in baseball over the past 5 seasons...

If Simmons and Machado stayed healthy and played everyday, they'd probably be the most valuable defenders in baseball.  But other than those guys and Kiermaier, I can't think of anybody similarly impressive to watch play than Heyward right now.  Again, my guess is, if you don't agree, you haven't watched enough of Heyward's games.

$180 mil for RF defense? How come Adam Everett didn't get paid that much? Best defensive SS

Dahs, you exemplify everything that is wrong with the saber crowd. You buy into these metrics, even when they're totally in contrast with concrete stats or common sense, and you act like you're all intelligent and enlightened while everyone else is an idiot who simply isn't bright enough to grasp their revolutionary value.

Player A gets to more balls, but makes an error every 82 chances.

Player B gets to fewer balls, but makes an error every 162 chances. And has a better arm.

You can choose to trust in your metrics, but if we're purely talking defense, I'm taking Player B all day, everyday, and I'm sure most would as well. I've watched Heyward play - he's good. But I've never watched him and thought "Wow! He's the best in the game!". And neither would you if it weren't for these precious "enlightened" metrics you've attributed as gospel.

And for what it's worth - having watched nearly every O's game during Nick's time here, he was vastly underrated defensively and one guy sabermetrics always got wrong. He, among others, is a big reason I don't put much stock on them. Another is JP Arencibia, who finished fifth among full-time catchers in dWAR in 2013 (ahead of Posey, Wieters, Ellis, Lucroy, Montero, Pierzynski and others), despite having 11 errors and 13 passed balls - both tops among qualifiers - and a CS% that was fourth from the bottom. Advanced metrics have value, but they're flawed. They're only valuable in conjunction with tangible statistics and actually watching players play.

If you're going to take them as gospel, even in cases where they're obviously wrong, I think it's pretty obvious who the stupid one is.

Posted by Jtpsops on 2/14/2016 1:32:00 PM (view original):
Dahs, you exemplify everything that is wrong with the saber crowd. You buy into these metrics, even when they're totally in contrast with concrete stats or common sense, and you act like you're all intelligent and enlightened while everyone else is an idiot who simply isn't bright enough to grasp their revolutionary value.

Player A gets to more balls, but makes an error every 82 chances.

Player B gets to fewer balls, but makes an error every 162 chances. And has a better arm.

You can choose to trust in your metrics, but if we're purely talking defense, I'm taking Player B all day, everyday, and I'm sure most would as well. I've watched Heyward play - he's good. But I've never watched him and thought "Wow! He's the best in the game!". And neither would you if it weren't for these precious "enlightened" metrics you've attributed as gospel.

And for what it's worth - having watched nearly every O's game during Nick's time here, he was vastly underrated defensively and one guy sabermetrics always got wrong. He, among others, is a big reason I don't put much stock on them. Another is JP Arencibia, who finished fifth among full-time catchers in dWAR in 2013 (ahead of Posey, Wieters, Ellis, Lucroy, Montero, Pierzynski and others), despite having 11 errors and 13 passed balls - both tops among qualifiers - and a CS% that was fourth from the bottom. Advanced metrics have value, but they're flawed. They're only valuable in conjunction with tangible statistics and actually watching players play.

If you're going to take them as gospel, even in cases where they're obviously wrong, I think it's pretty obvious who the stupid one is.

This
I think SABRmetrics are cool but, they're flawed. Hopefully that route efficiency fracker takes off. En we can really see who has the best range.
Even though I hate Mike Trout, (good player but band wagoners make me hope he fails), his route efficiency everytime I turned on ESPN was like 98% and he was running 19-21 MPH.
So hopefully there's a formula to judge range based on that.

We rely on stats because that is the medium with wich we choose to explain baseball players performance. Having said that. Nick Markakis was/is a star. Jason Heyward is a star with superstar potential.

Jtp - I missed your post until this just got bumped. I'll respond to it and we'll see if you come back to defend your absurd position. I'll use your numbers, since it's pretty obvious you didn't bother actually doing the math:

Ok, so Heyward has one error every 82 chances. In his last season with the Braves, he had a range factor of 2.56 chances/9 innings. The 2 years before that, it was more like 2.3, so let's split the difference a little and say he gets to 2.4 balls/9. Now let's say you play about 1300 innings. That's a little low, but in the normal range - about 145 or 150 games worth of defense. So based on that range factor, in 1300 innings, Heyward gets to 347 balls. One error per 82 chances means about 4 errors, so he makes 343 outs.

On the same team, the very next season, Markakis put up a range factor of 2.02. So in 1300 innings, he gets to 292 balls. One error per 162 chances gives him only 1 or 2 errors. So let's give him credit for 291 outs.

Now, you want to talk about arm. The tangible result of throwing arm is outfield assists. In 2014, Heyward had 10 of 'em. In 2015, on the same team, Markakis had 4 in more innings. So add in those extra 6 outs and Heyward is making 58 more outs per season. Again, this is using YOUR numbers. So your argument is that "most" would take player B. I'd say the contracts of players A and B beg to differ. Most would take A. And unless you're arguing that the value in preventing runner advancement of Markakis' arm is worth more than 58 outs, "most" would be right, and you would be wrong.

58 outs is 19 innings. That's more than 9 runs worth of advantage. I don't think the arms are all that different, either. And, by the way, if you look just MARGINALLY deeper, you'd find that Heyward made 6 errors in 2010, 6 in 2011, 5 in 2012, and 4 total in the past 3 seasons combined. Which is more than Markakis made in the same 3 years (1). But that 1 extra out given back per season is probably not very important relative to the extra 60 balls per year he's getting to.

I don't think the arms are as different as you think, either.

You don't like comparing apples to apples, do you? That season was Heyward's 5th...Markakis's in Atlanta was his 10th, so ya, I'd expect a 25 year old Heyward to be better than a 32 year old Markakis. Markakis had 72 assists his first 6 years in the league. Heyward has 43. I'd say there is a significant difference in arms. And FWIW, I never once said Markais was better than Heyward. Just that they were comparable. And they are.

But as I mentioned previously, Alex Gordon is a great comp for Heyward. Their career innings played and chances in the OF are nearly identical. Gordon has 10 fewer errors and 25 more assists in the same amount of chances. Yet Heyward's career dWAR is 9.8 in his 6 seasons to Gordon's 6.8 in his 6 seasons as an OFer. I'm not buying it. At all. I would take Gordon over Heyward in the OF defensively.

I love how you also didn't acknowledge my JP Arencibia example above. Advanced metrics had him as a better defender than Posey, Wieters, Ellis, Lucroy, Montero and Pierzynski in 2013, despite him leading the majors in passed balls and errors at his position and having the fourth worst caught stealing percentage.

I'm not saying advanced metrics have no value but to act like they're the be all, end all is ridiculous. You have to look at them in tandem with stats and common sense. I watched Markakis, almost every game, for years. Am I saying he was ever the best RF in the game? Not at all. But he's proof of how flawed advanced metrics are because they've pretty much always had him as a negative/below average defender, and he's far from it.

I just had another debate on dWAR recently, and I'll let you weigh on some stats, before I give you the players and their respective dWAR. All players play the same position. Rank these players based on the stats below, in terms of who you think had a better season. There doesn't need to be any rationalization or conditions. Just tell us how you'd rank the players below based solely on the numbers provided.

Player A: 482 chances, 23 errors (.952 fielding % vs league average of .956), RF/G of 3.06 (vs league average of 2.55)

Player B: 364 chances,15 errors (.959 fielding % vs league average of .962), RF/G of 3.14 (vs league average of 2.52)

Player C: 488 chances, 19 errors (.962 fielding % vs league average of .956), RF/G of 2.79 (vs league average of 2.55)

Player D: 362 chances, 12 errors (.967 fielding % vs league average of .956), RF/G of 2.57 (vs league average of 2.55)

Dahsdebater - go for it.

Posted by Jtpsops on 2/28/2016 11:14:00 PM (view original):
You don't like comparing apples to apples, do you? That season was Heyward's 5th...Markakis's in Atlanta was his 10th, so ya, I'd expect a 25 year old Heyward to be better than a 32 year old Markakis. Markakis had 72 assists his first 6 years in the league. Heyward has 43. I'd say there is a significant difference in arms. And FWIW, I never once said Markais was better than Heyward. Just that they were comparable. And they are.

But as I mentioned previously, Alex Gordon is a great comp for Heyward. Their career innings played and chances in the OF are nearly identical. Gordon has 10 fewer errors and 25 more assists in the same amount of chances. Yet Heyward's career dWAR is 9.8 in his 6 seasons to Gordon's 6.8 in his 6 seasons as an OFer. I'm not buying it. At all. I would take Gordon over Heyward in the OF defensively.

I love how you also didn't acknowledge my JP Arencibia example above. Advanced metrics had him as a better defender than Posey, Wieters, Ellis, Lucroy, Montero and Pierzynski in 2013, despite him leading the majors in passed balls and errors at his position and having the fourth worst caught stealing percentage.

I'm not saying advanced metrics have no value but to act like they're the be all, end all is ridiculous. You have to look at them in tandem with stats and common sense. I watched Markakis, almost every game, for years. Am I saying he was ever the best RF in the game? Not at all. But he's proof of how flawed advanced metrics are because they've pretty much always had him as a negative/below average defender, and he's far from it.

I just had another debate on dWAR recently, and I'll let you weigh on some stats, before I give you the players and their respective dWAR. All players play the same position. Rank these players based on the stats below, in terms of who you think had a better season. There doesn't need to be any rationalization or conditions. Just tell us how you'd rank the players below based solely on the numbers provided.

Player A: 482 chances, 23 errors (.952 fielding % vs league average of .956), RF/G of 3.06 (vs league average of 2.55)

Player B: 364 chances,15 errors (.959 fielding % vs league average of .962), RF/G of 3.14 (vs league average of 2.52)

Player C: 488 chances, 19 errors (.962 fielding % vs league average of .956), RF/G of 2.79 (vs league average of 2.55)

Player D: 362 chances, 12 errors (.967 fielding % vs league average of .956), RF/G of 2.57 (vs league average of 2.55)

If a stat only tells us what we already know from other stats, it's useless. The point of advanced defensive stats is to go beyond errors and range factor because those two stats don't tell us much about how good someone is.

Your ranking game is no better than if I gave you these four players and asked you to rank them:

Player A 80 Runs scored, 100 RBI
Player B 64 RS. 76 RBI
Player C 74 RS, 95 RBI
Player D 82 RS, 79 RBI

There's a lot more to offense than RS and RBI.

No, the point of advanced metrics is to create formulas from known stats, in order to create one or two "catch all" numbers that we can use to define someone's value. Not saying there's not value in it, but it's subjective. WAR, for example, is not even calculated the same way by everyone.

Case and point, from Baseball-Reference:

"There is no one way to determine WAR. There are hundreds of steps to make this calculation, and dozens of places where reasonable people can disagree on the best way to implement a particular part of the framework. We have taken the utmost care and study at each step in the process, and believe all of our choices are well reasoned and defensible. But WAR is necessarily an approximation and will never be as precise or accurate as one would like."

Metrics are formulas that someone created in an attempt to best define a player's value, and enough people agreed with it and accepted it as a useful tool. That doesn't mean it's concrete or without flaws.

Posted by Jtpsops on 2/29/2016 12:58:00 PM (view original):
No, the point of advanced metrics is to create formulas from known stats, in order to create one or two "catch all" numbers that we can use to define someone's value. Not saying there's not value in it, but it's subjective. WAR, for example, is not even calculated the same way by everyone.

Case and point, from Baseball-Reference:

"There is no one way to determine WAR. There are hundreds of steps to make this calculation, and dozens of places where reasonable people can disagree on the best way to implement a particular part of the framework. We have taken the utmost care and study at each step in the process, and believe all of our choices are well reasoned and defensible. But WAR is necessarily an approximation and will never be as precise or accurate as one would like."

Metrics are formulas that someone created in an attempt to best define a player's value, and enough people agreed with it and accepted it as a useful tool. That doesn't mean it's concrete or without flaws.

___________________

No, the point of advanced metrics is to create formulas from known stats

Nope. UZR, for instance, doesn't use fielding percentage or range factor (or any other "known stat"). It's a completely different stat.

it's subjective. WAR, for example, is not even calculated the same way by everyone

The fact that the formula that varies slightly between BR and fangraphs doesn't make the stat subjective.

Yes, WAR is an approximation. No one has ever denied that.

Metrics are formulas that someone created in an attempt to best define a player's value

You realize that ALL stats are metrics, right?

All stats are metrics, all metrics are not stats. Stats are tangible - they can be measured. Advanced metrics are man-made formulas - they are often subjective, not concrete. I can count up how many hits a player gets with 100% accuracy. You can't count up a guy's WAR and get a concrete number.

And your example above is again comparing apples to oranges. You're using the dahsdebater approach to debating. A hitter has no control over runs or RBI in the sense that he can't control if anyone is on base when he hits, or if other hitters drive him in. A fielder is directly responsible for how many balls he gets to/doesn't get to and how many errors he makes.

I can gain a good idea of how productive a fielder is by his total chances relative to the rest of the league and how many times he screws up a play. Runs and RBI don't really give you any measure of how good a hitter is, other than maybe his ability to hit situationally to bring in runners when they are on base. A better comparison would be if you gave me a hitters slash line and homers and told me to determine his value. From those numbers, I could get a pretty good idea.

In addition, my point with my example above was to show that in no way does dWAR correlate logically with defensive statistics. I guarantee if people ranked those 4 players above, they would not rank them in the same order dWAR does. If advanced metrics are correct all the time, they should always correlate with statistics.

That is my point. Sometimes stats and advanced metrics clash. Your side will always blindly say "well, no, it doesn't make sense - he does have pretty good stats. But WAR says he sucks, so he sucks!!" Personally, when there's a contradiction, I prefer to lean towards the side of concrete, measurable statistics over an "approximate" formula.

I didn't respond to your Arencibia question because it was irrelevant to the conversation. dWAR is a flawed stat, especially on BBR. But it's by far the least reliable in evaluating catchers. Knowing that, and since Arencibia had nothing to do with the question at hand, I ignored it.

As far as Gordon vs. Heyward goes, you've ignored too many factors yet again. You keep trying to oversimplify the problem. You can quickly glance at Heyward and Gordon and see that they have similar range factors and say, "hey look, they're the same." But that's remarkably ignorant. For the 6 years that Gordon has been an outfielder, the Royals pitching have averaged 7.07 K/9 and a .763 GB/FB ratio. Heyward's pitching staffs (2010-2014 Braves, 2015 Cardinals) have averaged 7.9 K/9 and a .898 GB/FB ratio. That means that Heywards teams got almost 12% fewer outs in play per game, and had over 7.5% fewer balls hit to the outfield. If you multiply those numbers, you'd expect Heyward's teams' outfielders to get about 18.5% fewer opportunities to make outs than Gordon's teams' outfielders. That makes Gordon and Heyward look a lot less similar.

This is reflected in fairly simple stats, too. Are you familiar with RZR (revised zone rating)? This is a simple number in modern baseball - the number of plays made by the fielder divided by the number of balls hit into his region of the field. Given how accurately modern cameras capture where the ball goes, this is a highly accurate stat for the past decade and a half or so. It's not a composite of anything. It's basically the fielding equivalent of batting average.

Heyward's RZR is .931. Gordon's is .913. The average COF is about .901. So Heyward is about 3 times as far above average in terms of making plays. RZR is still fairly simple. The more advanced metrics that play into FanGraphs dWAR consider more details, like where in the zone the ball is hit, and how hard, to determine how hard plays really are. But the basic numbers seem to be all you're comfortable with. So I'm giving you the basic numbers that show why Heyward is a better defender than Gordon.

FWIW, Markakis' RZR over his first 6 seasons was basically right on that league average of .901.