3/26/2004
From a message board conversation...
So, I was wondering if people more apt in the ways of sabrmetrics (Hi, Brendan!) could tell me if I have any glaring holes in my argument here:
Stats don't win championships, players do, if it were the other way around Mr. Beane would have about 4 WS rings right now.
Stats help choose the correct players that give you a chance to win. You read Moneyball...doesn't it strike you as insane how baseball scouting has been for years and years? Unless you have a NYY or Boston-sized budget to play with, then you have to get the absolute most value for money that you can. That's where OPS+ and ERA+ and Win Shares and all the rest of it come in.
But, all sabrmetrics can do is ensure that you GET to the playoffs. The only thing that determines playoff winners is luck. That's it. It's a crapshoot -- the sample size is so small, no other factor becomes statistically significant. Look at it this way: were the Florida Marlins the best team in baseball last year? Or in 1997? The Anaheim Angels were the best team in 2002? The answer to all of the above is "HELL NO!" However, look at the Marlins' run last year: 2 costly errors from Gold Glovers (JT Snow, Jose Cruz Jr.) got them past the Giants, the Cubs had that insane breakdown, and then the Yankees...well, they were a paper tiger that was extremely lucky to get past Boston. Luck, luck, and more luck.
Stats work very, VERY well when you have 162 games to look over (or 80 in other sports). That's what they are there for.
But still, I believe that you do need a type of attitude and mental toughness to be a closer.
Anyone who makes the major leagues should have this attribute. Those who slip through the cracks do so because of poor scouting and coaching.
So from what you are saying, Rivera would be just as good as SP as a closer.
No. Some pitchers are born to be starters (Hi, Derek Lowe!), and some are born to be closers. But, once that distinction is made, there is very little to separate one closer from another. Even in the case of a very good closer like, say, Troy Percival, you can argue that Weber and Donnelly and Rodriguez are just as good, if not better members of that bullpen. Say Percival goes down...you're saying one of those three can't come in and rack up the same gaudy save numbers? Closers are very, VERY fungible, and VASTLY overrated in the market (Hi, Billy Koch!).
How many leads has BK Kim blown against the pressure of playing against the Yankees as opposed to other less meaningful games.
That's why Bill James measures saves in three categories. If you see someone blowing easy saves, then they should probably be in AAA...if they can't get 3 outs with a lead and nobody on, then they damn sure aren't going to be any more useful in the 7th with guys on already.
Bottom line is, Yes Stats mean a lot, but would you pitch BK Kim in a Game 7 of an ALCS against the Yankees with a lead?
No, but not for something as ephemeral as "he can't handle pressure". I wouldn't pitch him (or even have him on my roster in the first place) because he walks a ton of guys, he gives up a lot of hits, and is just not a good pitcher in general. Arizona got their one good season out of him...and they should count themselves as lucky for that.
Now, on to Goose:
I think it is essential for a team to have someone that can close the door in the 8th and 9th innings to have a successful team nowadays.
To me, it's not the inning that's important, but the situation. I want my best reliever on the mound when the danger to my lead is at its greatest. Let's say I'm managing Philly, and I'm in the World Series vs. the AL champ. I'm winning in the 7th, but it's close, my starter is tiring, and there's a guy on. You can bet your ass that I'm going to put Billy Wagner in then, rather than depend on a B or B- reliever to get me through it, in the hope that I'm still winning when the 9th comes around.
As for the Hall of Famers you mentioned, for a lot of those guys, the save was a very different stat. You had to go like 2 or 3 innings to get it, and I think the biggest possible lead to be eligible was like 2. Sparky Lyle won a Cy Young, and only got 26 saves (or something in that neighborhood). Actually, I claim to be the stats man, I may as well use them.
Sparky Lyle, 1977: 72 G, 13-5, 26 SV, 2.17 ERA, 182 ERA+ (meaning he's about 52% better than league average for a reliever), 137 IP, 33 BB, 68 K.
Bobby Thigpen, 1990: 77 G, 4-6, 57 SV, 1.83 ERA, 210 ERA+, 88 2/3 IP, 32 BB, 70 K.
Thigpen's numbers are certainly flashier (and indicative why ERA+ isn't a be-all and end-all), but let's analyze this for a second. They have similar walk and strikeout numbers, but Lyle's ERA is a shade higher. And, of course, Thigpen has a kajillion saves there. But, notice how Lyle pitched around 50 MORE INNINGS than Thigpen did. Not only that, Lyle lost 1 less game, won a bunch more, and obviously was called on to put out actual fires, while Thigpen's numbers are VERY indicative of the dilettante "comes in when there's no real danger, and gets a bunch of cheap saves" kind of guy.
Not only that, but Lyle had a long, consistent career. Thigpen had 2 decent 30-save Rocky Biddle-esque seasons before the record-breaker, and then...wow...don't look below. He had three more seasons of consequence after that (with a tremendously steep rate of decline), and then he was GONE. Thanks for coming, Bobby.
And, when you look at Thigpen's Similarity Scores, it paints a VERY interesting picture. On one hand, he's similar to known busts Gregg Olson, Armando Benetiez, and Mel Rojas. He ALSO is similar to "proven closers" like Keith Foulke, Jason Isringhausen, and Ugueth Urbina. To me, one can draw the conclusion that Foulke and Isringhausen and the like have simply had better teams to draw on, and thus had more 3-run leads, that sorta thing.
Even the great Mariano Rivera...stick him on the Detroit Tigers, and he's a very nice pitcher, but not really value for money. Just for fun, let's compare Mariano Rivera's season last year, with...oh, I dunno...some middle-of-the-road guy. I know...how about that guy MacDougal from Kansas City?
Rivera, 2003: 64 G, 5-2, 40 SV, 6 BS, 1.66 ERA, 265 ERA+, 70 2/3 IP, 10 BB, 63 K
MacDougal, 2003: 68 G, 3-5, 27 SV, 8 BS, 4.08 ERA, 124 ERA+, 64 IP, 32 BB, 57 K
It's not even an issue that Rivera is the far superior pitcher. Hell, his ERA+ his about double, which would mean he's twice the pitcher MacDougal is. I'd go along with that (ERA+ is a bit clumsy, but it IS useful when making general assessments). But, let's look a little deeper.
The 2003 Yankees were 101-61, whilst the 2003 Royals were 83-79. Rivera, all told, had 45 positive outcomes (wins and saves), and 8 negative ones (yes, sometimes a blown save turns into a win, but I'm just going rough and ready here). So, Rivera positively affected 44.6% of his team's wins, (again, a real sabrmetrician would shoot some holes in my methodology, but I think it works well enough for this super-basic point) and neagtively affected 11.4% of the losses. Meanwhile, Macdougal had 30 positive outcomes, and 13 negative ones. So, that works out to be 36.1% of their wins, and 16.5% of the losses.
Ergo, the difference between an absolutely top-shelf closer and a generic Some Guy is about 8% on the winning side and 5% on the losing side (of course, there's other factors...Rivera won't exit his prime for another year or so, while MacDougal is just getting there), but to me, this all leads up to one point:
As long as you don't put a total zero in the closer's spot, then it won't make a gigantic difference whether you have Mariano Rivera, or just Some Guy.
Stats don't win championships, players do, if it were the other way around Mr. Beane would have about 4 WS rings right now.
Stats help choose the correct players that give you a chance to win. You read Moneyball...doesn't it strike you as insane how baseball scouting has been for years and years? Unless you have a NYY or Boston-sized budget to play with, then you have to get the absolute most value for money that you can. That's where OPS+ and ERA+ and Win Shares and all the rest of it come in.
But, all sabrmetrics can do is ensure that you GET to the playoffs. The only thing that determines playoff winners is luck. That's it. It's a crapshoot -- the sample size is so small, no other factor becomes statistically significant. Look at it this way: were the Florida Marlins the best team in baseball last year? Or in 1997? The Anaheim Angels were the best team in 2002? The answer to all of the above is "HELL NO!" However, look at the Marlins' run last year: 2 costly errors from Gold Glovers (JT Snow, Jose Cruz Jr.) got them past the Giants, the Cubs had that insane breakdown, and then the Yankees...well, they were a paper tiger that was extremely lucky to get past Boston. Luck, luck, and more luck.
Stats work very, VERY well when you have 162 games to look over (or 80 in other sports). That's what they are there for.
But still, I believe that you do need a type of attitude and mental toughness to be a closer.
Anyone who makes the major leagues should have this attribute. Those who slip through the cracks do so because of poor scouting and coaching.
So from what you are saying, Rivera would be just as good as SP as a closer.
No. Some pitchers are born to be starters (Hi, Derek Lowe!), and some are born to be closers. But, once that distinction is made, there is very little to separate one closer from another. Even in the case of a very good closer like, say, Troy Percival, you can argue that Weber and Donnelly and Rodriguez are just as good, if not better members of that bullpen. Say Percival goes down...you're saying one of those three can't come in and rack up the same gaudy save numbers? Closers are very, VERY fungible, and VASTLY overrated in the market (Hi, Billy Koch!).
How many leads has BK Kim blown against the pressure of playing against the Yankees as opposed to other less meaningful games.
That's why Bill James measures saves in three categories. If you see someone blowing easy saves, then they should probably be in AAA...if they can't get 3 outs with a lead and nobody on, then they damn sure aren't going to be any more useful in the 7th with guys on already.
Bottom line is, Yes Stats mean a lot, but would you pitch BK Kim in a Game 7 of an ALCS against the Yankees with a lead?
No, but not for something as ephemeral as "he can't handle pressure". I wouldn't pitch him (or even have him on my roster in the first place) because he walks a ton of guys, he gives up a lot of hits, and is just not a good pitcher in general. Arizona got their one good season out of him...and they should count themselves as lucky for that.
Now, on to Goose:
I think it is essential for a team to have someone that can close the door in the 8th and 9th innings to have a successful team nowadays.
To me, it's not the inning that's important, but the situation. I want my best reliever on the mound when the danger to my lead is at its greatest. Let's say I'm managing Philly, and I'm in the World Series vs. the AL champ. I'm winning in the 7th, but it's close, my starter is tiring, and there's a guy on. You can bet your ass that I'm going to put Billy Wagner in then, rather than depend on a B or B- reliever to get me through it, in the hope that I'm still winning when the 9th comes around.
As for the Hall of Famers you mentioned, for a lot of those guys, the save was a very different stat. You had to go like 2 or 3 innings to get it, and I think the biggest possible lead to be eligible was like 2. Sparky Lyle won a Cy Young, and only got 26 saves (or something in that neighborhood). Actually, I claim to be the stats man, I may as well use them.
Sparky Lyle, 1977: 72 G, 13-5, 26 SV, 2.17 ERA, 182 ERA+ (meaning he's about 52% better than league average for a reliever), 137 IP, 33 BB, 68 K.
Bobby Thigpen, 1990: 77 G, 4-6, 57 SV, 1.83 ERA, 210 ERA+, 88 2/3 IP, 32 BB, 70 K.
Thigpen's numbers are certainly flashier (and indicative why ERA+ isn't a be-all and end-all), but let's analyze this for a second. They have similar walk and strikeout numbers, but Lyle's ERA is a shade higher. And, of course, Thigpen has a kajillion saves there. But, notice how Lyle pitched around 50 MORE INNINGS than Thigpen did. Not only that, Lyle lost 1 less game, won a bunch more, and obviously was called on to put out actual fires, while Thigpen's numbers are VERY indicative of the dilettante "comes in when there's no real danger, and gets a bunch of cheap saves" kind of guy.
Not only that, but Lyle had a long, consistent career. Thigpen had 2 decent 30-save Rocky Biddle-esque seasons before the record-breaker, and then...wow...don't look below. He had three more seasons of consequence after that (with a tremendously steep rate of decline), and then he was GONE. Thanks for coming, Bobby.
And, when you look at Thigpen's Similarity Scores, it paints a VERY interesting picture. On one hand, he's similar to known busts Gregg Olson, Armando Benetiez, and Mel Rojas. He ALSO is similar to "proven closers" like Keith Foulke, Jason Isringhausen, and Ugueth Urbina. To me, one can draw the conclusion that Foulke and Isringhausen and the like have simply had better teams to draw on, and thus had more 3-run leads, that sorta thing.
Even the great Mariano Rivera...stick him on the Detroit Tigers, and he's a very nice pitcher, but not really value for money. Just for fun, let's compare Mariano Rivera's season last year, with...oh, I dunno...some middle-of-the-road guy. I know...how about that guy MacDougal from Kansas City?
Rivera, 2003: 64 G, 5-2, 40 SV, 6 BS, 1.66 ERA, 265 ERA+, 70 2/3 IP, 10 BB, 63 K
MacDougal, 2003: 68 G, 3-5, 27 SV, 8 BS, 4.08 ERA, 124 ERA+, 64 IP, 32 BB, 57 K
It's not even an issue that Rivera is the far superior pitcher. Hell, his ERA+ his about double, which would mean he's twice the pitcher MacDougal is. I'd go along with that (ERA+ is a bit clumsy, but it IS useful when making general assessments). But, let's look a little deeper.
The 2003 Yankees were 101-61, whilst the 2003 Royals were 83-79. Rivera, all told, had 45 positive outcomes (wins and saves), and 8 negative ones (yes, sometimes a blown save turns into a win, but I'm just going rough and ready here). So, Rivera positively affected 44.6% of his team's wins, (again, a real sabrmetrician would shoot some holes in my methodology, but I think it works well enough for this super-basic point) and neagtively affected 11.4% of the losses. Meanwhile, Macdougal had 30 positive outcomes, and 13 negative ones. So, that works out to be 36.1% of their wins, and 16.5% of the losses.
Ergo, the difference between an absolutely top-shelf closer and a generic Some Guy is about 8% on the winning side and 5% on the losing side (of course, there's other factors...Rivera won't exit his prime for another year or so, while MacDougal is just getting there), but to me, this all leads up to one point:
As long as you don't put a total zero in the closer's spot, then it won't make a gigantic difference whether you have Mariano Rivera, or just Some Guy.
