The Bleacher Bums Header

With a little luck

Posted at 11:49 AM on October 25, 2005 by Bob Collins (21 Comments)

First this disclaimer: There's nothing -- absolutely nothing -- wrong with having good luck. On the other hand, there is something wrong with not acknowledging it's part of the reason for your success.

And, of course, luck is usually a part of everybody's success. The thing with luck, though, is that it's not distributed evenly and -- as most of us know -- it's not consistently available over time. You could, for example, flip a coin 10 times in a row and have it appear heads 10 times. But flip it enough, and it will appear heads as often as it appears tails.

A few years ago, the Kansas Royals were the talk of the American League. Behind new manager Tony Pena, they raced to an early season lead and it was the feel-good story of the early summer. The problem is, look beyond the 10 coin flips, and you could reasonably predict the 10 to come. Now it's true that if you flip a coin 10 times and it comes up heads, the odds of it coming up heads on the 11th are the same 50-50.

The Royals had a mediocre second half and the next year, they were -- once again -- horrible. This season, wunderkind manager Tony Pena was fired.

I bring this all up because of yesterday's entry on Scott Podsednik, and the feel-good story of the summer in Chicago -- that Scott Podsednik has remade this squad and the ChiSox were able to shake off the remnants of underachieving seasons past to race to a terrific division title and pennant and, now it appears, World Series.

I say good for them. And deepest sympathies on what's about to happen to their dynasty.

By noting the extent to which the coin appeared "heads" for the White Sox, it does not mean that they didn't have fine pitching or some good hitters (although, it should be noted, the number of runs scored by the ChiSox this year was pretty mediocre). It means that the talent they did have played to its expected (and beyond) potential, and they had -- wait for it -- a tremendous run of luck.

The undeniable reality of baseball is this: a win is made up of a run scored and a run not allowed. The degree to which those two items relate has been researched for years in hundreds of thousands of games and has been well established (a run allowed is slightly more damaging than a run scored).

But no matter. Nothing else matters but runs. If it exists, it can be measured in runs. If it can't be measured in runs, it doesn't exist (i.e. the well-worn team "chemistry" nonsense).

Taking the number of runs each team in the Central Division scored and allowed this year, then, we can predict a reasonable expectation of the number of wins the team would get. Try it. Take any team in history and take the number of runs scored (squared) and divide it by the number of runs scored (squared) and the number of runs allowed (squared). What's revealed is the expected winning percentage. Multiply that by the number of games played and, voila!

Here then is the 2005 Central Division.

Team RS RA Pctg ExpW ExpL Diff
White Sox 741 645 .569 92 70 +7
Indians 790 642 .602 98 64 -5
Twins 688 662 .519 84 78 -1
Tigers 723 719 .503 81 81 -10
Royals 700 935 .359 58 104 -2

I'd say two teams -- the Twins and Royals -- got exactly what they deserved. One team -- the Tigers -- should've won more games considering their team's production. The White Sox were well ahead of the game, and the Indians well behind.

So what can we expect from next year? Who knows. The Pythagorean method doesn't predict the future any more than you can predict the next 10 coin flips. But it helps analyze what happened.

I'm rooting for the White Sox because I think if they win several things will happen. First, the players will -- as most championship players do -- forget how hard it was and lose focus. Second, the management will throw big money around trying to keep the team exactly as it was in its championship year and , third, the odds of the coin flips will return balance to performance results.

So what to do? If you're the Twins, you wake up to the fact your team isn't that good and you stop trying to win with the Cuddyers and Jones and Fords and go out and get some quality hitters because nothing -- not even luck -- can help you now.

If you're the Tigers, you make some improvements and you can expect to be the feel-good story of 2006.

If you're the Indians, you figure out a way to sign Kevin Millwood, hope a rightfielder falls into your lap, and show up for every game.

If you're the ChiSox, you enjoy the rubber chicken.


Comments (21)

I'll simplify your analysis further. The only thing that matters is scoring more runs than you allow in a game. Measuring the margin of victory is useless. A victory with by 12 run margin is worth just as much as a victory by a 1 run margin.

James' Pythagorean formula doesn't take into account anything but runs and treats them as though they existed in a vacuum. They don't, however. The reason the Sox won 7 games over their "expected wins", which is another way of saying they won a lot of close ballgames, is their excellent pitching; pitching that will be under contract next year. The White Sox aren't going anywhere. Neither is Detroit for that matter. Cleveland, however, would seem to be the early favorite for the division.

Posted by Scott Janssens | October 25, 2005 1:19 PM


The runs don't exist in a vacuum, however they ARE the only thing that matters.

// They don't, however. The reason the Sox won 7 games over their "expected wins", which is another way of saying they won a lot of close ballgames, is their excellent pitching;

Not necessarily true. You can win close ballgames and have lousy pitching.

//The White Sox aren't going anywhere. Neither is Detroit for that matter. Cleveland, however, would seem to be the early favorite for the division.

Well, we'll see you back here next year. If all things remain equal -- and I'm guessing they don't -- then the ChiSox should come back to the pack, the good pitching notwithstanding.

The White Sox were a good -- not a great -- team in the last half of the season. I'm guessing that's pretty much who the White Sox are.

The offense is relatively weak, which of course puts a lot of pressure on a pitching staff. But I love Mark Buehrle. And I love that kid that pitched a few turns in September. Beyond that, the ChiSox probably aren't a 100-team. 92-94, maybe.

But God help 'em if they depend on Scott Podsednik and think they've got a good offense. They don't.

Posted by Bob Collins | October 25, 2005 1:54 PM


Oh, and they're excellent pitching IS reflected in the runs allowed figure, just as the Indians excellent pitching (they actually were better than then the ChiSox pitching, and they were better than the ChiSox hitting). But their excellent pitching -- from a theoretical point of view -- wasn't, by itself, enough to make up for their lack of offense.

Only it was this season.

Posted by Bob Collins | October 25, 2005 1:56 PM


Runs are not the only thing that matters, wins are, more importantly wins against division opponents. Who you score (or don't score) against is as important as simply scoring. The Sox went 14-5 against Cleveland this season including sweeping them with bench and minor league players the last weekend of the season when the Indians were fighting for the wildcard. Tell me exactly how the runs the Indians scored against the Royals affected their playoff run.

The Sox may not win as many games next year, but I would expect them to win around 92. That's certainly enough to compete for the wild card if not the division.

Posted by Scott Janssens | October 25, 2005 2:27 PM


Runs are the most important because those are the only factors that contribute to wins.

I agree with your prediction on 92 wins.

I don't understand your question regarding the Indians and Royals. The number of runs the Indians scored against the Royals afffected their playoff run only in relationto the number of runs they gave up.

Posted by Bob Collins | October 25, 2005 4:20 PM


Runs are the most important because those are the only factors that contribute to wins.

I agree with your prediction on 92 wins.

I don't understand your question regarding the Indians and Royals. The number of runs the Indians scored against the Royals afffected their playoff run only in relation to the number of runs they gave up.

Can't wait til the World Series is over so we can get to the free agent signings and trades.

Posted by Bob Collins | October 25, 2005 4:27 PM


To clarify the Royals comment (bear with me):

Here are the scores of the games between the Sox and Indians (Sox first): 1-0, 4-3, 5-11, 2-1, 5-4, 6-8, 6-4, 6-5, 4-6, 1-0, 7-1, 7-5, 4-0, 5-7, 7-6, 0-8, 3-2, 4-3, 3-1

RF: 80, RA: 75.
Expected Record: 10-9
Actual Record: 14-5

When the Indians won they usually won by a large margin. In the game they won 8-0 there were 7 runs that contributed nothing to the victory and nothing to their playoff push. I'm not saying you should try to score more runs just that those extra runs don't mean much after the game is over. Beating the Sox 8-0 doesn't mitigate any of the nine 1-run losses they suffered.

When the Indians are busy trouncing the Royals running up their run total, what do those extra runs contribute to their playoff push? The Sox don't score as many runs against the Royals because their offense isn't as potent. (I actually haven't added up the runs against the Royals for either team, but it illustrates my point.) My argument is that those extra runs by the Indians are misleading and not informative. Those runs didn't help the Indians overcome the Sox, only put a beating on the Royals.

Simply using run differential isn't digging deep enough. Head to head the Sox only scored 5 more runs than the Indians, yet the Sox finished 9 games ahead of them in the season series. That's 3 games more than in the standings. This is my point, it's not the runs alone, but who those runs come against and how many wins those runs generate.

Wins are what counts. Usually more runs lead to wins, but not always. In the World Series they count wins not runs. The team that has scored the most runs has usually, but not always, won the World Series.

Posted by Scott Janssens | October 25, 2005 5:16 PM


Your answer is in your numbers. The Indians won 4 more games than the number of runs scored/runs allowed would normally suggest.

So what happened? Well, look again. There were 7 1-run games in that series..

Pythagorean formula doesn't attempt to put a value on a run scored in an 8-0 win vs. a 2-1 win...it seems to explain what happened. You know by the runs scored number alone that the Indians scored a bunch of runs against the Royals. So the theory itself ain't gonna shed any light there.

And yet, there we are with a +4 differential over 19 games. What does that tell us? Well, if it's +4 over just 19 games, it tells me the next 19...or maybe the 19 after that...or the 19 after that, are likely not going to result in a +4.

The fact is that if you score as many runs as you give up, you're probably not going to play much more than .500 ball.

That doesn't mean you definitely won't. That doesn't mean you can't. But nobody should be surprised if you don't.

Posted by Bob Collins | October 25, 2005 5:37 PM


Oh, by the way, here's the potential flaw in the argument. What if next year is the year the luck evens out? What if this year was? I've thought the ChiSox were a much better team than the Twins for the last couple of years. But they appeared to underachieve.

I haven't looked at the last couple of years, but it wouldn't surprise me a bit if their expected wins were lower than their actual wins.

Posted by Bob Collins | October 25, 2005 5:49 PM


The Pythagorean formula is a "best-fit" line. The expected wins formula has been developed by analyzing data from the past. Of course, there will be variation.

Taken to the extreme, a team could score 100 runs and allow 2000 and win 100 games, if in those 100 games, the score is 1-0. Not very likely, now is it? What the White Sox did this year wasn't very likely either, given the number of runs that they scored vs. the number of runs that they allowed. A good record in one run games doesn't necessarily mean that a team is clutch or even good. A team with a four run lead that gives up three of those runs wins by one. Is that clutch?

The Arizona Diamondbacks were 28-18 in one run games. Does that mean they were good? Well, they were 77-85. (They won 13 more than they were expected to win, which is unbelievable.) So a bad team wins a huge number of games more than expected. It appears that more than anything, they were lucky.

On the other hand, Oakland was 26-24 in one run games, but won six fewer games than expected (88 v. 94). Lucky or unlucky? One run games the difference?

Posted by SBG | October 25, 2005 9:29 PM


I can remember one Indians game with the Royals (since it was pointed out earlier). They won 13-7. One might say they won on the strength of their offense and if you looked at 13-7, you'd say....darn good hitting machine.

And they were. Except...

The reason they scored 13 runs is because they scored 11 runs in the 9th inning and had absolutely no business doing so.

That was a loss the team deserved. Only they won.

Belliard, if I recall correctly (and I do because I just looked it up) popped out for the third out in a 7-3 KC win...except Berroa dropped the ball.

Then Aaron Boone lofted an EASY fly ball for the final out to left....only the leftfielder dropped the ball.

For that game, there should've been another column in the standings.

The point? What, I have to have a point now?

It's good to be good. It's better to be lucky AND good. And it's better to have more luck then the other guy who's lucky and good.

But stay out of Vegas.

Posted by Bob Collins | October 26, 2005 10:29 AM


I agree with the Pythagorean analysis, but the question is why the White Sox won seven more games than they might have been expected to. You say that it was luck, and it might have been, but not necessarily. Too often we use the word "luck" to explain things that we can't measure or define. It could be that they were lucky, or it could be that they played particularly badly in some games and got blown out, or it could be that they hit or pitched well in the clutch, or that Ozzie had a particularly good year managing, or that team chemistry played a role. It could be a hundred things. This is what people mean when they talk about "intangibles": things that exist, and have a real effect, but which we haven't figured out how to measure in a qualitative way. I'm not arguing that the White Sox haven't been lucky, just that you can't assume that's the only reason they won more games than they would have been expected to.

Posted by Jeff | October 26, 2005 11:35 AM


The 7 games they won above expectations doesn't necessarily mean they won 7 games because of luck. It means that their total performance over 162 games was 7 games better than could be expected.

Hitting hte clutch, pitching etc. are *actually* all measureable. They're measurable by runs scored or runs allowed. So the fact they hit "in the clutch" is already in the measurement.

For the most part -- and I alluded to this before -- there's no such thing as intangibles that contribute to a win or a loss. Only one thing contributes to a win or loss (well, OK, two things): runs scored and runs allowed.

All of that is measurable. If a guy contributes to a run scored, it's measurable. It shows up in his performance.

If a guy hollers louder than anybody on the bench, he really doesn't contribute anything to a run scored or a run allowed. He just contributes to the noise.

Reminds me of a few years ago late in the season sitting behind the plate at the metrodome when Tom Prince came up. He was batting .229. The guy behind me turns the guy next to him and says, "he's only batting .229, but he's been better than that."

I wanted to -- I really wanted to -- turn around and say "no, sir, he really hasn't."

but I didn't.

Posted by Bob Collins | October 26, 2005 12:29 PM


By the way, that doesn't mean that it's not nice that a guy screams louder on a bench. Let's say Player A is walking to the plate and Player B is on the bench waving his towel.

Now, let's say that strictly because of the towel-waving, Player A is feeling so good, that he hits a home-run.

One might argue that player B's contribution is an intangible that can't be measured.

But it can.

It's measured via Player A.

The run is calculated as part of the measurement. So there's nothing out there that's an "intangible" that doesn't somehow find its way into the calculation because there are only two ways -- where wins and losses are concerned -- that something can have an impact on a game-- it either leads to a run scored. Or it leads to a run not scored.

Nothing else is involved.

Posted by Bob Collins | October 26, 2005 12:32 PM


// // They don't, however. The reason the Sox won 7 games over their "expected wins", which is another way of saying they won a lot of close ballgames, is their excellent pitching;

// Not necessarily true. You can win close ballgames and have lousy pitching.

I'll take that and go a step further. In 2004, the White Sox had a record of 28-18 in one-run ballgames. Only the Oakland A's (33-19) had a better record in those games.

The Sox finished 83-79, far behind the Twins in the Central.

In 2005, there were six teams that finished with a better-than-.500 record in one run games while finishing with an overall record below .500, and in some cases it wasn't even close:

Tampa Bay - 67-95 - 29-25
Seattle - 69-93 - 26-23
Chicago Cubs - 79-83 - 26-20
Arizona - 77-85 - 28-18
San Fran - 75-87 - 27-25
Colorado - 67-95 - 25-24

Dang, look at that one-run record for the Devil Rays! They must have had some great pitching...wait...

Tampa, Arizona, and Colorado were among the worst pitching teams in the majors, yet they still managed to do significantly better in one-run games than they did overall. Why?

Part of it may be due to what happens when you have a particularly bad offense or bad defense (but not both - note that the Royals don't make this list) - the three teams on this list that didn't have bad pitching had bad offenses. When you lose, you're likely to lose by a lot, either because your offense doesn't score or your pitching collapses. With a bad offense or pitching staff, it would seem less likely for you to blow another team out, so when you do have a chance to win, it's likely to be close.

The better explanation, though, is just that you happened to get your runs in games when you needed them. And I haven't seen anybody who has demonstrated that 'scoring when you need it' is anything other than random chance.

Posted by David Wintheiser | October 26, 2005 12:48 PM


"The 7 games they won above expectations doesn't necessarily mean they won 7 games because of luck. It means that their total performance over 162 games was 7 games better than could be expected."

You're treating the formula as though it measured true probability. If I flip a coin 100 times and it comes up heads 57 times then you can say it came up heads 7 more time than could be expected. However, This isn't like that.

Bob, your recollection of the Indians/Royals game supports my point. Let's switch games however. The first series of the season for the Sox was against the Indians. The scores were 1-0, 4-3, 5-11. In the third game the Sox had a 5-2 lead going to the 9th. Shingo Takatsu came in to close and promptly gave up three solo home runs. The Indians then scored 6 runs in the 11th. The Indians outscored the Sox 15-10 in the series. Your approach says the Sox were lucky for winning one extra game as judging by run differential they should only be expected to have won one game. In fact only a lack of performance, not luck, prevented a sweep.

The Sox didn't win 99 games instead of 92 because they were lucky. They won 99 games because they did the things necessary to win 99 games. The Indians were not a better team than the Sox because they had a better run differential or a higher number of expected wins. The Sox demonstrated their superiority by going 14-5 against them this season including a four game sweep right after the all-star game and a three game sweep during the last weekend of the season when the Indians were supposedly at their peak.

I think the Sox, with their pitching, could repeat that feat just as easily as not next season.

Posted by Scott Janssens | October 26, 2005 1:33 PM


//You're treating the formula as though it measured true probability.

I said in the original missive, I believe that it doesn't measure anything but performance.

//If I flip a coin 100 times and it comes up heads 57 times then you can say it came up heads 7 more time than could be expected. However, This isn't like that.

I didn't say it was. I noted the coin flip analogy only to point out that the chances of something happening is relatively constant to the extent it can be measured. We know as a fact that runs scored and runs allowed are the only real factors in a win. We know the extent to which each contributes to a win.

//Bob, your recollection of the Indians/Royals game supports my point.

I'm not really sure what the point was, though. I think the formula has very little to do -- or at least very little meaning -- over a game here or there.

// Your approach says the Sox were lucky for winning one extra game as judging by run

Well, first of all, it's not my approach. But second, if there were only a one game differential, I would never characterize the difference as significant, let alone the result of luck.

//differential they should only be expected to have won one game. In fact only a lack of performance, not luck, prevented a sweep.

I don't know how to respond to that. I'd never use the theory to analyze three games. I wouldn't need to if I watched the game. I'd be able to figure out why they won three games. That's harder over 162 games.

//The Sox didn't win 99 games instead of 92 because they were lucky. They won 99 games because they did the things necessary to win 99 games.

I think that's wrong. I don't think teams always win games they deserve to win. I think sometimes things just happen. People ae in the right place at the right time.

But you're missing my point. The ChiSox won what they won fair and square. However -- and this is the point -- I wouldn't expect them to win 99 games again -- all things being equal. Now, I'm not sure they are equal, of course, because I don't know what's going to happen with Konerko and I think they have one stud pitcher about to join the team and Garland has just flat-out arrived.

So they may have better offense and better pitching AND win more games next year. But, like I said, I wouldn't be surprised if they fall back to 92, which is a more reasonable expectation given their numbers.

// The Indians were not a better team than the Sox

I didn't say they were a better team. They had better pitching and their offense was better. But yet they still finished second. Was that their own doing? Of course it was. Should the ChiSox be in the World Series? Of course they should.

Do I believe that fate has been kind to the ChiSox? Absolutely. Grab it while you can.

And, by the way, The Tribe wasn't at their peak because they'd just lost four straight games to two last-place teams.

And they choked.

Posted by Bob Collins | October 26, 2005 1:58 PM


I agree that "intangibles" ultimately have to show up in the runs scored or runs allowed columns to exist. My point, however, is that there is no way to objectively measure the exact extent to which these intangibles contributed to those columns, nor do we know the extent to which that might have translated into more wins for a team than might have been expected. It is entirely possible that attitude or team chemistry might have contributed to their record in one-run games, which, in turn, would contribute to them winning more games than expected.

It is also entirely possible that it was just luck. That's the point. We don't really know, and we can't measure it.

Posted by Jeff | October 26, 2005 4:10 PM


//It is entirely possible that attitude or team chemistry might have contributed to their record in one-run games, which, in turn, would contribute to them winning more games than expected

That's where we'll have to agree to disagree. Wins and losses are made up of only one thing: runs. So if it exists, it shows up in runs one way or the other. As long as the theory looks at runs, the intangibles are included in the theory insofar as they contributed to a run (or the prevention thereof). If they didn't contribute to the scoring of a run or the prevention of a run, they're irrelevant.

The theory itself has been well proved WHAT the combination of runs scored vs. runs allowed should so. If there's a flaw in it, then -- sure -- that's worth looking at. But I haven't seen any evidence of it.

To me the whole intangible argument is like the Scott Podsednik evaluation. If the guy doesn't create runs, then he really isn't much of a reason for a team's success.

OTOH, someone in the newsroom here -- a ChiSox fan (hey, where have you all been for 45 years?) -- said that having Scott Podsednik on the team forced the White Sox to adopt a style of play tha was successful.

Hmmm... that's worth noodling on because, frankly, I'd take Carlos Lee over Scott Podsednik on my team in a second. And while ChiSox fans may have tired of waiting for the big homer to win games...hey...how've you won the last two World Series games?

Posted by Bob Collins | October 26, 2005 5:32 PM


Bob,
You're getting famous....MSNBC is picking you up. Cool. They have IMUS too, don't they?

Posted by cheryl Collins | October 27, 2005 9:59 PM


Yeah, I saw that. Kinda mangled the point (Sabermetricians are offended by the ChiSox. Huh?)
point spelled the name right.

Posted by Bob Collins | October 28, 2005 1:37 PM


Post a comment

The following HTML tags are allowed in your comments:
+ Bold: <b>Text</b>
+ Italic: <i>Text</i>
+ Link: <a href="http://url" target="_blank">Link</a>
Fields marked with * are required.


Comment Preview appears above this form upon pressing the "preview" button. Edit your comment and press "preview" again, until you are satisfied with your comment.

Your comment may not appear on the blog until several minutes after it was submitted.

October 2005
S M T W T F S
            1
2 3 4 5 6 7 8
9 10 11 12 13 14 15
16 17 18 19 20 21 22
23 24 25 26 27 28 29
30 31          


Master Archive

MPR News
Radio

Listen Now

On Air

As It Happens

Other Radio Streams from MPR

Classical MPR
Radio Heartland

Services