April 08, 2008
Significance?
The Tigers lose to the Red Sox 5-0. At 0-7, it's time to seriously reconsider the notion that the Tigers are a playoff caliber team. A 90 win team has a .555 winning percentage. The chance of a .555 team going 0-7 is .0035, or 0.35%. In other words, playoff teams tend not to lose seven games in a row.
They're not barely over two runs per game, having scored 15 on the year. It's tough to blame this on the absence of Granderson, since Inge has hit well in his stead.
If the Tigers were in the AL West or NL Central they might not need to worry so much, but boy are Indians fans happy right now.
In 1962 the Mets, in their first year in existence, went 0-9 to start the season. The 2003 Tigers tied that record, so they have the pedigree to make a run at a new record.
Hey, is this a follow-up on the probability quiz? ;-)
First of all, your calculation works only if the games are probabilistically independent, which is very doubtful.
Second, suppose the games are independent events. The chance of a .400 team going 0-7 is (by my quick calculation eight times is .028, which is eight times as great as your .0035. So, it's not as if we had one hypothesis (the Tigers are bad) that makes the evidence very likely, and another (the Tigers are good) that makes it very unlikely. It's very unlikely in either case.
What we should believe now about the Tigers' prospects depends on our priors, of course. (I'm an orthodox Bayesian.) Clearly the losing streak is *some* significant evidence that Detroit isn't as good as we thought, but I wouldn't be writing them out of October yet.
Yes, James, which is why I said tend not to go 0-7.
Wait a second. Yes, the odds of a .555 team losing any particular 7 games in a row is .00345. But there are 156 different possible groups of 7 games in a 162 game season. So the odds are actually .539, or a little better than even that such a team will lose 7 games in a row sometime during the season.
On further review, I think I calculated that number wrong. I think it should be (1-.00345)*156, or .583.
David's number gives the odds of a .555 team losing its first 7 games (or last 7 games, or any particular set of 7 games). But that chance is very low for any team, even a terrible one.
David, yes, they *tend* not to go 0-7. But my point was: .400 teams also *tend* not to go 0-7.
Tor: but David gave the chance of a team starting the season 0-7. Of course, the chance that a team has *some* 0-7 stretch is much higher.
Tor,
This second calculation is correct. I'll have to go back and look to see how many team with 90 or more wins lost seven in a row. There's a certain amount of overlap in losing streaks, so I'm not sure that 156 is the right number to use for chances. Since even the worst teams are going to win at last 40 games, there really may only be 116 chances to lose seven games in a row. If that's the case, then the chance of a 90 win team losing seven in a row is about 0.33. I'm guessing if we do the research we'll see a number closer to that.
But for a .400 team, 0 wins is within the 95% confidence interval.
It's within the 95% interval? Hm. I got 2.8%.
Well, the 95% confidence interval goes from 2.5% to 97.5%.
Oh, of course. Sorry.
(We Bayesians generally frown upon confidence intervals, you know.)
I believe that the 1983 Orioles had two 7-game losing streaks on their way to the world championship.
Wow. That's true, and they won 98 games.
That Cabrera plays a classy 3B. He misplayed the throw from Polanco into an E and made 2 errors on the 1 play on the ball hit by Lugo - Lugo somehow gets a H and they rule an error on the throw.
Remember those predictions of a 1,000-run season for the Tigers offense? Well, after their mighty production of fifteen runs in seven games, they'll have to score 6.35 runs per game the rest of the way to make the thousand mark.
In a season where the average AL team is scoring a little over four runs a game - yes, scoring will probably perk up - the probability here has got to be pretty low.
Orioles 1983 lost games 37 - 43 and 105 - 111. In the second skid, they lost several to Chicago, which ended up with 99 wins (but lost to the O's in the ALCS.)
Of course the 1983 season is famous for the Lenn Sakata catching, Gary Roenicke playing third and John Lowenstein playing second game in which they beat Toronto 7 - 4. Tippy Martinez got the win picking off 3 baserunners. And Lenn Sakata hit the winning 3 run homer in the 9th.
1983 was an odd (but thrilling) season for the O's.
This is a little more basic, but I did some 'back of the napkin' math this morning and here are the key numbers I came up with...
Over the last 10 years, the wildcard team won 95.5 games (high - '01 A's: 102, Low - '00 Mariners: 91). To win 95 games you need to have a 58.6% winning percentage over 162 games. If you drop the first seven games, you now need to win a rate 61.6% to hit the 95 win mark. Now obviously you don't need to win 95 to get in, and you could still miss the playoffs with 96, but it's still a pretty reasonable number to look to as a goal.
To put the 61.6% into perspective, the last team to hit a winning percentage that high over the course of a season was the '05 Cardinals (100 wins - 61.7%). And the last AL team to do it was the '04 Yankees with 101 wins and a 62.3%.
So basically, it's not unrealistic for a team to win at that percentage, but it doesn't happen every year either.