SI.com: Why the O's Post Season Dream is all but Impossible

[Lucky Jim]

For all the talk about run differentials lately, yesterday's game doesn't fit into that argument nicely because we trailed by 5 runs twice and came back to win by one run. Had we geeked out another 1-run game and held it for 5-6 innings, I could see that being another factor of how it can't last but last night was an example of how this team can overcome big deficits to win games.

This doesn't really matter re: pythag, Sean. This team does a pretty good job of overcoming big deficits. On the flip side, everyone is saying that we get blown out a lot. Maybe it's just that we tend to have big deficits a lot and have managed to overcome a few?

[callahan09]

Sorry in advance for taking a mathematical approach to this, but I'm a mathematician. We all know that pythag gives you an average expectation of wins vs. losses. And as some have already pointed out without wading into the numbers, the O's exceeding their pythag record is a result of inconsistent starting pitching and a really good pen that allows them to win a lot of close games.

Here's a breakdown of the run distributions both for and against for the Orioles this season, not including tonight:

Offense Defense

Runs Scored Count Percentage Runs Allowed Count Percentage

0 8 7.34% 0 6 5.50%

1 10 9.17% 1 15 13.76%

2 17 15.60% 2 15 13.76%

3 18 16.51% 3 11 10.09%

4 12 11.01% 4 13 11.93%

5 14 12.84% 5 9 8.26%

6 10 9.17% 6 16 14.68%

7 5 4.59% 7 5 4.59%

8 7 6.42% 8 6 5.50%

9 3 2.75% 9 5 4.59%

10 3 2.75% 10 2 1.83%

11 1 0.92% 11 1 0.92%

12 1 0.92% 12 1 0.92%

13 1 0.92%

14 2 1.83%

15 0 0.00%

16 0 0.00%

17 0 0.00%

18 0 0.00%

19 1 0.92%

Sorry the columns are screwy. Anyway, if you take those numbers and do the calculations, given the Orioles performance this season, you would expect them to outscore the opposition in 42.19% of games. By the same token, you'd expect the opposition to outscore them in 47.86% of games. This leaves roughly 10 % of the games where you'd expect the Orioles to score the same number of runs as the opposition. If you just say the O's are equally likely to win these games as the opposition, then you get an expected win percentage of 47.16%, about 3% greater than their pythag record. So that suggests the pythag record is at least a bit harsh. So looking at the differentials on a game by game basis more closely:

Run Diff Count Percentage

-12 2 1.83%

-11 1 0.92%

-10 0 0.00%

-9 2 1.83%

-8 0 0.00%

-7 4 3.67%

-6 3 2.75%

-5 7 6.42%

-4 7 6.42%

-3 8 7.34%

-2 11 10.09%

-1 6 5.50%

0 0 0.00%

1 21 19.27%

2 16 14.68%

3 6 5.50%

4 2 1.83%

5 3 2.75%

6 8 7.34%

7 0 0.00%

8 1 0.92%

9 1 0.92%

So basically the O's have been about 4 times as likely to win by 1 as to lose by 1 and one and 50% more likely to win by 2 than to lose by 2. So splitting that extra 10% evenly doesn't really seem adequate for analyzing this team. I know I haven't really any major conclusions, just thought some, like me, would be interested in this analysis. I'd love to see a comparison of this type of breakdown on a league wide and historical basis, but I just don't have the time. Definitely seems like the O's are good at getting a decent start and winning tight games. And not so good when they get behind big, likely early in the game from a poor start by one of their starters.

You can get all fancy with the numbers and the calculations if you want, but the most meaningful way to come up with an expected percentage of games that the Orioles will outscore the opposition is to look at their win loss record. That is the up-to-date, real-world, actual percentage of games that they outscored the opposition, and that's a good enough predictor for me.

[square634]

You can get all fancy with the numbers and the calculations if you want, but the most meaningful way to come up with an expected percentage of games that the Orioles will outscore the opposition is to look at their win loss record. That is the up-to-date, real-world, actual percentage of games that they outscored the opposition, and that's a good enough predictor for me.

You can get all fancy with the numbers and the calculations if you want, but the most meaningful way to come up with an expected percentage of coin flips that will come up as heads is to look at the current percentage of heads after 110 flips. That is the up-to-date, real-world, actual percentage of flips that wound up as heads, and that's a good enough predictor for me.

[Loophole]

For all the talk about run differentials lately, yesterday's game doesn't fit into that argument nicely because we trailed by 5 runs twice and came back to win by one run. Had we geeked out another 1-run game and held it for 5-6 innings, I could see that being another factor of how it can't last but last night was an example of how this team can overcome big deficits to win games.

The only reason yesterday's game doesn't fit is because it was extra innings at home, and there's only one win situation (walk-off home run with 1+ men on) that would result in more than a one-run lead.

The fact that we came back doesn't matter; a better team with a better run differential wouldn't have allowed a big deficit in the first place. Other teams with ~60 wins at this point win games 7-0 or 7-1 or 7-2, to go along with their one-run wins. This team rarely does that, and you can't rely on miraculous comeback wins all season long.

[Moose Milligan]

IMO, the goal is and always has been .500. Forrest Gumping our way into a playoff spot would be awesome...it's been a fun ride but I'm with Drungo, it really can't last.

If it somehow does, it'll be incredibly enjoyable. But still, a .500 season would be incredibly enjoyable, too.

[Flash- bd]

You can get all fancy with the numbers and the calculations if you want, but the most meaningful way to come up with an expected percentage of coin flips that will come up as heads is to look at the current percentage of heads after 110 flips. That is the up-to-date, real-world, actual percentage of flips that wound up as heads, and that's a good enough predictor for me.

Are you serious with this?

[weams]

Are you serious with this?

To infinity, and beyond.

[Lucky Jim]

Are you serious with this?

I can almost guarantee that he's not.

[Loophole]

Are you serious with this?

Obviously coins aren't a good analogy for ball teams, but they sort of are for one-run games, seeing as they have an expected probability of 50% with perhaps a slight tick in either direction if a team is very good or very bad.

[CmonOs]

Whatever works for you. I don't necessarily disagree. I was really hoping to stumble upon the secret to 2012 Orioles Magic.

[square634]

I can almost guarantee that he's not.

Sorry, made this post and then haven't been reading the board closely for a while. I guess I'm not sure what is meant about whether I was serious.

To clarify: my point is that what has happened before is not the best prediction of what will happen next.