Schoenfield's ESPN Sweet Spot featuring the O's today

[Dark Helmet]

The grey area is that outlined by many posters on here who utilize pythag correctly. The grey area is probability .

Ok, but at what point does probability change. When does it start to take into account that Tillman, Gonzalez etc. are actually better than who they replaced?

[Lucky Jim]

Ok, but at what point does probability change. When does it start to take into account that Tillman, Gonzalez etc. are actually better than who they replaced?

You can take that into account at any time. A pythagorean calculation isn't going to take into account anything other than runs, however.

[SilentJames]

This makes no sense. Literally, none.

I worded it poorly.

If you go fiddling with the run diff's of the yankee, or looking at the one-run record of the reds, it wont tell you anything because their seasons are following Pythagorean norms.

But the orioles clearly are not following Pythagorean norms. This entire team is a statistical outlier. So saying that team x is this, or team y is that really is meaningless.

The orioles are going to look very off by comparison on a lot of levels because their season is not following Pythagorean norms. Neither is it following statistical norms, or what people perceive to be statistical norms anyway.

Sent from my DROID BIONIC using Tapatalk 2

[SilentJames]

You can take that into account at any time. A pythagorean calculation isn't going to take into account anything other than runs, however.

Exactly. And it sort of assumes that all of the runs are more or lesser spread out evenly over time. Whereas the orioles have lost big bunches of runs in multiple two-four game spurts. Therefore, throwing off their overall run totals.

Sent from my DROID BIONIC using Tapatalk 2

[Lucky Jim]

I worded it poorly.

If you go fiddling with the run diff's of the yankee, or looking at the one-run record of the reds, it wont tell you anything because their seasons are following Pythagorean norms.

But the orioles clearly are not following Pythagorean norms. This entire team is a statistical outlier. So saying that team x is this, or team y is that really is meaningless.

The orioles are going to look very off by comparison on a lot of levels because their season is not following Pythagorean norms. Neither is it following statistical norms, or what people perceive to be statistical norms anyway.

Sent from my DROID BIONIC using Tapatalk 2

Everyone knows that the Orioles have been good in one-run and two-run games. That's not news. It's one particular way teams outperform their pythag. If you want to say that the enormous gulf between their pythag and their record is based on some skill, you need to evaluate each component in the context of what other teams do, and see how it holds up. There's no escaping that. Your rationale is not solid.

[Dark Helmet]

You can take that into account at any time. A pythagorean calculation isn't going to take into account anything other than runs, however.

Oh ok, thanks. Seems like it's a stat thats too black and white too me then, IMO. I just don't see a stat as a good guideline to go by, if it doesn't take into account different variables. I have to say, this has been an interesting thread, whether or not you agree with these types of formulas.

[LookinUp]

Isn't that essentially what people are doing when they cite our RD as a reason for an impending collapse? They aren't throwing concrete numbers out there but they're sure begin suggestive.

Yes, if by people you mean ESPN and SI. People on this site aren't doing that.

Through the haze of these two threads, I have come to three conclusions:

1. All things being equal, the pythag is generally a good predictor of outcomes.

2. Changes in the composition of a roster/streakiness of players have the potential make the pythag a less valuable* predictor of future success.

3. The pythag does not determine outcomes. It just assigns probabilities.

*the degree to which factors like changes in the composition in a roster affect how well the pythag predicts future success will greatly vary depending on the amount of value those changes give/take. In this regard, the pythag haters could actually use this function to make a case for future success based on the changes in the composition of our roster. Unfortunately, I haven't seen the data, but I do think the pythag's prediction for our future success would be less than we wish for, and probably less than we need to make the playoffs. This data point leads me to hope that DD continues to upgrade the roster.

[DrungoHazewood]

So its either right or wrong, with probable being the gray area. But it doesnt take unlikely into account? Im only asking because id be lying if i said i completely understood how the pythag works.

Maybe someone needs to give the 30-second summary:

A team's runs scored and runs allowed are good indicators of how often a team wins. If you use the formula (Runs^2)/(Runs^2 + Allowed^2) that's a pretty fair approximation of winning percentage. In most seasons most teams will be within 5 games of this prediction over a full season. For example, in 2011 the Padres and Royals under-shot their prediction by 7 and 8 games, and were the only MLB teams more than 6 off. 22 of the 30 teams were within +/- four games.

In all of recorded history (i.e. retrosheet's score database that goes back, currently, to 1901), there is only one team that overshot their projection by 14 games (the 1905 Tigers), and only a very small handful that have been over by 10 or more games.

Basically, it's a rule of thumb, a simple guideline. Since a sizeable majority of teams win and lose according to this baseball Pythagorean Theorem it is often useful to observe the outliers. They will usually fall back in line after a long enough sample.

There appears to be some correlation between good bullpens and overachieving your Pythag. But by no means is that a hard-and-fast rule.

[LookinUp]

Ok, but at what point does probability change. When does it start to take into account that Tillman, Gonzalez etc. are actually better than who they replaced?

I'm pretty sure the poff% posted on ESPN takes more into account than the pythagorean. W/L records and games remaining are factored in...I think. I also think the O's poff% has improved over the last week despite a relatively small change in our pythag.

[SilentJames]

Everyone knows that the Orioles have been good in one-run and two-run games. That's not news. It's one particular way teams outperform their pythag. If you want to say that the enormous gulf between their pythag and their record is based on some skill, you need to evaluate each component in the context of what other teams do, and see how it holds up. There's no escaping that. Your rationale is not solid.

You could dissect any team's run differential like I did and discover anomalies but if the team is more or less following their expected Pythagorean record then it really wont tell you much. Because after the dissection they will still be, more or less, on the same track.

But teams like the orioles and red sox who are either drastically over/under performing despite the numbers on the surface digging deep can reveal the trends.

After spending a bit looking through these numbers and sketching it all out, I will concede that the orioles were very lucky in June, and the 23 game stretch (game 67-90) is (june into july) is the biggest culprit and royally threw off the team's overall numbers.

The offense was dead

The rotation was in complete turmoil

The bullpen was cracking a bit

The question is, have the last 20 games been enough to convince you that the team has come out of that?

Sent from my DROID BIONIC using Tapatalk 2

[Dark Helmet]

Maybe someone needs to give the 30-second summary:

A team's runs scored and runs allowed are good indicators of how often a team wins. If you use the formula (Runs^2)/(Runs^2 + Allowed^2) that's a pretty fair approximation of winning percentage. In most seasons most teams will be within 5 games of this prediction over a full season. For example, in 2011 the Padres and Royals under-shot their prediction by 7 and 8 games, and were the only MLB teams more than 6 off. 22 of the 30 teams were within +/- four games.

In all of recorded history (i.e. retrosheet's score database that goes back, currently, to 1901), there is only one team that overshot their projection by 14 games (the 1905 Tigers), and only a very small handful that have been over by 10 or more games.

Basically, it's a rule of thumb, a simple guideline. Since a sizeable majority of teams win and lose according to this baseball Pythagorean Theorem it is often useful to observe the outliers. They will usually fall back in line after a long enough sample.

There appears to be some correlation between good bullpens and overachieving your Pythag. But by no means is that a hard-and-fast rule.

Thanks for the summary. I get where it's coming from. You give up more runs than you score, common sense says you lose. It just seems like it bunches up all the numbers over the season, to come to a conclusion. Just off the top of my head, we've had three or four games we lost by 7+ runs. Should the probabilities say we're a bad team, because of four or five very bad games? The Giants and Dodgers are at +17 and +18 I think. A few bad games like we've had before, and they could be in the negative. But is it probable that both will miss the playoffs?

[BRobinsonfan]

It's usually at this point someone brings in the whole we've changed the culture of losing "argument." :roll eyes:

[Dark Helmet]

It's usually at this point someone brings in the whole we've changed the culture of losing "argument." :roll eyes:

In my case, I don't base W/L on stats alone. But, I don't disregard those who do. I think the culture around the team does have an affect to a degree. Jim Palmer gives alot of credit to Earl Weaver in that aspect.

[Pickles]

I haven't read the entirety of this thread so forgive me if this has already been pointed out but in no way, shape, or form, does Pythag claim to be predicative. It doesn't make any claims in that regard. It's does not say, "Since the Orioles have been outscored by 59 (or whatever) runs to this point in the season, they will continue to be outscored at such a rate." It simply is saying, "Given the runs differential to this point in the season the Orioles record should be close to X."

If the Orioles have improved internally, specifically w the additions to the starting rotation, that should be reflected in the run differential going forward, and the Orioles record will start to come closer to their expected record.

[CA-ORIOLE]

I haven't read the entirety of this thread so forgive me if this has already been pointed out but in no way, shape, or form, does Pythag claim to be predicative. It doesn't make any claims in that regard. It's does not say, "Since the Orioles have been outscored by 59 (or whatever) runs to this point in the season, they will continue to be outscored at such a rate." It simply is saying, "Given the runs differential to this point in the season the Orioles record should be close to X."

If the Orioles have improved internally, specifically w the additions to the starting rotation, that should be reflected in the run differential going forward, and the Orioles record will start to come closer to their expected record.

No it does not.

simply is saying, "Given the runs differential to this point in the season the Orioles record should be close to X."

Correct.