[Justin7]

If a team is projected to score 5 runs in a 9-inning game, their per inning scoring distribution might look like:
0: 70.2%
1: 15.7%
2: 7.5%
3: 3.6%
4: 1.7%
5: 1.3%

This changes somewhat for home teams that are tied in the bottom of the 9th, or extra innings.

Does anyone know what the scoring distribution for home teams in extra innings is? The game total and league (AL/NL) don't matter...

[tomcowley]

How's that water-carrying gig working out for you?

[Justin7]

I'm not quite sure what you mean, Tom.

[boondoggle]

tom is indicating that it is futile.

[mathdotcom]

Again I can see why you'd like to think about these issues but even if someone provides the answer to your question, I cannot imagine how it would be useful in practice. Basically I think the gap between the conditional and unconditional expectations of these variables is huge.

[Justin7]

Again I can see why you'd like to think about these issues but even if someone provides the answer to your question, I cannot imagine how it would be useful in practice. Basically I think the gap between the conditional and unconditional expectations of these variables is huge.

If you get this right, you can price alternate lines and live betting more efficiently.

[mathdotcom]

Yeah if you get it right, but if someone answered with such a distribution I don't see how you could properly use such aggregate numbers.
I imagine at best it would just give you an idea whether a more thorough analysis would bear fruit.

BOL

[Justin7]

Yeah if you get it right, but if someone answered with such a distribution I don't see how you could properly use such aggregate numbers.
I imagine at best it would just give you an idea whether a more thorough analysis would bear fruit.

BOL

thanks mathy. I did get the distribution. I don't really want to go into the methods here, but you can continue this in pm if you like.

[MonkeyF0cker]

If you get this right, you can price alternate lines and live betting more efficiently.

No. You really can't.

[Justin7]

No. You really can't.

Perhaps you can't.

[MonkeyF0cker]

Right. It's your money.

I'm just glad there are people like you out there.

Good luck.

[Dark Horse]

How can you judge something if you don't know the method? This may just be about a little piece of the puzzle, used to fill up small existing gaps. The underlying method may be far superior.

I wouldn't blindly trust data provided by others, however.

[mathdotcom]

Because the unconditional distribution is pretty much useless as it ignores a number of extremely relevant factors that need to be conditioned on.

[Justin7]

Because the unconditional distribution is pretty much useless as it ignores a number of extremely relevant factors that need to be conditioned on.

But, it is a component. If you already know the scoring distribution in a given half-inning that will be played to a full 3-outs (given whatever factors you choose to use to determine that distribution), and you have a half-inning distribution for a bottom 9/extra innings tied situation, you can construct distributions for other scenarios.

[Dark Horse]

nm.

[mathdotcom]

But, it is a component. If you already know the scoring distribution in a given half-inning that will be played to a full 3-outs (given whatever factors you choose to use to determine that distribution), and you have a half-inning distribution for a bottom 9/extra innings tied situation, you can construct distributions for other scenarios.

Bolded part is key, and those factors will probably give you a distribution much different than the one you asked for in the thread.

[mathdotcom]

nm.

Are you related to Wrecktangle?

[Justin7]

Bolded part is key, and those factors will probably give you a distribution much different than the one you asked for in the thread.

Thanks for the warning. Those are obviously other components of the puzzle, but I already have those at a satisfactory level.

[Dark Horse]

Are you related to Wrecktangle?

I just can't make up my mind if it's funny or sad that kids with so much to learn have already convinced themselves they know it all. Forgot the name of the kid that lost his entire bankroll. One of those hot shots from the sect.

[mathdotcom]

Yeah obviously a real hot shot

Just when I thought Wrecktangle was the only 'there is no wrong way to approach a question' type...

There ARE right and wrong answers to stats questions.

[MonkeyF0cker]

Thanks for the warning. Those are obviously other components of the puzzle, but I already have those at a satisfactory level.

This is such an absurd, posturing statement.

[mathdotcom]

Sometimes I have trouble sleeping worrying about J7's BR