Multinomial Logit Models

[Dark Target]

Either you post on PTT or my phone is bugged

Can't think of any other way you knew I was there.

[Ian Meyers]

Originally Posted by Dark Target

I can gaurantee that Multinomial Logistic Regression is EXTREMELY suited to calculating win probabilities.

I can also gaurantee that other than the first 2 years, Benter (nor Woods and the rest) never struggled with implementing it.

Thanks for your input. I was 98% sure it's what most (all?) teams are using. I know who the teams are, some through a direct relationship, others by reputation. I could never ask how they do what they do and no one would tell me if I did ask.

I don't expect to be able to build a profitable black box model but the 'rating' a MNL logit generates will help my handicapping in other ways.

Happy New Year to all.

[sjk]

I was hoping there would be more discussion of the relevant variables and calculations.

[Native Texan III]

Originally Posted by Dark Target

I can gaurantee that Multinomial Logistic Regression is EXTREMELY suited to calculating win probabilities.

I can also gaurantee that other than the first 2 years, Benter (nor Woods and the rest) never struggled with implementing it.

This is from someone that posts the spelling wrong twice. (guar·an·tee). You cannot guarantee anything except that you have no clue about the topic. If you did, you would understand the huge drawbacks involved and why Benter abandoned it. The topic is decades old now and things have long moved on.

Benter actually spelled out in his papers the huge implementation issues he had in the early days - they nearly parted company over it. Of course, you have never read, let alone understood, the papers.

[Dave Schwartz]

This is from someone that posts the spelling wrong twice. (guar·an·tee). You cannot guarantee anything except that you have no clue about the topic. If you did, you would understand the huge drawbacks involved and why Benter abandoned it. The topic is decades old now and things have long moved on.

Mr. Texan,

Well, I can GUARANTEE that I understand the topic very well and Mr. Dark Target is absolutely right on the money.

Beyond that my NDAs will kick in and I'll say no more.


Regards,
Dave Schwartz

[Dark Target]

As with Dave i will not say no more on the topic other than to say, my trusty Google Chrome did not pick up my spelling mistake... twice

I've read the paper in question back to front several times.

1. Benter did not abandon it due to its draw backs, he "left" Hong Kong for other reasons which i'm sure with a bit of investigation you can find out why.

2. Bill and Alan did not nearly part ways because of implementation issues. They did part ways due to disagreements on funding etc...

Have a nice New Year.