Now that Ive found an edge...How do I apply it to Kelly?

**AgainstAllOdds** · 01-19-09, 12:40 AM

(Assuming the numbers above are correct)...then my expected growth is 0.0973% and my expected profit would be 0.1901% and I should be betting 8.7200% of my bankroll.

Is this correct?

**smitch124** · 01-19-09, 01:28 AM

right on the button AAO...

**durito** · 01-19-09, 09:30 AM

How did you come up with the original numbers?

**AgainstAllOdds** · 01-19-09, 11:26 AM

Read this article...http://forum.sbrforum.com/handicappe...tml#post250260 And this...http://forum.sbrforum.com/handicappe...tml#post258665 as well as some others and then started using a calulator, spreadsheets and some of Ganch's/SBR Tools.

**durito** · 01-19-09, 11:39 AM

I'm well familiar with those articles.

I'm just curious as to how you determined that Syracuse has a 81.74% win chance.

**donjuan** · 01-19-09, 12:16 PM

I'm fairly certain AAO used the half point calculator or a similar tool to come up with that. I would advise against this.

**AgainstAllOdds** · 01-19-09, 02:43 PM

Originally posted by donjuan

I'm fairly certain AAO used the half point calculator or a similar tool to come up with that. I would advise against this.

Yea I used that and the ML-Spread Converter...After I read and started trying doing some of them on my own...I found the Tools do pretty much the same thing in half the time I was doing it.

Is there a really bad flaw in this?

**durito** · 01-19-09, 04:15 PM

Originally posted by AgainstAllOdds

Is there a really bad flaw in this?

Yes, you will end up broke. Though in this particular case had you happened to play the Pitt money line at -400 I'd say you have a pretty good play based on the current market odds.

The ml/spread converter is a good tool to estimate probabilities in games with smaller <-4 or so point spreads. But I wouldn't take it for anything more than that and certainly wouldn't bet off it. There are however money line/point spread mis-pricings that you can find and profit from if you look hard enough.

**donjuan** · 01-19-09, 05:24 PM

Yea I used that and the ML-Spread Converter...After I read and started trying doing some of them on my own...I found the Tools do pretty much the same thing in half the time I was doing it.

Is there a really bad flaw in this?

The problem with it is that it assumes the push percentage of the 2 is the same for both an 8 point favorite as it is for a 2 point favorite when doing the spread/moneyline conversion.

**B1GER1C828** · 01-19-09, 08:18 PM

Where can you find the tools? i tried looking but no luck. Can anyone help me out?

**durito** · 01-19-09, 08:20 PM

SBR Forum - Sports Betting Talk - Sportsbook Review Forum

http://www.sbrforum.com/Betting+Tools/Home.aspx

Sports betting and handicapping forum: discuss picks, odds, and predictions for upcoming games and results on latest bets.

**roasthawg** · 01-20-09, 07:29 PM

Keep it simple...just bet 1/40 of your bankroll on every game you play.

**jtuck** · 01-20-09, 07:33 PM

Originally posted by roasthawg

Keep it simple...just bet 1/40 of your bankroll on every game you play.

Awful advice. If you are willing and able to take the time to learn something that will maximize bankroll growth, why not?

**roasthawg** · 01-20-09, 07:39 PM

Originally posted by jtuck

Awful advice. If you are willing and able to take the time to learn something that will maximize bankroll growth, why not?

Because it's all an estimation anyways...you have to estimate your edge to use kelly and there's no way of knowing what percentage of your games you'll win. At 53.5% and -110 odds kelly is close enough to 1/40 of your bankroll that it becomes moot.

**donjuan** · 01-21-09, 01:08 AM

you have to estimate your edge to use kelly and there's no way of knowing what percentage of your games you'll win.

If you don't have a reasonable estimate of your edge, why do you think you have any edge?

**roasthawg** · 01-21-09, 11:39 AM

Originally posted by donjuan

If you don't have a reasonable estimate of your edge, why do you think you have any edge?

Sure...but a "reasonable" estimation could be between 53% and 55%. The difference in kelly betting is enormous in that 2%.

The main purpose of kelly in relation to sports gambling as I see it is to compound your winnings by ensuring that you bet progressively larger amounts the more you win. Betting 1/40 of your bankroll is an easy way to do this.

**durito** · 01-21-09, 12:02 PM

So you have two bets at -110

One with an estimated win of 53%-55%

and one with an estimated win of 60-62%

You want to bet the same on both?

**roasthawg** · 01-21-09, 12:13 PM

Originally posted by durito

So you have two bets at -110

One with an estimated win of 53%-55%

and one with an estimated win of 60-62%

You want to bet the same on both?

No, you want to recalculate your edge on the one you have at 60-62%, lol...you're probably wrong!

**Peep** · 01-21-09, 12:43 PM

Originally posted by roasthawg

Because it's all an estimation anyways...you have to estimate your edge to use kelly and there's no way of knowing what percentage of your games you'll win. At 53.5% and -110 odds kelly is close enough to 1/40 of your bankroll that it becomes moot.

Well, I am with you anyways roasthawg.

I use 1/50 of my bankroll.

I really don't know how to calculate an edge, only know if I have one by whether or not my bankroll grows.

I could come up with a THEORETICAL edge. But I wouldn't trust it, too many variables for my tastes. Nice to say that 6 pt favorites in NBA win 73.6% of their games. So if I get -230 I have an edge.

But how big a one? Totals, Home/Road, days off, playing styles all factor in. But to what extent?

KISS-----Keep It Simple Stupid!