Something for the mathematicians

Discuss anything related to using the program (eg. triggered betting tactics)

Moderator: 2020vision

Something for the mathematicians

Postby doris_day » Wed Jun 16, 2010 12:14 pm

This forum has been so useful to me over the years, I'm going to present those interested with a concept that I've found to be very valuable. I'm not going to say much about it other than what it is and how it can be calculated. Its then up to those who want to experiment with it to do just that.

The concept is 'Entropy' and its a measure of the 'spread' of prices in a market.
Say the probability of winning (as measured by their price) of each market participant is x then Entropy is calculated thus: =Sum(xlogx)

Go play with it. Its more useful than you might first think......:)
'He was looking for the card so high and wild he'd never need to deal another' - Leonard Cohen
User avatar
doris_day
 
Posts: 968
Joined: Fri Nov 02, 2007 12:34 am

Postby osknows » Wed Jun 16, 2010 1:07 pm

this is a bit 'random' :)
User avatar
osknows
 
Posts: 946
Joined: Wed Jul 29, 2009 12:01 am

Postby jokerjoe » Wed Jun 16, 2010 2:47 pm

Thanks, that's very useful. I was trying to figure out how to measure the spread of prices the other day, what I came up with is not nearly as neat!
User avatar
jokerjoe
 
Posts: 122
Joined: Wed May 09, 2007 12:00 pm

Postby alrodopial » Wed Jun 16, 2010 5:45 pm

So the entropy of two events from odds:
3.5 -> 28.6%
1.23 -> 81.3%

is 28.6*log(28.6)+81.3*log(81.3) = 28.6*1.456+81.3*1.91=198.8 ?

or the probability is usedas 0.286 ?
alrodopial
 
Posts: 1384
Joined: Wed Dec 06, 2006 9:59 pm

Postby doris_day » Wed Jun 16, 2010 6:10 pm

0.286
'He was looking for the card so high and wild he'd never need to deal another' - Leonard Cohen
User avatar
doris_day
 
Posts: 968
Joined: Fri Nov 02, 2007 12:34 am

Re: Something for the mathematicians

Postby xraymitch » Sat Jun 19, 2010 7:18 pm

doris_day wrote:This forum has been so useful to me over the years, I'm going to present those interested with a concept that I've found to be very valuable. I'm not going to say much about it other than what it is and how it can be calculated. Its then up to those who want to experiment with it to do just that.

The concept is 'Entropy' and its a measure of the 'spread' of prices in a market.
Say the probability of winning (as measured by their price) of each market participant is x then Entropy is calculated thus: =Sum(xlogx)

Go play with it. Its more useful than you might first think......:)


Thankyou for that Doris it is what I am looking for. Give me a month or so and I will report back here.
xraymitch
 
Posts: 410
Joined: Wed Jun 25, 2008 7:06 am
Location: UK

Postby Lewis » Thu Nov 04, 2010 10:47 pm

Excuse my ignorance but how do I write "xlogx" in excel.

I am straining back to remember about logs and remember there is normally a base in addition to the number.

[Thanks for this looks like the 'entropy' could prove useful]
Lewis
 
Posts: 50
Joined: Sun Mar 28, 2010 10:15 pm

Postby Lewis » Thu Nov 04, 2010 10:50 pm

Actually....

=X(ln(x)) does this look right? natural log is coming back to me know
Lewis
 
Posts: 50
Joined: Sun Mar 28, 2010 10:15 pm

Postby Royville » Sat Nov 06, 2010 12:59 pm

Hi,

I saw this thread when it first came out and I couldn't get my head around it then!

Would some kind person like to explain what this is all about please - in layman's terms.

I'm very intrigued and would love to know/learn more.

Regards
Royville
Royville
 
Posts: 65
Joined: Wed Apr 21, 2010 4:50 pm
Location: UK Midlands

Postby doris_day » Mon Nov 08, 2010 5:00 pm

The explanation is in the first post really.

Entropy quantifies the 'spread' of prices in a market. For instance if you had five runners and they were all priced around 6.0 you would have one number and if in another market you had a short priced favourite of, say, 1.8 and all the rest around 12.0 then you'd get a totally different number, both reflecting the respective 'spread' of prices. Its sometimes valuable to be able to quantify the spread of prices as some algorithms work better in one type of market rather than another.
'He was looking for the card so high and wild he'd never need to deal another' - Leonard Cohen
User avatar
doris_day
 
Posts: 968
Joined: Fri Nov 02, 2007 12:34 am

Postby Royville » Mon Nov 08, 2010 6:01 pm

Thanks for the reply Doris.

I don't suppose you could simplify it any more than that really! :?

Still lost though...

Royville
Royville
 
Posts: 65
Joined: Wed Apr 21, 2010 4:50 pm
Location: UK Midlands


Return to Discussion

Who is online

Users browsing this forum: No registered users and 108 guests

Sports betting software from Gruss Software


The strength of Gruss Software is that it’s been designed by one of you, a frustrated sports punter, and then developed by listening to dozens of like-minded enthusiasts.

Gruss is owned and run by brothers Gary and Mark Russell. Gary discovered Betfair in 2004 and soon realised that using bespoke software to place bets was much more efficient than merely placing them through the website.

Gary built his own software and then enhanced its features after trialling it through other Betfair users and reacting to their improvement ideas, something that still happens today.

He started making a small monthly charge so he could work on it full-time and then recruited Mark to help develop the products and Gruss Software was born.

We think it’s the best of its kind and so do a lot of our customers. But you can never stand still in this game and we’ll continue to improve the software if any more great ideas emerge.