Well at last I am starting to get my charting application together for trading on the betting exchanges. (Only a year late)
It has some features that have already earned money on the exchanges and I will be working further on it to enhance the risk management and provide trend indicators.
At the moment it is really only at alpha level but if anyone is interested then do feel free to comment.
Thursday, May 27, 2010
Thursday, May 13, 2010
Predicting The Score
I have been spending quite a bit of time at the bet your life blog over the past couple of weeks mainly in the chatroom.
Talking to other traders while trading adds a good dimension to the activity and the guys up there are good fun and willing to talk about some ideas.
Which brings me onto this post.
The other night I was discussing the Correct Score Markets and I decided to show them my price predictor for this market.
It's a very simple excel sheet where you enter what you believe to be the average goals for the home and away team in the table at the top.
While I just enter the average of the last 6 home matches for the home and away teams you could look at other factors that could modify the average.
i.e. If a striker is injured then deducted whatever goals he has scored from the average goal calculation.
You could also weight the goals scored by looking at who they were scored against.
i.e. A goal against Manchester United is worth 1.2 while a goal against Hull is only worth 0.7 but that is getting into a Rateform approach which I will talk about more later.
Take it for a drive and compare it to the prices on the exchanges.
The other night I was discussing the Correct Score Markets and I decided to show them my price predictor for this market.
It's a very simple excel sheet where you enter what you believe to be the average goals for the home and away team in the table at the top.
The second table takes those averages and displays the probability of the home and the away team scoring a specific amount of goals based on a Poisson distribution.
The third table then takes the probabilities of specific goal counts and merges them together to create the odds for a specific scoreline.
While I just enter the average of the last 6 home matches for the home and away teams you could look at other factors that could modify the average.
i.e. If a striker is injured then deducted whatever goals he has scored from the average goal calculation.
You could also weight the goals scored by looking at who they were scored against.
i.e. A goal against Manchester United is worth 1.2 while a goal against Hull is only worth 0.7 but that is getting into a Rateform approach which I will talk about more later.
Take it for a drive and compare it to the prices on the exchanges.
I guarantee you will be surprised by the correlation between the predicted values from the excel sheet and the prices offered.
You will notice however that the most probable outcome is under priced on the exchange which could be an opportunity.....
Tuesday, February 9, 2010
The Initial Investigation into Fuzzy Logic
OK so I started looked into the fuzzy model offered in the paper from the previous post and I found that I wasn't happy with it.
The main problem I see is the approach of using big win, little win, draw, little loss big loss based on goals scored misses the key point for me:
A big win against Burnley cannot be as significant as a big win against Chelsea.
Researching the problem I found it had been approached in a strategy called RateForm.
The rateform model given in that article is a generalization of the ELO system used in chess and I was lucky enough to have some old code for calculating chess ratings based on an article by Vincent Bisset of Malahide Chess Club on the Irish Chess Union website.
There are also some recommedations as to how this system could be tweaked to cater for home bias in this article from chessbase.
So after installing Visual Studio 2010 and taking Silverlight for a spin I now have an application that rates teams performance over time based on results against the opposition they play.
The screenshot shows the output for the past year in hard hitting Rugby League Football.
Now I am going to see how this plays into the fuzzy model.
The main problem I see is the approach of using big win, little win, draw, little loss big loss based on goals scored misses the key point for me:
A big win against Burnley cannot be as significant as a big win against Chelsea.
Researching the problem I found it had been approached in a strategy called RateForm.
The rateform model given in that article is a generalization of the ELO system used in chess and I was lucky enough to have some old code for calculating chess ratings based on an article by Vincent Bisset of Malahide Chess Club on the Irish Chess Union website.
There are also some recommedations as to how this system could be tweaked to cater for home bias in this article from chessbase.
So after installing Visual Studio 2010 and taking Silverlight for a spin I now have an application that rates teams performance over time based on results against the opposition they play.
The screenshot shows the output for the past year in hard hitting Rugby League Football.
Now I am going to see how this plays into the fuzzy model.
Thursday, January 21, 2010
An interesting article on prediction models
I found this PDF on how to predict the result of association football matches this week.
You can download a copy here.
Its a great insight into how you might go about implementing fuzzy models and refine the results.
The basic premise is that you categorize results into big win, little win, draw little loss and big loss.
You then group these outputs with the results of the previous matches between the actual teams that are due to play and that gives a prediction model.
Smart tuning methods are then implemented on top of this model and your away in a hat!!
Although the whole point of the approach is to simplify analysis I think it could be worth incorporating some home and aways as weighting.
But what is REALLY interesting in this paper is the last paragraph on the first page.
"In the practice of prediction making the football experts and fans usually make good decisions using simple reasonings
on the common sense level."
This gave me the idea that it could be worth investigating if the model could be used to predict betting trends instead of actual results.
Watch this space!
You can download a copy here.
Its a great insight into how you might go about implementing fuzzy models and refine the results.
The basic premise is that you categorize results into big win, little win, draw little loss and big loss.
You then group these outputs with the results of the previous matches between the actual teams that are due to play and that gives a prediction model.
Smart tuning methods are then implemented on top of this model and your away in a hat!!
Although the whole point of the approach is to simplify analysis I think it could be worth incorporating some home and aways as weighting.
But what is REALLY interesting in this paper is the last paragraph on the first page.
"In the practice of prediction making the football experts and fans usually make good decisions using simple reasonings
on the common sense level."
This gave me the idea that it could be worth investigating if the model could be used to predict betting trends instead of actual results.
Watch this space!
Tuesday, May 26, 2009
The hard righthand side!
In response to Leon on dontgiveupyourjob (A very interesting blog if there ever was one!!)
I thought I would post my opinion on his latest question.
"Does this class as a swing trade?"
In my view if he got in exactly where he positioned A in the diagram then he took a speculative position.(A Bet NOT a trade)
If he got in four candles further on from the A when the Selection had broken the resistance of 4.7/8 as indicated at the start of the chart along with a stop loss around the previous 4.4 bottom with an initial 3-1 ratio sell at 6.0 then I would have said he was in a fairly risk free swing/scalp whatever you want to call it but at least it would be a trade.(Although I would have a question as to what led him to believe it would make 6 or even the 6.6 heights but that is for another day)
This may all sound like mumbo jumbo but if you look at the graph below I think it illustrates the point very well.... with the hard right handside in place what in the market is telling you to buy or sell?
In my humble opinion nothing...
There is nothing in that picture that gives me any reason to believe I should do something.
I am not saying that there is anything wrong with watching for news and predicting how the market will react but from my perspective this is closer to studying form and other fundamentals than it is to technical analysis.
Monday, May 11, 2009
Concurrent Programming Agents
The traditional approach to devising possible trades is to analyse a set of charts to identify trends and then place trades placed on that trend.
The skill in this is analysing the chart data to see if trends exist and how much they are potentially worth and how much risk is involved.
At the moment Microsoft are helping to tackle the problem by providing a Domain Specific Language to help deal with concurrent programming. Thus enabling access to the power required to perform the market analysis.
The Axum Parallel Programing Model helps deal with this and is worth keeping an eye on.
The skill in this is analysing the chart data to see if trends exist and how much they are potentially worth and how much risk is involved.
At the moment Microsoft are helping to tackle the problem by providing a Domain Specific Language to help deal with concurrent programming. Thus enabling access to the power required to perform the market analysis.
The Axum Parallel Programing Model helps deal with this and is worth keeping an eye on.
Friday, May 8, 2009
An interesting overview of the internals at Betfair
As well as trading I also write software for a living.
If you use an API at a betting exchange its always worth while understanding how they have put it together.
You can get a very good over view of the betfair exchange at this address
There is a lot of detail here but at the very least it should make you curse less when the exchange is slow because you will understand the unique challenges they face.
If you use an API at a betting exchange its always worth while understanding how they have put it together.
You can get a very good over view of the betfair exchange at this address
There is a lot of detail here but at the very least it should make you curse less when the exchange is slow because you will understand the unique challenges they face.
Subscribe to:
Posts (Atom)