Discussion around sports betting analysis

Division

I see there is a large number of tips posted here relative to any discussion regarding how to increase your edge through methods such as qualitative or quantitative analysis.

If you would be interested in joining a private study based group post here with some details of where you feel you are at in terms of being able to analyse a match to determine if there is a bet that represents value.

If you have other related skills such as programming that would be useful in analysing data and would be willing to trade this in exchange for more knowledgeable bettors providing the theory let us know.

If you have any suggestions on how this could be structured then pitch in.

Nuts102

Division wrote: »

I see there is a large number of tips posted here relative to any discussion regarding how to increase your edge through methods such as qualitative or quantitative analysis.

If you would be interested in joining a private study based group post here with some details of where you feel you are at in terms of being able to analyse a match to determine if there is a bet that represents value.

If you have other related skills such as programming that would be useful in analysing data and would be willing to trade this in exchange for more knowledgeable bettors providing the theory let us know.

If you have any suggestions on how this could be structured then pitch in.

What type of programming skills are you referring too?

I would be pretty decent at SQL and Excel.

Division

It depends on the amount of information the group would be able to collect regularly as to the best approach. I'm not too knowledgeable but Python has been used by other groups producing good models. Someone that would be good at making scrapers to get some info from various sources would be useful.

People who can come up with a way to interpret the stats would be useful. A group that can discuss the weightings to give to each stat and some others who are knowledgeable about different leagues to add the bit of intuition that will make a good model a great model for betting purposes.

You can see from the bets and banter thread that overall the bets are losing a lot of money as they are based on few if any stats and too much guesswork. I'm looking for ways to bring stats to the forefront of the decision making process.

Division

Ultimately I am wondering if the skills and desire are on this forum for a group to form with the skills to make an advanced model in order to make an output table similar to something like this;



If the skills are there then there is a lot of money that could be won.

Nuts102

Division wrote: »

Ultimately I am wondering if the skills and desire are on this forum for a group to form with the skills to make an advanced model in order to make an output table similar to something like this;



If the skills are there then there is a lot of money that could be won.

Before anyone can answer you about putting this together, what I would need anyway is.

An Excel file with the data
Rules around how you want to get the percentages, for example what are you using to calculate for win percentage

I can do all the calculations once I have the data and the rules around each stat.

Then how are you getting the data if it is not scraped then it is very manual work especially if you are looking at multiple leagues.

Even with all the data, you need someone to produce the stats for each fixture by manually putting in the fixture.

Then you need someone to analyze the fixtures and come up with the bets.

It would require at least 3 people

Someone who can scrape the data
Someone to build the stats and enter the fixtures
Someone to Analyse

Division

Nuts102 wrote: »

the rules around each stat

This is by far the most crucial aspect to get something that could be used across leagues and will take a lot of time in all likelihood even with focusing on one market.

Once that is sorted everything else will fall into place quickly.

Nuts102

Division wrote: »

This is by far the most crucial aspect to get something that could be used across leagues and will take a lot of time in all likelihood even with focusing on one market.

Once that is sorted everything else will fall into place quickly.

The first part is getting the data, and then once you can see what can be used a few people together can work together to build rules.

I have worked with someone on Boards making a start on building something.

Have you found somewhere to download free data?

Nuts102

I have attached Championship Data with some calculations done in Excel.

Can someone have a look at what is their on sheet 1 and come up with ideas.

I have made an effort at things on other tabs.

Nuts102

Nuts102 wrote: »

I have attached Championship Data with some calculations done in Excel.

Can someone have a look at what is their on sheet 1 and come up with ideas.

I have made an effort at things on other tabs.

Can anyone go through that and come up with anything for tonight's championship games.

In work till 7 so won't get a chance.

Would be interesting to see what anyone can come up with.

wylo

As a programmer Id be interested in contributing, I have written scrapers/crawlers before.

Division

I'm in contact with Opta to see what stats we can get and for what cost. Potentially this can be sourced elsewhere cheaper.

We still need at least 2 people who have an educated opinion on how stats should be interpreted to generate outcome percentages for the win-draw-win and over/under betting markets.

I'm also looking for a good programmer to do a more basic bit of work to analyse and present football data if anyone wants to PM me.