Football system - Bet sample size?

dsmythy

I'm using a system for football at the moment. Top four English divisions to be precise. So far this season it has been turning a profit however i don't really know what to make of this fact.

So i was wondering how big a sample size of recommended bets should there be before you can make a conclusion as to a system's future success and what do you think should be the minimum profit yield percentage you would want to achieve to make it viable and worthwhile?

thedini

dsmythy wrote: »

I'm using a system for football at the moment. Top four English divisions to be precise. So far this season it has been turning a profit however i don't really know what to make of this fact.

So i was wondering how big a sample size of recommended bets should there be before you can make a conclusion as to a system's future success and what do you think should be the minimum profit yield percentage you would want to achieve to make it viable and worthwhile?

whilst there isn't a definitive answer with regard to sample size it needs to be reasonably large so as to give an accurate indication as to whether or not it works. with regards to roi this will be depend on the individual. if you are betting in large amounts/large no. of bets then obviously a small roi % will still make it very worthwhile.

ArmaniJeanss

dsmythy wrote: »

So i was wondering how big a sample size of recommended bets should there be before you can make a conclusion as to a system's future success and what do you think should be the minimum profit yield percentage you would want to achieve to make it viable and worthwhile?

One of the key problems with obvious systems is that they tend to have a shelf-life.
For example there has less draws in English football this season than usual so there has been money to be made laying the draw.

However even after 500 games of this being a profitable system one of two things will fcuk it up medium term. Either bookies/Betfairians will increase the price at which you must lay the draw therefore savaging your ROI, or else there will be a gradual regression to the mean in the amount of draws.

Ditto over/unders bets if thats your system.

However if you've discovered something really obscure (like a trend in HT/FT results in boring midtable games) then theres no reason why it shouldn't be long term profitable. Anything approaching 5% ROI would be brilliant imo.

Hobart

ArmaniJeanss wrote: »

One of the key problems with obvious systems is that they tend to have a shelf-life.

ArmaniJeanss, I'd really be interested in what you mean by "obvious" systems. Do you have an example?

For example there has less draws in English football this season than usual so there has been money to be made laying the draw.

there have been less no score draws as far as I can see, but I see no buck in the trend in terms of draws, again, do you have the stats to back this up?

However even after 500 games of this being a profitable system one of two things will fcuk it up medium term. Either bookies/Betfairians will increase the price at which you must lay the draw therefore savaging your ROI, or else there will be a gradual regression to the mean in the amount of draws.

Ditto over/unders bets if thats your system.

However if you've discovered something really obscure (like a trend in HT/FT results in boring midtable games) then theres no reason why it shouldn't be long term profitable. Anything approaching 5% ROI would be brilliant imo.

I wish we all had that tbh...

dsmythy

ArmaniJeanss wrote: »

One of the key problems with obvious systems is that they tend to have a shelf-life.
For example there has less draws in English football this season than usual so there has been money to be made laying the draw.

However even after 500 games of this being a profitable system one of two things will fcuk it up medium term. Either bookies/Betfairians will increase the price at which you must lay the draw therefore savaging your ROI, or else there will be a gradual regression to the mean in the amount of draws.

Ditto over/unders bets if thats your system.

However if you've discovered something really obscure (like a trend in HT/FT results in boring midtable games) then theres no reason why it shouldn't be long term profitable. Anything approaching 5% ROI would be brilliant imo.

It's not on any special markets, just straight win bets. I make ratings for the teams and then using the ratings put them through a formula to find recommended odds for the outcome of a match. There is a bit of human input also.

It's interesting you mention the draws being down this season as the formula purposely makes the draw near unbackable meaning so far this season i've not backed a draw once, so that has obviously helped so far. Saying this i have come close to backing draws in the last couple of weeks so i think that price push might already be happening.

I presume with this form of ratings system the 5% ROI figure is also one to aim for come seasons end?

ArmaniJeanss

Hobart wrote: »

ArmaniJeanss, I'd really be interested in what you mean by "obvious" systems. Do you have an example?
there have been less no score draws as far as I can see, but I see no buck in the trend in terms of draws, again, do you have the stats to back this up?

By obvious system I mean something like draws. Theres been 129 games in the premier league this year and only 26 draws. Which is 1 in 5.

Traditionally theres been approx 2 draws in every 9 games (I don't have the figures to back this up but am fairly sure of it and will try to hunt them down) so draws are less frequent and laying them in the EPL this season would have been a cashcow.

But my pessimistic theory is that either this will regress back towards its long term trend, or else the price at which you can lay a draw will increase, so suddenly a winning system of laying the draw would stop being a winning system.

Ditto with something like the incredible amount of goals we are seeing in La Liga this season (the average is over 3 per game I think). But everyone is aware of it and the over/unders price or lines will change, if they haven't already.

So a longterm successful system needs to be on something obscure such that it doesn't leap out at too many people at once. Lets make one up, say you notice that away teams at 5/1+ 90minute win price in divisions 1 and 2 are consistently getting to half time no worse than level. Backing the outsider +.5 HT on AsiansHandicaps is giving you 7% ROI. You could cheerfully continue to do these bet sure in the knowledge that you are one of only a handful of punters who have noticed this trend.

Hopefully the OP has spotted something obscure like this.

HTH

hotspur

I suggest sitting down and learning some fundamental aspects of sports betting from scratch. If you have got to the point of using systems and tracking profitability and ROI without understanding the nature of probability and statistics in the first place then you have made a slight misstep. I say this because if I recall from other threads you are not an idiot, I wouldn't even bother recommending it to the average mug punter.

When you study the subject a bit you will understand that asking how big a sample should be to be significant actually makes no sense in the abstract. You will have a sample size, an outcome, and a probability of that outcome occurring by chance. The first thing you have to do is set a probability of your results occurring by chance that you are happy with.

In the social sciences we choose p<.05 or p<.01. p=.05 means the results will occur 5% of the time due to statistical variance. p=.01 it will happen 1% of the time. You are not going to be affecting people's lives if you are wrong so you will likely be inclined to accept a less demanding figure as evidence that your system is working. It is entirely up to you what level you are willing and comfortable in accepting.

The more impressive your results are then the smaller the sample size has to be in order to be confidant in the statistical significance of them. There's a big difference between a win rate of 80% over 200 bets and 52% in terms of being confidant about the significance of your system which produced these results and the probability that it has occurred by chance. So while as big a sample size as possible is always helpful there is no way of abstracting it from your winning percentage.

I would suggest that you don't ask for simple figures from people but rather take the time and effort to understand this vital aspect from the ground up. The actual business of serious sports betting is boring hard work once you get the fundamentals of it down, but practically nobody on this forum as far I can tell has those fundamentals down. I'm sure there's a couple but I doubt you would need two hands to count them. And remember shopping for odds is always the most important aspect of sports betting.

Dodge

hotspur wrote: »

The actual business of serious sports betting is boring hard work once you get the fundamentals of it down, but practically nobody on this forum as far I can tell has those fundamentals down. I'm sure there's a couple but I doubt you would need two hands to count them. And remember shopping for odds is always the most important aspect of sports betting.

I'd imagine there's quite a few. I'd agree that its not always apparent from posts here, but then a "serious" gambling topic rarely comes up.