Do I have a genuine edge or is this just statistical variance?

[ShaunWhite]

The subject was also touched upon here ..viewtopic.php?f=2&t=18478&hilit=steyx&start=10#p182933

It didn't get far but I did post a crude spreadsheet with an example of how to use the excel trend analysis functions. Correlation to trend seems to be a reasonable measure of how close your trend is to the underlying results. It was just something I only played around with so its not really in any shape for public consumption. Might help, might not.

[pythonic]

About Joseph Buchdahl, I have read his older book and most of his articles and while he himself obviously has no edge apart from writing and providing information and is generally very sceptical about profitable betting, I found that what he writes about risk management and statistics is always pretty solid.

LinusP: the average implied probability from the odds that betfair gives me at the time I enter the market is around 12%. The spread is very large, though. Over my 1800 bets, the lowest probability I took was around 1%, and the highest was around 75%.

Given the average taken odds imply that the chances are around 12%, how many matches would I need to confirm my strategy is profitable?

With 12% I get a standard deviation for yield (edge) of about 6.4% which means that you have a 66% chance that your true yield is equal to your observed yield +- 6.4% and a 95% chance that your true yield is equal to your observed yield +- 12.8%.
If you have a lot of very high and/or very low odds in your sample then this could be somewhat different but mostly this approximation is pretty robust.

You could try to exclude the higher odds for a lower variance but be careful about data mining odds ranges.
Generally data mining is always a problem with data samples from the past and it's difficult to eliminate that completely.

Otherwise I would suggest to try the strategy live with small stakes first and slowly scale it up when it looks like your profit estimations were correct.

[Speculator_3]

Given the average taken odds imply that the chances are around 12%, how many matches would I need to confirm my strategy is profitable?

Sample sizes may be chosen in several ways:

1. Using experience – small samples, though sometimes unavoidable, can result in wide confidence intervals and risk of errors in statistical hypothesis testing.

2. Using a target variance for an estimate to be derived from the sample eventually obtained, i.e. if a high precision is required (narrow confidence interval) this translates to a low target variance of the estimator.

3. Using a target for the power of a statistical test to be applied once the sample is collected.

4. Using a confidence level, i.e. the larger the required confidence level, the larger the sample size (given a constant precision requirement).

You won't be suprised to know that I didn't know that by heart, I grabbed that from the Wiki page on sample size determination but I've never been able to understand it fully. We seem suprisingly short of mathematicians on here who can put it into a formula with just the parameters necessary for what we do. 100s of years of experience kicking around on here and nobody knows any semi-serious maths, it's strange. Either that or they're not the sort who help much.

Being constructive though I think a good place to start is 100 of the least likely outcome...ish?
200 @ 50%
400 @ 25%
800 @ 12% etc

But there's clearly a lot more to it than that. "As soon as the charts look reasonable" is my other goto on sample size and I suspect that goes under #1 above.

It's funny that you say the forum is short of mathematicians, because I am one! (Got my PhD in the subject). However, my PhD was in pure maths, completely unrelated to statistics; in fact throughout my education, I never had interest in the subject and lacked formal exposure to the concepts you mention (confidence intervals, for example). I am just learning about these topics now.

One really needs a statistics / data science background for such questions. It's entirely possible that quite a few people, who have studied a degree in mathematics, would have no clue about hypothesis testing, confidence intervals, and so on.

[pythonic]

It's funny that you say the forum is short of mathematicians, because I am one!

Same here

[ShaunWhite]

It's funny that you say the forum is short of mathematicians, because I am one!

Same here

ooooo now that is interesting I might need to quiz you about some basic notation at some point. My maths only goes as far as when I fell asleep doing differentiation of integrals about 35 years ago.

[pythonic]

Feel free, Shaun!

[ShaunWhite]

Feel free, Shaun!

I'll tuck that away like a Get Out of Jail Free card and perhaps use it sometime. Thanks.

[Euler]

It's actually quite nice to have some deeper discussion on the forum. If you feel we need to create a new area for it, let me know.

[Euler]

Have you read his book? Its hard to recommend due to the underlying feeling that he believes making money from gambling is impossible and anyone who says they are is a liar but then thats why he is probably writing rather than betting..

He has written some good stuff but I think he falls into an intelligence trap where you convince yourself that because you can't do it, nobody can. I've had many meetings or 'discussions' like that in 20 years. But that in itself tells you where an edge is in the market. About three years or so ago somebody really went at me with a similar argument so I invited him to the office, but he seemed to think I was going to beat him up or something so wouldn't come. All I was going to do was show him a couple of things that would have made him think again.

The upshot, I'll give a lot of credit and respect to those that use their intelligence to solve a problem rather than argue a point.

[gazuty]

The upshot, I'll give a lot of credit and respect to those that use their intelligence to solve a problem rather than argue a point.

+1

[gazuty]

If you think have an edge and given your mathematical brain skip over here - https://stattrek.com/online-calculator/binomial.aspx

[Speculator_3]

If you think have an edge and given your mathematical brain skip over here - https://stattrek.com/online-calculator/binomial.aspx

Thanks.

[Speculator_3]

Thought I'd post an update.

Since I wrote this thread back in May, I have made over 2000 bets - with real money, small stakes. This brings up the total of bets I monitored to over 4000. And I'm in profit.

Not as large as I expected from my paper trading experiment earlier, but that's my own fault. There were a number of "bad bets" I made (panicking when missing an opportunity and getting on the market at the wrong time, ending up losing several relatively large stakes), and also many times when I disobeyed my initial trading strategy and did not enter the market at all (again due to psychological fear of losing the profits I'd gained just before). If I did all these things correctly, I estimate my initial bank would have grown 2.5 times, or even more.

One thing I note, is that I changed my strategy somewhat: I enter the market at the same "signal" as before, but instead of leaving the trade open till the end, I usually trade out around a third of my stake (just enough to break even or make a small loss) if a certain next event in the match changes the odds in my favour, and let the rest of the stake run. This tends to minimise the longer losing streaks I'd have otherwise - if I let the initial bet open till the end.

Still puzzled as to why I'm profitable. There seems to be some kind of a consistent pattern to my bets: typically, I make several rare winning bets in a quick succession, which radically increases my bank size. Then this is followed by a long period of straight losses or me just breaking even on my trades. These long periods bring my bank size down....But then these "several rare winning bets" occur and bring my bank up again - to a higher overall amount than it was before. And then the cycle repeats.

[PeterLe]

Just a thought to the maths guys...Why not try automation (if you're not trying it already). I think if you try it you will be hooked and would have thought it would be right up your street.
As long as the logic is sound its relatively easy to set something up and leave it to do its thing.
Its easy to compare and contrast knowing exactly why you entered and why you exited.
next day review/tweek if necessary and repeat
I left four programs running this evening and went playing snooker with a couple of mates. Three programs had very small losses, the other a surprising positive.
Im not saying it is easy per se, but I was never really any good at manual trading and thought Id give the automated strategies a chance, that was 12 years ago and still loving the challenge!
regards
Peter
PS Lost at Snooker

[BrianDee]

There seems to be some kind of a consistent pattern to my bets: typically, I make several rare winning bets in a quick succession, which radically increases my bank size. Then this is followed by a long period of straight losses or me just breaking even on my trades. These long periods bring my bank size down....But then these "several rare winning bets" occur and bring my bank up again - to a higher overall amount than it was before. And then the cycle repeats.

I had one of these strategies last winter - and binned it.
Went back to it and found the additional risk factor to remove and it's doing really well now