leftism Posted December 21, 2010 Share Posted December 21, 2010 Hey guys, i'm currently backtesting football lay system which over the past 5 seasons shows good profitability year on year. I'm trying to tweak my odds criteria for selection and have a question... Is it better to reduce my odds range down to say less than 6, which reduces my strike rate, and reduces my number of bets BUT increases my ROI overall? Alternatively i can increase my odds range to an acceptable limit of say 11, which increases my strike rate, increases my overall number of bets, increases my overall profits (since the rate at which i chuck money into the system increases), but my ROI is reduced... Any thoughts???? P.S- I shall be most likely sticking up a separate thread with the system picks once i get the odds criteria nailed down. Quote Link to comment Share on other sites More sharing options...
clay747 Posted December 21, 2010 Share Posted December 21, 2010 Re: ROI or strike rate I would go for the higher ROI and work out a staking plan to maximise the profits. If you lower your ROI in favour of more bets, you will increase the variance in your results, so your stake per bet will have to be lower or your risk of ruin will increase. Quote Link to comment Share on other sites More sharing options...
AJ Posted December 21, 2010 Share Posted December 21, 2010 Re: ROI or strike rate At the end of the day what matters is profit and variance. If you can have an roi of 6% instead of 7% and put much larger sums through it, that's a good thing. Quote Link to comment Share on other sites More sharing options...
Matthew Posted December 21, 2010 Share Posted December 21, 2010 Re: ROI or strike rate If you lower your ROI in favour of more bets' date=' you will increase the variance in your results[/quote'] Is that true? Surely the more bets you place in a season, the smoother the ride. Reducing the number of bets will increase the likelihood of getting a bad end result that is purely down to random variation around the expected result. Of course if your reducing your bets from 3000 to 2500 in a season then it's not a problem, but if you are going from 500 to 200 then I'd say this is a problem. Normal variance doesn't settle down in even 200 bets. My suggestion would be to take your historic results and randomly select a subset and see the swing in the results you observe. Even if the whole sample (say a 1000 bets) shows a yield of +7%, I'm sure a subset of 200 bets could easily, on a number of occasions, show a loss. I don't lay over 10.0. At 1pt per lay on a 100pt bank, losing close to 10% in one bet is a tough pill to swallow. Clay747 is correct when he says for you to be wary of too many bets - your risk of ruin increases. Good luck - backtest backtest backtest - it's the only way to have a vague idea of what is coming up. PM when you put your thread up so I don't miss it - sounds good. Got any stats to share on performance as it currently stands before you start playing around with your odds criteria? Be wary of retrofitting and seeking the best result by testing things until you get the right answer. GL Matt Quote Link to comment Share on other sites More sharing options...
slapdash Posted December 22, 2010 Share Posted December 22, 2010 Re: ROI or strike rate If you lower your ROI in favour of more bets, you will increase the variance in your results, so your stake per bet will have to be lower or your risk of ruin will increase. To be slightly pedantic, lowering your ROI doesn't affect your variance very much. What it does do is to reduce the variance needed to wipe out your expected winnings. For example, over 100 bets, with an expected profit of 100 units and standard deviation of 50 units, you're more likely to make a loss than with an expected profit of 10 units and standard deviation of 20 units, even though your variance is larger. Lowering your ROI might increase your risk of ruin (depends on exact figures, staking plan, etc.), but not really because of increased variance. Actually, with the same ROI (assuming you're calculating this as profit divided by amount risked for lay bets?) then by laying at higher odds you should be able to make more profit per bet for the same level of risk. Given that your ROI seems to be lower for the higher odds lays, it really depends on the exact figures which effect is more significant: higher odds (good) or lower ROI (bad). Quote Link to comment Share on other sites More sharing options...
leftism Posted December 22, 2010 Author Share Posted December 22, 2010 Re: ROI or strike rate Thanks a lot for the advice guys! I was always under the impression that the only way to reduce variance was to increase the sample size, although its been a while since i've done statistics at Uni. Anyways, from backtesting it does appear that sticking to tighter odds reduces fluctuations in overall bankroll. Just to give people an idea of the system i'm backtesting. I use Matlab to backtest and collate data from the amazing website football-data.co.uk. Sample leagues are pretty much all first tier european leagues apart from the German Bundesliga which does not seem to fit with any systems i've tested. I've backtested to 06-07 season. I took lay odds as: (backing odds from B365 +15% to account for the spread and betfair commision). So with all selection criteria the same apart from odds range, the results are like this (1000e starting bankroll staking to win 50e on each bet): Odds Quote Link to comment Share on other sites More sharing options...
leftism Posted December 22, 2010 Author Share Posted December 22, 2010 Re: ROI or strike rate Actually, with the same ROI (assuming you're calculating this as profit divided by amount risked for lay bets?) then by laying at higher odds you should be able to make more profit per bet for the same level of risk. Given that your ROI seems to be lower for the higher odds lays, it really depends on the exact figures which effect is more significant: higher odds (good) or lower ROI (bad). Sorry i don't understand how laying to higher odds could give you more profit per bet. If i lay 10 euro at odds of 5 i will risk 40 euro. If i lay 4.45 euro at odds of 10 i will risk 40 euro. So my profit is over twice as high if i lay at odds of 5... Quote Link to comment Share on other sites More sharing options...
slapdash Posted December 22, 2010 Share Posted December 22, 2010 Re: ROI or strike rate Sorry i don't understand how laying to higher odds could give you more profit per bet. If i lay 10 euro at odds of 5 i will risk 40 euro. If i lay 4.45 euro at odds of 10 i will risk 40 euro. So my profit is over twice as high if i lay at odds of 5... Sorry, by "risk" I didn't mean the amount you were risking on each individual bet, but something like the risk of ruin. In your example, if your ROI is the same then laying at odds of 5 is "riskier" than laying at odds of 10, as you're more likely to have a long unlucky spell. The higher the odds at which you're laying, the safer it is to risk more per bet given the same ROI, for any reasonable definition of "safer" (e.g., lower risk of ruin). Quote Link to comment Share on other sites More sharing options...
leftism Posted December 22, 2010 Author Share Posted December 22, 2010 Re: ROI or strike rate Sorry, by "risk" I didn't mean the amount you were risking on each individual bet, but something like the risk of ruin. In your example, if your ROI is the same then laying at odds of 5 is "riskier" than laying at odds of 10, as you're more likely to have a long unlucky spell. The higher the odds at which you're laying, the safer it is to risk more per bet given the same ROI, for any reasonable definition of "safer" (e.g., lower risk of ruin). Ah right... Yeah i'm still at a loss as to my best approach with this. More backtesting using different staking plans and odds ranges required. I like the idea of randomly formulating smaller samples from my overall group and may spend a bit of time doing that just to see. So far in backtesting, the longest losing streak has been 3, but that could be 30% of your bank if you're laying to fixed liability of 10%... Far too risky in my opinion Quote Link to comment Share on other sites More sharing options...
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.