What Sample size would be needed to validate a strategy?

Learn sports betting strategies and discuss key factors to consider when placing a bet.
Post Reply
redbrix
Posts: 13
Joined: Fri Apr 17, 2009 12:45 pm

Hi all

I have been trading an in play horse racing strategy that has showed a decent consistent profit recently.

I have developed a consistent approach for it and have 200 trades completed. The strike rate is 80% and expected value on a £10 stake is £1.20. I am really happy with the process but not sure whether to keep stakes small and get more trades completed or start to slowly raise the stakes.

Would you consider 200 trades since early Jan a decent sample size or would you want to see more (how many?) before scaling up?

Thanks all!
User avatar
Derek27
Posts: 23471
Joined: Wed Aug 30, 2017 11:44 am
Location: UK

It's a question that's often asked and the answer (if there is one) depends on a number of factors and varies substantially on how much risk you're taking. There's certainly no harm in increasing stakes in proportion to your bank as you're only increasing the risk of losing what you've already won.

I haven't read the link below so I'm not sure if it's applicable to you, but it might be worth a look.

viewtopic.php?p=54639#p54639
rik
Posts: 1583
Joined: Sat Jan 25, 2014 5:16 am
Location: London

scale up gradually in relation to your bank size?
User avatar
ruthlessimon
Posts: 2094
Joined: Wed Mar 23, 2016 3:54 pm

The conventional responses are gonna be "it depends. more trades, more time, more out-samples etc"

But I'm gonna pose some left field questions

(don't answer them btw; just think about em) (also I'm coming from a preoff context, not inplay)

-Can you explain the inefficiency you're exploiting?
-Do the optimisations/logic make sense?
-How much leeway is available before we go -ve?

For me, those questions are massive dictators of longevity. but they're tricky questions (especially the 2nd one; because genuine edges tend to have counterintuitive elements).

If we can nail those questions - the strategy can be scaled hard & quickly imho. But if we can't, (most likely); then it's the long game of gradual reinvestment
User avatar
ShaunWhite
Posts: 9731
Joined: Sat Sep 03, 2016 3:42 am

If it's working then push on with the stake. I've had automated strategies fail after a few hundered selections and had them fail after 100,000+ selections. Likewise strategies that fail for the first few hundered selections can become profitable later if the underlying principal has a +ve expectancy.

The link Derek/Euler posted is interesting but only really applies to situations where the underlying opinion/survey result doesn't change over time, whereby in markets that change or evolve require weighting recent results higher in significance than the oldest results.

The bottom line is like all trading it's a gamble, you look at your recent results and gamble on how much you want to invest that day given the likely return/likely loss and liklihood of each of those outcomes occurring. These probably aren't the clear cut answers you want but definitive answers simply don't exist whether you're trading a few quid on sports or billions in managed assets. Experience and 'feel' are just as much part of automated trading as they are when trading manually. Automation isn't a magic way to enter a world of certainty even though it appears to be based purely on statistics.
redbrix
Posts: 13
Joined: Fri Apr 17, 2009 12:45 pm

Thanks to all for your responses, I appreciate the feedback.

I am risking 4% of equity per trade (so £10 stake on a starting balance of £250)
The account has nearly doubled so I plan to now increase the stake but remain under the 4% of equity risked.

The stats show the longest run of consecutive losers so far to be 4, so the 4% of equity seems manageable. I guess my real concern is whether 200 trades would reflect longer term averages for consecutive losing trades. The only way to find out is to continue to trade it and track the results. This is part of the process...

Thanks again!
User avatar
firlandsfarm
Posts: 2688
Joined: Sat May 03, 2014 8:20 am

Personally I think you have to look at at least 12 months for any system so that you capture all the seasons and even then there could be 'special circumstances'. Let's look at Going … below is a table of NH races in January for the last 5 years and the number where the Going was Heavy …
JanHeavyGoingRaces.JPG
Clearly a system that was profitable in 2016, 2018 and 2020 may not have been profitable in 2017 and 2019 (or vice-versa) because of the number races with Heavy Going. (Yes, 2018 surprised me as well!)

But you can't wait 5 years before you jump into a system so my advice is don't grow your stakes too quickly and don't ever think all is OK (I'm in total agreement with Shaun). Monitor daily and review regularly.
You do not have the required permissions to view the files attached to this post.
rik
Posts: 1583
Joined: Sat Jan 25, 2014 5:16 am
Location: London

for sure agree if you can easily define your edge thats a big plus
User avatar
Kai
Posts: 6092
Joined: Tue Jan 20, 2015 12:21 pm

redbrix wrote:
Thu Feb 13, 2020 9:36 pm
I have developed a consistent approach for it and have 200 trades completed. The strike rate is 80% and expected value on a £10 stake is £1.20.
Well done, it sounds good so far.

I love speculating (on preoff markets) as much as the next trader, but I feel that the inplay markets are the real heart and soul of a betting exchange, or at least they should be, racing markets in particular. There is just so much speed, uncertainty and greed at times in this market and that, for me personally, is the perfect storm of opportunity.
Post Reply

Return to “Betfair trading strategies”