Which Staking Plan To Use?

Happy New Year Everyone,

I have a strategy for picking selections and have some stats that I have been collecting and I wondered if anyone would recommend a staking plan that could make money from the selections.

Total markets:  4,357
Wins:               2,938 (67.4%)
Loses:              1,419 (32.6%)

Loses in a row:

10 Loses   1
9 Loses     0
8 Loses     0
7 Loses     2
6 Loses     4
5 Loses     4
4 Loses     28
3 Loses     69
2 Loses     190
1 Loses     652

With a £50 stake the average win is £20.75 (which works out to be about odds of 1.44).

Just using the same stake is not a winning strategy I know, but does anyone think looking at the stats above there could be a staking plan that could end up in profit?

Thanks,

Whitey

Happy New Year Everyone,

I have a strategy for picking selections and have some stats that I have been collecting and I wondered if anyone would recommend a staking plan that could make money from the selections.

Total markets:  4,357
Wins:               2,938 (67.4%)
Loses:              1,419 (32.6%)

Loses in a row:

10 Loses   1
9 Loses     0
8 Loses     0
7 Loses     2
6 Loses     4
5 Loses     4
4 Loses     28
3 Loses     69
2 Loses     190
1 Loses     652

With a £50 stake the average win is £20.75 (which works out to be about odds of 1.44).

Just using the same stake is not a winning strategy I know, but does anyone think looking at the stats above there could be a staking plan that could end up in profit?

Thanks,

Whitey

1.44 needs a minimum strike rate of 69.4 to be profitable .. and this is before Betfair commission.

If you can't break even at flat stakes after commission, a staking plan will only delay the inevitable bust.

You're close to something good, just need to try a little harder.

I don't know if this will help you.  I compiled a lot of data over several months and was exploring the best way to make it work profitably.  I came across Betting-Dev.Com and asked them to give me some answers with their Money Management Service. 

What I got blew me away: they sent me the results from sixteen (16!!) systems in graph and number form so that I could see clearly the risks, highest stakes, losses and winnings and more.  I can see great returns but at a higher risk and a cluster of sound returns.  Seven systems would have broken the bank at some stage and one was an outright loser.  Powerful stuff that got rid of a lot of guesswork and the more data, the better.

As I remember it they asked for a month at least and in CSV format.  What was especially interesting was that you could be looking at a system that looked unprofitable but could well be profitable using another staking plan.  For example, I could have been testing Lay %, which showed losses of £29,000 whereas Maria showed a return of £3300!  Made me want to back and look at other systems I've discarded.  Good luck

It's Always possible to fit a system to be profitible due to historic data. You would construct a system from a sample set of historic data, then test it on the rest (the data that have not been used). Then if the system is profitble of that data it might be profitble, you need to show that i can stand a standard deviation of at least 1 sigma.

Well said Tupp,
Few people seem to realise this simple but vital lesson in model building.
Too many opt for staking plans that increase stakes to get them out of trouble.
The fact is, staking plans won't work when a level stake loss is shown. A staking plan may but not necessarily help increase profits if a level stake profit is first determined.
I know this as I spent at least two decades trying to prove the weapons of math instruction were wrong. Professors of statistics finally convinced me of my erroneous ways.
So please all you forumites take heed and try to develop strategies based upon level stake profits, it will save you a lot of frustration and money.
R

It looks like you have a negative expectaion (negative value). So in general one would say you can't win in long term no matter what you do. But I wouldn't give up until I made a simulation. You need two things for a simple simulaton the distribution of losses and the distribution of the odds. In your text you have given the distribution of the losses and the average odds is it normal distrubuted around 1.44?

I Think the most importent thing here is how the odds is distributed, I mean depending on that it could still be profitable (if we assume that the probability is constant anyway).

I could help with a simulation if you want? But in that case I need the distribution of the odds.

I came across Betting-Dev.Com and asked them to give me some answers with their Money Management Service. 

What I got blew me away: they sent me the results from sixteen (16!!) systems in graph and number form so that I could see clearly the risks, highest stakes, losses and winnings and more.  I can see great returns but at a higher risk and a cluster of sound returns.  Seven systems would have broken the bank at some stage and one was an outright loser.  Powerful stuff that got rid of a lot of guesswork and the more data, the better.

Hi,

How much did they charge for the Money Management Services?


Larp

There is a piece of software called the Staking Machine.  I dont have it but know people that do.  You just feed in the list of bets (odds & outcomes) and it will compare a whole raft of staking systems. (Back or lay).

I would not be surprised if this is not what BettingDev.com used.  I entirely agree with the above though.  If my historic results do not show positive results at level stakes I just forget them.  Also please don't ignore the effects of BF commission and PC, as that can be massive.

Once I identify a system that works historically on level stakes I go live with it until I cannot stand any further losses, then drop it !!  

Ho hum!    

It's all about finding value i.e you want to find an odds that's higher then the true probability of the outcome - Easier said than done.

Once I identify a system that works historically on level stakes I go live with it until I cannot stand any further losses, then drop it !!  

What is your explanation for this behavior ( the value seems to disappear why) ?
 

I always imagine, that, despite best attempts not to, that I have back-fitted the data in some way. 

ie  I have fired a bunch of arrows at a tree, gone and painted a target round the best bunched arrows, then claimed what a great shot I am. 

(Either that or just that trends in data often reverse for no apparent reason).

Obviously sometime there are good practical reasons - like spotting on spreadsheet that if you lay a load of high priced runners, then hedge any that dip to say below 20 in play you make a profit overall.  In reality even if the trigger fires in a hedge when the price dips to 20 it may only get matched at 18 or 16 or lower, as it is in play and things are changing fast.  This kind of difference is enough to kill a good theoretical system.

ie  I have fired a bunch of arrows at a tree, gone and painted a target round the best bunched arrows, then claimed what a great shot I am. 

A really funny and vivid description of the art of back-fitting 

I belive most systems fails sooner or later due to Market Efficiency. In a perfect market all information is within the odds, so we would end up with expected value corresponding to -commision. The market at betfair is not perfect, but I belive even with a "weak" efficiency sooner or later the market will catch up and the system will no longer have any value.

One way to come around that would be to accept the above and quit the system when you have been catched up. As I understand is your strategy.

Yes Tupp, I believe you are right. With the advent of fast moving technology and state of the art algorithms, markets are efficient anywhere that racing matters. From Australia through Asia HK and Japan to America and now the UK. Without Betfair here in the UK system players would be constantly donating to the Bookmaking Foundation.
At least with Betfair one can, with some help from statistics and speed figures, come up with a system that could break even long term after commissions. At least that is a good start, and with some extra work, which one could learn, such as (video) race reading, trainer intentions trainer jockey relationships and some of the other finer points of handicapping and hugely important getting value on your bets, it may well be possible to turn it into a profit. We are forever hopeful.
R

one could learn, such as (video) race reading, trainer intentions trainer jockey relationships and some of the other finer points of handicapping and hugely important getting value on your bets, it may well be possible to turn it into a profit. We are forever hopeful.
R

Having more information than the market or to manage existing information more successfully than the market, will definitely give you the edge that we all want to have.
It requires both knowledge and time, much of what you describe is difficult to get into a trigger I suppose.

Sorry for the late reply!

Thanks for the advice everyone