Systems which work that don't

[psycho040253]

This may sound like the rantings of a mad man. They may well be but humour me for a little while.

What’s the first question we ask when a system stops working? ‘Why has it stopped working?’ There’s an implication in this statement and that is that it worked in the first place. Why do we think that the system worked? Because, it produced a profit and, if it did produce a profit, then, it must have worked, right? Well, maybe – but, maybe not. There is also a likelihood that system never worked in the first place and during the period in which the system was initially run, it just happened to be a lucky period for the system. The loss that the system is now producing may actually be the norm.

Perhaps, rather than asking why a system has stopped working when it fails, we should be asking why a system works when it does. If we can’t figure out the answer, maybe the system, long-term, doesn’t work.
Maybe if we look at things this way, perhaps there would be fewer disasters.

Psycho

[Alpha322]

This may sound like the rantings of a mad man. They may well be but humour me for a little while.

What’s the first question we ask when a system stops working? ‘Why has it stopped working?’ There’s an implication in this statement and that is that it worked in the first place. Why do we think that the system worked? Because, it produced a profit and, if it did produce a profit, then, it must have worked, right? Well, maybe – but, maybe not. There is also a likelihood that system never worked in the first place and during the period in which the system was initially run, it just happened to be a lucky period for the system. The loss that the system is now producing may actually be the norm.

Perhaps, rather than asking why a system has stopped working when it fails, we should be asking why a system works when it does. If we can’t figure out the answer, maybe the system, long-term, doesn’t work.
Maybe if we look at things this way, perhaps there would be fewer disasters.

Psycho

I dont belive in systems, if they existed we all would be Milioners , i belive more in methods and approach, ie (Patience plays a big part), dont show your hand to early or to late.
I have a method how i trade football matches, but they are not the same as how others approach football matches, some lose but overall its about winning the majority of the time.
Bookmakers have a approach & method also, they work on percentages, a view i have Taken that works. When people trade (newbies especially, i for one done this)focus about monetary funds and systems, approach and method what Dr Van Tharp disscussed works, you realise when you stop topping up your Bankroll every month

[CaerMyrddin]

I dont believe in systems myself too, the market is really well balanced most of the times and finding value across them is really hard.

In my first days when I discovered the lay the draw back 0-0 I made some simple calculations and soon discovered it had to break even, unless I offered prices instead of taking them, what kickly drove me to simply drove me to trading them out before kick of. But, if I was to trade these particular outcomes, why not the others?

[LeTiss]

I think systems become very difficult once you pay PC. By and large, BF's markets are pretty accurate, so 20% deductions leave most systems floundering.

I think being able to recognise trends, is far more important than devising systems.

[mister man]

it depends how long you trialled the system for.

1 month 2 months 6 months...
e.g 6 months with only 150 bets i would suggest is not as reliable as 3 months with 500 bets.

are their seasonal variations in which case a longer trial is needed.
whats the strike rate, if its 25% a sudden downturn isnt as bad as one with a 95% strike rate etc, as youd expect longer losing sequences.
maybe a slight tweak in selection criteria is needed because the markets changed slightly in some way.

some systems that dont work,do ???

good debating point.

[freddy]

Losing systems can make money for a very long period of time but likewise winning systems can lose money for a very long period.

You just need an awful lot of data as proof to begin with,
but even then things are changing all the time so you have to stay ahead of the game.

Most people who think they have a winning system are just being fooled by randomness imo .

[psycho040253]

Hi Guys

Many thanks for even bothering to read my post let alone for bothering to answer it.

I'm beginning to realise that there's a pretty good bunch of folk on this forum.

In my previous post, I ought to have written 'methods/systems etc' rather than just systems because the same arguments apply.

What I was really getting at was that we always ask why things went wrong when they do go wrong but rarely do we ask why things go right when they do go right.

When things go wrong and we lose, we bitch and moan, at least I do. Then, we get down to some serious analytical work in an attempt to figure out where and why things went wrong. When we find the cause, we then try and fix it. Yet, when things go right and we win oodles of moolah, we lie back, smile and think 'what a clever bugger I am'. Usually, just before the aweshit God comes a calling.

Psycho

[xitian]

I think whenever developing some sort of system you need to be very clear about the assumptions you're making and your reasonings behind why you think the system will work. That is effectively the information you hold above the rest of the market which you exploit to make a profit.

If rigorous backtests and real trading is successful, but the system stops working, then you should backtest again using the failed period's data to double-check the system is operating correctly. If the real trading doesn't match the backtest (i.e. the backtest says it should have been profitable, but your real trading flopped) then you have an implementation problem (such as loosing too much to spread costs or something).

If the backtest also flopped in a similar way to the real trading then there's a good chance that the assumptions you've made no longer hold true for whatever reason (or perhaps was never exactly true).

I don't believe trading systems where I don't have a clear reasoning behind them.

[psycho040253]

I don't believe trading systems where I don't have a clear reasoning behind them.

Well said that man!

A lot of people that I come across say 'It works. Who cares why!'

My answer is: You will, when it stops working.

Psycho

[psycho040253]

Most people who think they have a winning system are just being fooled by randomness imo .

Couldn't agree more.

Psycho

[Yantraman]

Interesting thread. I have a spread sheet which has revealed a back to lay "system" which in theory should work and be very profitable.

But I dont have the balls to play it as I just cannot find a reason why it shouldnt fail.

[Alpha322]

I dont believe in systems myself too, the market is really well balanced most of the times and finding value across them is really hard.

This is what i mean about not showing your hand to early or to late

[Alpha322]

I think being able to recognise trends, is far more important than devising systems.

Fully agree

[Alpha322]

I don't believe trading systems where I don't have a clear reasoning behind them.

Well said that man!

A lot of people that I come across say 'It works. Who cares why!'

My answer is: You will, when it stops working.

Psycho

And how these people on the internet get away with fooling newbies that they have the holy grail

[psycho040253]

Interesting thread. I have a spread sheet which has revealed a back to lay "system" which in theory should work and be very profitable.

But I dont have the balls to play it as I just cannot find a reason why it shouldnt fail.

Yantraman

Have you run a chi-squared against it? Basically, it provides an indication of whether or not the system is likely to be just random/lucky.

If you post the number of bets, the strike rate and the average odds, I'll work it out and post the result.

Psycho