An experiment

[trader_dante]

i'm travelling right now....will start the experiment when i get back.

Ssingh,

I just found your thread and had a proper read.

I've been doing a slightly different version of this experiment for a while now with LIVE money.

Market and direction (long or short) is decided completely randomly.

Stops are fixed.

Targets are discretionary. (slightly different to your plan to fix your target)

I have made 40 trades in 7 weeks.

Won just 47.5% of them.

The account is up 234.75%.

It can be done.

Good luck.

 

[rathcoole_exile]

Tom, what you going to blow the 24 quid on ?


sorry .......

 

[TheBramble]

well if you were trading test and retests of fork lines it would have made some sense .
im no expert but a little bit brave :cheesy:

I’m sure any one of a million ‘explanations’ for why the price did what it did would fit today’s action.

Do your ‘forks’ work every day on every instrument?

 

[Hotch]

Ssingh,

I just found your thread and had a proper read.

I've been doing a slightly different version of this experiment for a while now with LIVE money.

Market and direction (long or short) is decided completely randomly.

Stops are fixed.

Targets are discretionary. (slightly different to your plan to fix your target)

I have made 40 trades in 7 weeks.

Won just 47.5% of them.

The account is up 234.75%.

It can be done.

Good luck.

Quite similar to by results, told ya all!

I think I'm gonna start a cult...

 

[trader_dante]

Tom, what you going to blow the 24 quid on ?


sorry .......

lol rath.

I hate percentages, the only reason I wrote it down was because I meant to mention that I used a standard 5% risk per trade although I forgot to put that in.

The actual amount profit made so far is £539

 

[TheBramble]

I have made 40 trades in 7 weeks.

Won just 47.5% of them.

The account is up 234.75%.

Interesting and heartening that there is (at least?) one other serious trader giving this experiment a serious review, and in your case Tom, already trading a similar version – presumably with experimental size live dosh.

What interests me is the likely Pwl on this class of system. My own very limited data show an unrealistically (?) high Pw. Again, with any trading system the run is only going to approximate to a mean after a significant number o f trades. But given my current position on the validity of trading system statistics, I’m not about to get too fussed about that. But very keen to get others’ data if they’re looking at similar systems.

In your case Tom you’re exercising discretionary targets, but fixed stops which is going to make the Aw:Al variable across all trades. Again, not an issue, just interested in your approach.

We have the Aw: and Al fixed (in SSingh’s version) which makes the only variable on the expectancy side the Pwl. No substitute for real market testing unfortunately – which just takes time.

It’s a real effort to stop yourself ‘improving’ somebody else’s system, but this one, as I’ve said before, is so devoid of pretension and bullsh!t it cries out to be tested ‘as is’.

 

[MrGecko]

Market and direction (long or short) is decided completely randomly.

Stops are fixed.

Targets are discretionary. (slightly different to your plan to fix your target)

TD what you're doing is a totally different kettle of fish if you're using discrtionary exits.

(mind you, kudos for doing it, but it's a different set of circumstances to the OP's)

 

[price777999]

I’m sure any one of a million ‘explanations’ for why the price did what it did would fit today’s action.

Do your ‘forks’ work every day on every instrument?

NO they dont ..... i think i will start fliping a coin

 

[TheBramble]

TD what you're doing is a totally different kettle of fish if you're using discrtionary exits.

Not quite TOTALLY different Mr. G.

All Tom is doing is allowing discretionary choice on Target. The initial direction of the trade is random (as is the fixed stop part), which is what most found pointless/ridiculous at the outset.

 

[TheBramble]

Targets are discretionary. (slightly different to your plan to fix your target).

Tom, are you setting the target at trade entry and leaving it or are you allowing yourself to dynamically change the target while the trade is running?

 

[trader_dante]

Tom, are you setting the target at trade entry and leaving it or are you allowing yourself to dynamically change the target while the trade is running?

Hi Bramble,

I've been doing this for quite some time and toyed with starting a journal on it many months back but looks like I got pipped to the post (so to speak) by a few others. I would still start a journal if there was enough interest but just expect to receive a lot of "yawwwn" type posts. lol

I'm using dice instead of coins.

One throw of the dice gives a number between 1 - 6 and gives me the market to trade.

1: Crude Oil
2: Dax
3: Eur/Usd
4: Bund
5: Gold
6. Wheat

(so, equities, bonds, fx and commodities (softs, metals and energies) are covered. A nice broad spectrum.

Then once the market has been chosen, I throw again.

Even number: Long
Odd number: Short

The trade is entered bang on 8am each morning for all trades.

One trade per day.

The stop is always 40 ticks/pips.

The target is totally dynamic. I use my reading of the market and interpretation of price action to decide whether to take any profits quickly or whether to run it. If I find myself long Bunds, for example, and the ECB announce quantitative easing, I am not going to be coming out once an arbitrary profit target has been hit. That is not sensible.

The rule I have used is that the stop cannot be moved until the trade has offered at least a 0.5:1 (e.g. 20 ticks profit is on offer). At that point it can be moved/trailed if so desired.

The trade can run as long as it takes until I am stopped out. Some have run for several days.

As soon as I am out, the dice is thrown and I start again the next trading day.

What I was trying to prove is whether a good trader can make money using a system that throws them randomly into a market (both the market and the direction are random) when ALL they have any power or control over is their exits.

The reason I give control over the exits is I am not trying to prove that the markets are random per say, I am simply trying to prove that entries (trade reasoning) is largely meaningless.

 

[KillPhil08]

One could argue that if you did the same for entries, eg. Long or short but you get to choose where to do it with a profit take of 40 pips is also possible.

Much like that fxpulp offer.

 

[price777999]

im finding it a bit disconserting that experianced traders are interested in a coin fliping strategy .

 

[MrGecko]

The target is totally dynamic. I use my reading of the market and interpretation of price action to decide whether to take any profits quickly or whether to run it. If I find myself long Bunds, for example, and the ECB announce quantitative easing, I am not going to be coming out once an arbitrary profit target has been hit. That is not sensible.

[...]


What I was trying to prove is whether a good trader can make money using a system that throws them randomly into a market (both the market and the direction are random) when ALL they have any power or control over is their exits.

Thats an important difference; the fact that you are making money with this strategy demonstrates that markets are not normally distributed all of the time, and you use your expertise to select which of the positions you are in are likely to be normally distributed (and close them out at the opportune moment) and which are not - i.e. those instances where your experience has shown that the distribution is skewed in some way - quite literally, the "high probability trades".

If you took this element of the system out, then you could put every position the OP's system into one of three baskets:

1) probabilities aren't skewed - Long OR Short

2) probabilities are skewed, and positioned correctly

3) As above, positioned incorrectly.

Now, assume that all the trades on basket 1) sum to nil (no reason not to it's noise, CLT, mean = 0) We are left with trades of type 2) and 3).

There will be instances where the OP's system does land a home run and picks up the 70 ticks (or whatever it was) of the trades in basket 2) - but they will be few and far between. Say the baskets 2) and 3) are evenly weighted (as we have no grounds to suggest otherwise, you are positioned correctly as many times as incorrectly) - imagine if we were to superimpose the distributions skewed to the upside and the distributions to the downside - the resulting distribution is symmetrical*. Then, we are long just as many times as we are short, with a symmetrical distribution - a situation identical to the trades in basket 1). The times you hit a home run are eaten up by the - more frequent - instances of you being stopped out.

* If you choose the right instrument, you could probably get your way around this by choosing an asset that had a large enough drift rate, like commodities - basically saying that the probabilities are more likely to be skewed in one direction that the other ("trading with the trend"). You will probably find fat tails too.

 

[Hotch]

Just think I should make a point.

I wasn't trying to prove that the markets are random, I'm not even sure if trading randomly would do it. I was proving that maths works, and that you don't need good R:R, tight stops etc.

Pft, picking your exits, that's cheating.

 

[SSINGH]

statisitically one can not lose 100% of the time.....casinos.....spread betting firms win generally win over time because the skew the odds in their favour by tinkering with core parameters.....plus with traders being influenced by emotion...trading more than their account permits is a losing combination.

in essence it should be possible for one to discover and establish very crude optimal parameters so that we win over time....

Trading ruthlessly without emotion....with fear....trading with simple set rules, parameters skewed in the traders favour ...should statistically and ultimately "OVER TIME" result in champions,.

The problem with human beings is that we ge too emotional about money....we want more and we want it yesterday....all down to the conditioning the we have under gone by the effects of marketing in the society that we live in.....

Look at Warren Buffet.... he is very down to earth.....and doesn't get sucked in to the fluff of societies marketing....his success is atributable to TIME and VALUE...a simple systematic approach....this guy always has bullets in his gun....while other fire blanks in a crisis.

 

[MrGecko]

parameters skewed in the traders favour

Pray tell, how?

 

[rathcoole_exile]

I am simply trying to prove that entries (trade reasoning) is largely meaningless.

absolutely

but you can jut hear the majority wailing about their "overbought/sold" stochastics

 

[TheBramble]

im finding it a bit disconserting that experianced traders are interested in a coin fliping strategy .

It is quite possibly that experienced traders are willing to explore the boundaries of ‘what is’ and ‘what may or may not be’ that makes them experienced traders.

It’s only disconcerting if you allow what you think you already know to become dogma and belief.

My view on price action in the lower timeframes predisposes me to allow that for the most part it is random. So I am interested in any research in that area and especially so if it shows any possibility of empirically disproving that hypothesis. Taking a static and analysis-free approach to trade execution (as per SSingh) is interesting to me from that standpoint. I can run a lot of simulations and estimate with reasonable accuracy the likely drawdown, expectancy and optimum stop/target size. But that is all quite irrelevant to what will actually happen in the reality of the market. I don’t trust statistics and will only give current performance the weight it deserves – basically, you’re only as good as your last trade. Execution data based on this class of system can only be accumulated empirically, directly from the markets and not from any theoretical model. Which is why experiments like these are so valuable and so expensive (time).

Perhaps it’s a cynicism born of many years experience that has gently led me away from the premature ejaculation of positive expectation to the objective arrogance of positive empiricism. I don’t think most traders’ use of current statistical models serve them very usefully. You can see this quite clearly in the markets when you shove the data through the standard pricing models – and wonder at the divergence between that result and what you’re actually able to trade it at...When you trade for your living you tend to get a different take on what works than do most textbooks.

What you believe limits you. Stay loose and open and you’ll get more of an intuition for what’s really going on. You’ll also get sticks and stones thrown at you, but it’s a small price to pay for being more ‘right’ then them doing the throwing. LOL.

 

[rathcoole_exile]

Hey Singh, any updates on your experiment ?