Coin Flip Trading

[Liquid validity]

Here's the issue.

If you tested this coin flip system out, you would ABSOLUTELY be able to find a management technique that turned it into a profitable system historically.

For example, there's a book that "proves" money management can work on random entry. The problem is they only tested it on trending market.

So what they did was not to prove that random entries work BUT to prove that a trend following system works on trending markets, even with a random entry. To use the system profitably, you still have to make a call on whether the market will trend or not.

So - when you start making adjustments, how do you avoid changing your random system into a market specific system?

True.
Simple answer is to forwards test it along with other tests to rule out
anomalous sample distribution.

Actually I agree about the trending aspect and market specific tactics.
I run mine on EURUSD, which exhibits mean reversion characteristics
as opposed to trending behavior.

Thats the key - a market specific approach is unavoidable anyway.
What I do probably wouldn't work on stocks for instance, or even indices.
Currencies are much less likely to have long term trends or go to zero.
OK, fair point that the Euro may, or may not,get broken up at some point,
but that won't happen overnight...

In fact flash has a broadly similar take on it, not entry wise obviously,
but he has adopted a counter trend methodology that does work well
with currencies for the above reasons.

Any approach has to be market specific.
Same method different market, stop size for one will be market specific.
That is without going into session timing and other key timed events.

Random entry for someone who is available (either by choice or circumstance)
is lunacy, providing they are more efficient in the long run...

At the risk of repeating myself, due to binary logic, random entry is a good basis for automation,
if you choose, or are unable to be present at the optimum session times, nothing more.

 

[kimo'sabby]

Aye, for the most part. I'd say that you've got to be consistent in your overall strategy, but not necessarily wholly consistent in its application depending on the discretionary elements in it. To some extent there's a weighing of pros and cons to be done and you might not necessarily arrive at the same answer each time.

The pros and cons should be allowed for within the method.

 

[scholfield]

Is there any spreadbet platforms (or otherwise) that would easily allow your to break up your stoploss in the way being demonstrated in this thread?
I'm guessing not. It's a bit of a headache having to manually place all of your stops in intervals.

 

[DionysusToast]

In simple terms the logic of my strategy suggests that if I see x then there is a statistical expectation that y will follow. That sets the primary "rules". There are a few associated "conditions", both for and against, which cannot necessarily be described in such clear cut "rule" terms and are matters of judgement.

You seem to be suggesting that rules and discretion are mutually exclusive.

What you describe above (y follows x & statistics) is nothing to do with rules based approaches but everything to do with mechanical based approaches to trading.

You could have a rule that says "consider that the market has changed direction if there is overwhelming volume on the countertrend move, accompanied with a larger than average swing". This is a trading rule it has no "y follows x" element at all. It is 100% discretionary. There is no mechanical element. It needs no mechanical element. It needs no set of mechanical primary rules.

 

[barjon]

.......You could have a rule that says "consider that the market has changed direction if there is overwhelming volume on the countertrend move, accompanied with a larger than average swing". This is a trading rule it has no "y follows x" element at all. It is 100% discretionary. There is no mechanical element. It needs no mechanical element. It needs no set of mechanical primary rules........

Well, we could argue semantics all night.

In your example "rule" you are in effect describing x - "if you see overwhelming volume on the counter trend move, accompanied with a larger than average swing" - with the expectation of y following - "the market has changed direction".

If you do no more than trade every time you spot those conditions then that is a mechanical application. If, however, you go on to decide when you might trade in the context of that rule, then what you are doing is discretionary.

Not sure it matters all that much, we each know how we go about it and it doesn't really matter what we call it.

 

[kimo'sabby]

Not sure it matters all that much, we each know how we go about it and it doesn't really matter what we call it.

Without knowing each others specific detail (the void), I suppose semantics will rule the day.

 

[random12345]

True.
Simple answer is to forwards test it along with other tests to rule out
anomalous sample distribution.

Actually I agree about the trending aspect and market specific tactics.
I run mine on EURUSD, which exhibits mean reversion characteristics
as opposed to trending behavior.

Thats the key - a market specific approach is unavoidable anyway.
What I do probably wouldn't work on stocks for instance, or even indices.
Currencies are much less likely to have long term trends or go to zero.
OK, fair point that the Euro may, or may not,get broken up at some point,
but that won't happen overnight...

In fact flash has a broadly similar take on it, not entry wise obviously,
but he has adopted a counter trend methodology that does work well
with currencies for the above reasons.

Any approach has to be market specific.
Same method different market, stop size for one will be market specific.
That is without going into session timing and other key timed events.

Random entry for someone who is available (either by choice or circumstance)
is lunacy, providing they are more efficient in the long run...

At the risk of repeating myself, due to binary logic, random entry is a good basis for automation,
if you choose, or are unable to be present at the optimum session times, nothing more.

Nice post LV. Market specific tactics are all important as far as I'm concerned - markets phase for sure, but currency pairs have a remarkable ability to exhibit enduring patterns and relative volatility thereon (the actual reasons for this are complex - number and size of speculators, relative parity of true currency circulation of the cross, CB aggressiveness, regularity of hedging orders from Asia etc etc) - I use diametrically opposed methods to trade the EU and the UJ. Despite the dollar correlation they are chalk and cheese with regards to intra-day movement and have been for some time. The EU - sudden move - consolidation - sudden move - consolidation - night time sleep. The UJ - all sorts of craziness all day long.

However, I don't actually care what the reasons are for them or any other market exhibiting whatever behaviour they do, nor have I ever cared to think too much about randomness in markets. Macro men confuse me generally, I'm a bit simple. Just do what works.

 

[Liquid validity]

Nice post LV. Market specific tactics are all important as far as I'm concerned - markets phase for sure, but currency pairs have a remarkable ability to exhibit enduring patterns and relative volatility thereon (the actual reasons for this are complex - number and size of speculators, relative parity of true currency circulation of the cross, CB aggressiveness, regularity of hedging orders from Asia etc etc) - I use diametrically opposed methods to trade the EU and the UJ. Despite the dollar correlation they are chalk and cheese with regards to intra-day movement and have been for some time. The EU - sudden move - consolidation - sudden move - consolidation - night time sleep. The UJ - all sorts of craziness all day long.

However, I don't actually care what the reasons are for them or any other market exhibiting whatever behaviour they do, nor have I ever cared to think too much about randomness in markets. Macro men confuse me generally, I'm a bit simple. Just do what works.

Good post, dug this thread up as I've been busy lately, and it deserved a reply
Completely agree with all your points, particularly the bold highlight.

Agree on randomness as well actually, my take on it is more the relationship
between automation and price action.

For me, thus far, adopting an element of randomness has yielded the best results,
simply as I do not possess the skill to program anything that can read price any better.
Thats largely down to issues such as intra bar granularity.

As far as markets are concerned, they have their moments of randomness.
Overall, they are far from random

 

[android8]

Let's start again?
Would you actually flip a real coin every day, Heads on my lucky £2 coin is Buy!!!!

Well I going to use Excel, =RANDBETWEEN(1,100)
and then =IF(B2>=50,"Buy","Sell")
and then Press F9

 

[android8]

O.K. GMT 6:40,

Buy Eur/Usd, TP 8 pips, SL 64 pips

Sell Gbp/Usd, TP 8 pips, SL 64 pips

--------------------------------------------------------

Oh that's nice -56 pips on day one.

 

[android8]

Buy Gbp/Usd, TP 8 pips, SL 64 pips

Sell Eur/Usd, TP 8 pips, SL 64 pips

-----------------------------------------------

+16, Now ( -40 )

 

[Forexmospherian]

Let's start again?
Would you actually flip a real coin every day, Heads on my lucky £2 coin is Buy!!!!

Well I going to use Excel, =RANDBETWEEN(1,100)
and then =IF(B2>=50,"Buy","Sell")
and then Press F9

Hi Android

Well Happy New Year to you - but its a shame as it just will not be profitable year doing this exercise - and I am sure you will realise it when you sober up sometime to day.

Lets look at what you are doing for a start -

First you are using a traditional way of getting what should be approximately 50/50 results - - ie over a 1000 coin tosses you should in theory have roughly 500 "heads" results and 500 "tail" results.

OK - you then try and connect it to buying and selling - yet again- no real connection or association - so making the possibilities of winning even less - and then to finally ruin it - you do a naff money management trick of a small target - ie 8 pips and a large stop - ie 64 pips

A massive recipe for disaster - so we can tell you the results in advance whether over 100 trades or even 10,000 trades - ie a LOSS.

Your MM means you need a 90% win ratio - to just about make a small profit or break even - and even if you coin as a load of bias in it - its not going to throw more than 75% of the same - and then with you connecting to the market - which could be totally out of sync with your coin

I don't think you will have extremes - ie less than 20% wins and 80% losses - giving you massive losses - or even 80% wins - giving you just losses - but there is only one word that seems to stand out - ie losses

Is there a way to get this to work ?

Shame - but......... NO

Still I look forward to your results as there will be days you get 100% wins - lol

Regards

Forexmospherian

 

[android8]

Hello and Thank You, Forexmospherian

This is just the response I was looking for,
now before you pass me off as some nutcase, one of the things I'm looking is a system that loses as much as possible.

Lets look at what you are doing for a start,

A massive recipe for disaster....

But What I'm I Not Doing !!!!!!!!!!!!!

We don't have to look far for a 100 losing systems, so why can't we turn one around, in my case if it's such a disaster waiting to happen, why can't we sell when were meant to buy and have a S.L. of 8 pips, and a T.P. of 64 pips, Yes Total Success

Please may I have Your thoughts

Android8

 

[Forexmospherian]

Hi Android8

Yes I would love to give you my thoughts - mainly based on my trading journeys in the Forex retail market over the last 11+ years.

Ideally just spend between 5k -10k hrs studying PA and price structure on live small frame charts.

Its takes many years - but by the end of it - you will be able to read price better and take trades that will have a 60%+ probability of working - and then with the correct MM - you will make money on a fairly consistent basis.

If you have not got say 3-4 yrs full time to do this - or 6 -10 yrs part time - you will really never get there - purely based on the fact that under 20% of retail forex traders are consistently profitable and its certainly under 10% who are churning out profits - week in week out.

OK - if you say - I have not got that silly time to do that - I want to be an expert in 18 months or 2 years - well you will fail on anything you might do - although I might give you a 1-3% chance of making it - if you are good and lucky ;-)

So how can we get you into that 1-3% chance of making money ongoing - without doing the time and study and conditioning needed to get there -

1. Don't relate taking a trade to non related things - ie coin tossing - days of the week with an e in them - fridge magnets - raining or dry - time of the day the dog poo's - etc etc.

2. Be realistic with your targets and stops if you are guessing - ie 15 pip stop and say 20 -35 pip target have a lot more chances of working then 8 pip stop and 64 pip target.

I should know this fairly well as all my trades are with stops under 7 pips max - many taken with 3-5 pip stops when spreads and commissions are under 1 pip cost. The tighter the stop the quicker to a RR over 1 = which is what you need with over 60% success rate really.

Forget 20 pip stops and 40 -60 pip targets - 50 /50 again - need luck and normally the market will either do 22 -28 pip and then take you out the other way. Remember the market makers have supercomputers to work out all the minute manipulations needed to "**** up" most stop / target ratios ;-)

ie 5 pip stop - and 8 pip target - and 60%+ win ratio and say 15 -30 trades a day - and you make money.

5 pip stop and 11 pip target is hard to do - more difficult to achieve than 5 pips stop and 6 pip target - but that's not efficient enough for what you need - 10 pips stop and 18 -25 pip target is OK - but that limits your trades you can take.

Now how can we get you taking say 15 -30 trades a day to achieve small targets.?

How about using small frame charts - ie tick to 5 mins and then look at say 3 pairs - with 2 pairs taking the same direction - with the 3 rd pair being a similar direction correlated pair - like say EU with GU or with EA or EJ - as examples.

Then every hr on and around the hour - take 1 or 2 trades - along with the 3 trade on the half hr on the different pair.

You will need at least 4 hrs a day - but that can be over say 10 hrs from start of London Open to past London Close.

Decide you trade directions - ie buys or sells - on the opposite to whats happened in last 30 mins as long the the pair as moved over 20 pips ;-)

Do this say 100 times - and please tell me what results you get ;-)

Good Luck

Regards

F

 

[Pat494]

OK so we don't take commission and slippage into consideration.

What if we double up after each 100 point loss ?
And have an inexhaustable supply of money i.e. borrow a QE printing press, thanks Benny

 

[android8]

Thank You Forexmospherian

You Have been more realistic and a lot off help, I will try out something more along your lines and let you know the out come.
Pat494's head must still be sore from the night before, but hay you always get one...

Android8

 

[spreader_legger]

Hi Forexmorphorian. I guess your system in principle would work with DMA .... not sure about spread betting .... especially with such a narrow TP & SL in place. Moreover in fast moving markets you might find yourself substantially off your S/L before you can take the hit.

As for a fixed T/P or S/L I personally am not enamored by anything fixed in a fluid and ever changing market .... but each to our own

Android .... in principle you can make money on a coin flip entry .... but then it will come down to your money management.

 

[android8]

Sell Gbp/Usd, TP 8 pips, SL 56 pips

Sell Eur/Usd, TP 8 pips, SL 56 pips

-----------------------------------------------

+16, Now ( -24 )

 

[Forexmospherian]

Hi Forexmorphorian. I guess your system in principle would work with DMA .... not sure about spread betting .... especially with such a narrow TP & SL in place. Moreover in fast moving markets you might find yourself substantially off your S/L before you can take the hit.

As for a fixed T/P or S/L I personally am not enamored by anything fixed in a fluid and ever changing market .... but each to our own

Android .... in principle you can make money on a coin flip entry .... but then it will come down to your money management.

Hi S _ L

I would agree with most of what you have said.

I do not do automation etc - but instead prefer to sit in front of the charts - 4 - 8 hrs normally a day and trade in the "now" ;-) ( and also learn more ;-) )

I am full time so its no problem and even a 35 hr week seems easy compared to the 50 -70 hr weeks in the corporate world.

I have fluid targets - the market is dynamic - but all stops have to be under 7 pips at entry - ( set in stone - normally under 5 pips on key pairs with spread in there as well) - otherwise - no trade - will wait for next one

Success win rate averages over 70 % + on over last 7000+ live trades (have taken over 13k live trades in total)

Its amazing when you have a larger retail account ( say over $50k ) and you trade a lot and then complain to your broker about negative slippage all the while.

it just seems to disappear and now and again - you even get positive slippage ;-))

I wonder why ? - yes i am sure we all know ;-))

Regards

F

 

[spreader_legger]

Yup. Fast moving markets can also move in one's favor. Well done by the way! I am glad you have found success. Personally I do not trade full time nor do I want to (it's really quite boring. .... And the girls are not all that interested either ;-)) I tend to trade when the opportunity presents itself.

By the way, I tend to have no stops. As the pros say - stops are for buses, spreads are for sandwiches
.