Martingale recovery

[emptyjar]

I traded FX a few years ago, and made and lost relatively inconsequential amounts doing it. I've started again recently, and decided the best strategy for what I want is to use the martingale (aka doubling down) loss recovery system.

My post is to ask, firstly, how many retail FX traders use this with success, secondly, whether I will (for a reason I'll elaborate upon in a minute) be better served moving away from my current provider (IG) and thirdly, how to calculate my probability of winning (so I know when to quit).

The way I am trading at the moment is this. I'll put a SL and a limit order on for 20 pips away from where the market is when I enter at £1 a point (nb: am only trading £500 at the moment, as a test, I will up it a lot if I can get my percentages sorted out). If it loses I re-enter the market with the same SL & limit orders of 20 pips in each direction, this time at £2 a point; then £4, £8 etc etc.

My first issue is that the bid-ask spread I am receiving from IG means I am already a few pips down when I enter the market. So, it doesn't take long for me to get wiped if the positions moves against me (I'm already 15% down on a position just by opening it). To counteract this I started trading a bigger SL/limit, but this slowed down my trading speed quite a lot (which is contrary to my current philosophy- I'm trying to make quick return on whatever I put in). Furthermore, I could trade 100 pips at £1 a point, but my observation (unscientific perhaps :cheesy is that if all I ever need is a 20 pip correction to make my profit (by doubling down in a falling market my average entry price falls) then I'm more likely to see this than I am a 100 pip correction.

Anyway, what sort of interface do I need in order to be able to actually just enter a bid/accept an offer on the market myself? Admittedly I've never seen this feature for retail FX traders, but I have seen equities market traders who have a scrolling page with all the bids and offers on it and they just click to accept a trade. Presumably, given the proliferation of firms like IG, it is not possible to do this easily for forex as a small time trader or else everyone would be doing it and cutting out these bucket shops.

Another irritation with IG is that if I set an SL & limit order it prevents me from capturing the upside. So, what I would like to do is let's suppose I enter at 1.5010 and set an SL at 1.4990 and a limit at 1.5030- I would like some way of once it hits 1.5030 the limit order isn't immediately fulfilled, so once the market hits 1.5031 let's say, a sell order is placed at 1.5030- in other words I'm able to capture the upside but avoid the downside (notwithstanding slippage). So, if it hits 1.51 by the time I next actually look at my account, great, but in a worst case scenario I'll
have the profit I aimed for.

Tl;dr- how do I work out my probability of making a 25% return on my capital using a martingale recovery system. Assuming I can withstanding 6 successive losses before going bust. Secondly, what sort of interface do I need to bid directly on the market to cut out the spread.

Thanks.

 

[timsk]

Hi emptyjar,
Welcome to T2W.

Most members will advise you in the strongest terms possible not to use a Martingale recovery system - for two reasons. Firstly, it assumes that the trader has an account with an infinite amount of capital in it. If the first trade is at £5.00 per point, by trade No. 10 you’ll have to bet £2,560.00 to recover your losses and, by trade No. 20, you’ll have to bet a whopping £2,621,440.00 to recover your losses. Very few retail traders have this sort of money to gamble with. The second problem with it is that most brokers will have a house limit as to how much you can bet per trade, regardless of the size of your account.

The common riposte to these two flaws with Martingale is that the probability of experiencing 10 or more consecutive losing trades is very low. This may or may not be the case, depending on the success ratio of the methodology employed. Even so, it’s a risk not worth taking, not least because as a trader, you need to do everything in your power to minimise risk - not take unnecessarily large ones. In the famous words of John Maynard Keynes, "Markets can remain irrational longer than you can remain solvent." In other words, put Martingale to the test and, sooner or later, you will blow up your account and become insolvent.

One last point to keep in mind is the risk model adopted by most institutional traders. This requires them to decrease their risk exposure as their account equity falls (i.e. after losses), and to increase it as their account equity builds (i.e. after profits). That's the very antithesis of Martingale and is the best way to grow your account and to avoid blowing up. Sorry emptyjar, but Martingale is not the holy grail - don’t fall for it!
Tim.

 

[alexaherself]

Assuming I can withstanding 6 successive losses before going bust.

The idea of a run of consecutive losers is leading you astray, here. Your frame of reference is a mistaken one.

Either your trading system has an edge (positive next expectation), or it doesn't.

If it does, you don't need to increase the stakes when losing (and it will eventually be suicidal to do so: that's the sure way to lose money even with a system that does have an edge!).

If it doesn't, then you can't give it one by escalating the stakes during a bad run: all you can do that way is re-arrange the deckchairs on the Titanic while actually accelerating the onset of the inevitable disaster.

I offer you three items of advice, here ...

(i) Keep re-reading Timsk's excellent post just above until you understand all of it so well that you could explain it to someone else in your sleep;

(ii) You need advice about staking, position-sizing and money management: I strongly suggest reading both these two books slowly and carefully: (a) Trade Your Way to Financial Freedom (Van K. Tharp); (b) Profitability and Systematic Trading (Michael Harris). They both explain clearly and in great detail why what you're currently proposing to do will inevitably result in disaster, whether the system actually works or not;

(iii) The futility of this kind of position-sizing isn't the fact that it eventually loses you all your money (though that's undoubtedly true, too): it's the fact that you can't even learn anything about the viability of your system from it, because the inevitability of its outcome predicates that no evidence about how good the system is will have been adduced at all by the outcome - so it isn't even a valid test of a system.

 

[emptyjar]

The idea of a run of consecutive losers is leading you astray, here. Your frame of reference is a mistaken one.

Either your trading system has an edge (positive next expectation), or it doesn't.

If it does, you don't need to increase the stakes when losing (and it will eventually be suicidal to do so: that's the sure way to lose money even with a system that does have an edge!).

If it doesn't, then you can't give it one by escalating the stakes during a bad run: all you can do that way is re-arrange the deckchairs on the Titanic while actually accelerating the onset of the inevitable disaster.

I offer you three items of advice, here ...

(i) Keep re-reading Timsk's excellent post just above until you understand all of it so well that you could explain it to someone else in your sleep;

(ii) You need advice about staking, position-sizing and money management: I strongly suggest reading both these two books slowly and carefully: (a) Trade Your Way to Financial Freedom (Van K. Tharp); (b) Profitability and Systematic Trading (Michael Harris). They both explain clearly and in great detail why what you're currently proposing to do will inevitably result in disaster, whether the system actually works or not;

(iii) The futility of this kind of position-sizing isn't the fact that it eventually loses you all your money (though that's undoubtedly true, too): it's the fact that you can't even learn anything about the viability of your system from it, because the inevitability of its outcome predicates that no evidence about how good the system is will have been adduced at all by the outcome - so it isn't even a valid test of a system.

Hi. Thanks for your response.

No, I am aware that my likelkyhood of winning will impact my probability of hitting a losing streak. So, if I won 65% of the time I'd be in a far better position than simply tossing a coin (or trying to follow trends as I do now)- it is partially for this reason that I want to eliminate/cut the spread I am paying. If a position has to go against me only 17 points to lose, but has to go 23 points to win then that's a pretty appalling "house edge" and I'd be better off taking my money to the blackjack tables. That said, I just wanted to work out my probabilities assuming I win/lose 50/50. Because if I had a system that won 2/3rds of the time I'd be a trader full-time, not grabbing nickels in front of steamrollers.

Just to be clear, and I apologise because my first post was a bit rambling, but I don't see this as a long term strategy to increase my wealth; I gamble in other forms too (casino/bookies) and just want to add this to my repertoire. So, my plan is to find a better underlying system than the one I have now, and ideally one which gives me better odds than the casino, and to simply follow martingale through until I either hit my planned profit or go bust. For reference, on this present account my plan is to hit a 25% return because it's a small amount. When I do it with much larger sums I will only be seeking a 5-15% return (as I would if I backed a favorite in sports betting), or even less (1%) if it's a very large sum to me.

Cheers for the book recommendations, I'll get them. In fairness, I don't think money management is my likely downfall: I've never "bottled" out of a trade, and wouldn't. If my system says do X then I would do it. I have seen a lot of criticism of martingale online on the basis that it is psychologically quite hard to implement to the last for some folk, perhaps because they are betting the farm so to speak- which I would never do. I try to regard money I trade/gamble with as not being mine anymore, so when I lose it it doesn't hurt so much.

Either way, when I do it with a more worthwhile amount I'll link to my account (however people do that) and you can break out the popcorn and watch the show.

 

[alexaherself]

I just wanted to work out my probabilities assuming I win/lose 50/50.

If you have a 50% win-rate, then over the course of 1,000 trades, there's a greater than 25% chance that you'll have a losing run of 11 or more, and a greater than 1% chance that you'll have a losing run of 15 or more. How will that work out for you?

 

[timsk]

If you have a 50% win-rate, then over the course of 1,000 trades, there's a greater than 25% chance that you'll have a losing run of 11 or more, and a greater than 1% chance that you'll have a losing run of 15 or more. How will that work out for you?

I suspect Alexa's question at the end is a rhetorical one and, certainly, not aimed at me. Nonetheless, I thought it would be interesting to run some numbers. What I've done is to imagine that emptyjar has some astonishing good luck before experiencing the sequence of 11 consecutive losing trades. So, of his 1,000 trades with 50:50 win rate, I've assumed the first 500 trades are all winners, each netting him £20.00 profit, trading at £1.00 per pip. This means that when he experiences his first loss on trade #501, his account has grown by £10k which, when added to his opening balance of £500.00, gives him a healthy PnL total of £10,500.00. Surely, with a balance of this size, his account could withstand eleven consecutive losing trades? The table below illustrates what happens . . .



The interesting thing to note is that by the 9th trade his account is almost completely wiped out and, certainly, there won't be anything like enough margin to place the 10th trade at £512.00 per pip. Alexa doesn't mention the probability of experiencing nine consecutive losing trades but, needless to say, it will be significantly higher than the 25% probability of having eleven losing trades. Even when putting everything in emptyjar's favour with an unrealistic gain arising from 500 consecutive winners, the Martingale experiment does not end well.
Tim.

 

[emptyjar]

I suspect Alexa's question at the end is a rhetorical one and, certainly, not aimed at me. Nonetheless, I thought it would be interesting to run some numbers. What I've done is to imagine that emptyjar has some astonishing good luck before experiencing the sequence of 11 consecutive losing trades. So, of his 1,000 trades with 50:50 win rate, I've assumed the first 500 trades are all winners, each netting him £20.00 profit, trading at £1.00 per pip. This means that when he experiences his first loss on trade #501, his account has grown by £10k which, when added to his opening balance of £500.00, gives him a healthy PnL total of £10,500.00. Surely, with a balance of this size, his account could withstand eleven consecutive losing trades? The table below illustrates what happens . . .

View attachment 203586

The interesting thing to note is that by the 9th trade his account is almost completely wiped out and, certainly, there won't be anything like enough margin to place the 10th trade at £512.00 per pip. Alexa doesn't mention the probability of experiencing nine consecutive losing trades but, needless to say, it will be significantly higher than the 25% probability of having eleven losing trades. Even when putting everything in emptyjar's favour with an unrealistic gain arising from 500 consecutive winners, the Martingale experiment does not end well.
Tim.

That was interesting to see, I know my probability of hitting a 9 or 11 trade losing streak is high with 1,000 trades. So, my intention is to trade £10,000 in the following way: I risk £20 per trade on my initial trade, and thus can sustain 8 consecutive losses (though even with 8 losses I won't lose the full £10,000).

My trading strategy will be to stop when either I have made £1,000 in profit, conceded 8 consecutive losses or made 100 trades- whichever comes first. So, if I get some rum luck and it goes LOSS-LOSS-LOSS-LOSS-PROFIT several times then 80% of the 100 trade maximum will be concerned with securing £20 of profit each- in which case I'll make less than my planned £1,000. But, with 100 trades, I presume, my chances of hitting 8 losses in a row must be considerably slimmer?

Best case scenario-£1,000 profit
Worst case scenario-£5,100 loss (or more with slippage)

The £10,000 will really only be there to ensure I have sufficient margin to enter all of the trades (my largest possible trade will risk £128 a pip- or £2,560 assuming a 20 pip stop loss).

 

[tomorton]

Please don't use real £ on this, even though its an interesting exercise.

The chances of an improbable series of results are actually greater with a smaller sample size. It is the greater sample size that makes a"typical" outcome more probable.

e.g. appr 10% of the world's population is left-handed. If you take the first 10 of your friends, colleagues and neighbours that come into your mind, I would feel safe estimating that more than 1 will be left-handed. Same for the next 10 and the next. But if you could look at the world's entire population, the 10% figure will be very true.

 

[timsk]

. . . The £10,000 will really only be there to ensure I have sufficient margin to enter all of the trades (my largest possible trade will risk £128 a pip- or £2,560 assuming a 20 pip stop loss).

Hi emptyjar,
Well, so long as you understand the risks you're taking and, if it goes wrong, you can walk away from it with a smile on your face - then the best of luck to you. The idea of risking £10k to make £1k certainly doesn't appeal to me!

Before doing it, may I suggest you give serious thought to Tom's comments above and, also, to imagine how you'll feel towards the back end of a losing streak - if you're unlucky enough to hit one. If it were me, placing the 8th trade at £128.00 per pip - I'd be cacking myself!

Tim.

 

[emptyjar]

Please don't use real £ on this, even though its an interesting exercise.

The chances of an improbable series of results are actually greater with a smaller sample size. It is the greater sample size that makes a"typical" outcome more probable.

e.g. appr 10% of the world's population is left-handed. If you take the first 10 of your friends, colleagues and neighbours that come into your mind, I would feel safe estimating that more than 1 will be left-handed. Same for the next 10 and the next. But if you could look at the world's entire population, the 10% figure will be very true.

I guess I didn't consider the 'law of large numbers' angle. I'm not statistician, but wouldn't having a small sample size simply mean my results will be less predictable, rather than that the risk will increase? So, with the 10 friends example- there's a good probability 2 may be left handed, i.e. above the statistical norm, but also a probability that 0 will be, i.e. below the statistical norm?

I know from my experience with martingale in casinos that it's necessary to fight the human tendency to discount the unlikely as the impossible. So I've experienced 7 losing hands in a row playing blackjack-it felt like awful luck at the time, but the reality was that this is almost guaranteed to happen every X number of hands played.

My point with 100 trades being that if I decrease the size of my trading enough then I must be limiting, though not eliminating, the risk of having 8 losing trades in a row. If I only did 20 trades then it would seem very unlikely to hit 8 losers, if I did only 7 trades then it would be impossible (but of course I'd be unable to recoup my losses if I hit several losers in a row). If I did 0 trades I couldn't lose, but nor could I win, and likewise if I did 10,000 trades then, I presume, the odds of me hitting 8 consecutive losses would approach 100% and it would just be a question of when in the 10,000 trades I hit the streak whether I made an overall profit or an overall loss.

 

[tomorton]

Please stop now emptyjar, before you convince yourself you're right, and therefore you have to do this.

 

[mpups]

...my plan is to hit a 25% return because it's a small amount. When I do it with much larger sums I will only be seeking a 5-15% return....

Or in other words you are considering it a long term strategy of making money.

It wont work mate.

 

[Sonicscooter]

If only trading was so easy.

 

[prabhanshuraman]

Nice discussion guys....good to be here...!!

 

[tomorton]

I think it is very risky to trade with any type of martingale strategy. This strategy will eventually eat up your account and thus it is very risky. One must first try this strategy on the demo accounts provided by some reputed company like "Signalstime" and then only proceed with real trading.

Waste of time to even demo a Martingale strategy. Its not just "very risky", it doesn't work - its about as relevant to trading as reading tea-leaves or tossing a coin.

mcrlogitech is just trying to drive traffic towards his own site, most of his messages doing this have already been deleted.

His messages on trading are vague and straight out of a child's level book on trading, of absolutely no value to any trader coming to T2W with a genuine need. There is no evidence he has any personal experience in trading of any value. Before it was deleted, his first message on here was only from August when he was apparently wishing to start trading and needed a strategy. Take no trading advice from this timewaster.

 

[Pat494]

Martingale is a killer of your bank account. Don't even think of it.

 

[ffsear]

As displayed above (depending on what you trade) By trade 7 or 8 you'll be unlikely to cover the initial margin payment anyway and its game over!