Using Martingale with a proven strategy.

afedoro · Oct 22, 2005

Quote from ssigma:

Now, the reason I am considering adding contracts after losses is to limit my total exposure for each trade.

I've found that if I initially start with 1 contract and add 1 after each loss, my total perfomance per year is roughly the same as trading 6 contracts all the time.

If this is the case, doesn't it make sense to use this martingale type money management method so I am not fully exposed everytime I'm in a trade?

Thoughts, comments?
More...

Well, your logic kind of makes sense for 99.9% of the time but the issue here is that you still have to take into consideration the remaining 0.1% of the time (please don't attach importance to the exact numbers), since it can and will happen.

When trading, you have to have a strategy that can handle the full distribution of possible outcomes, including tails. To top it off, the fact is that fat tails are common in finance so any strategy that cannot handle tail events will eventually blow up.

I hope this makes sense! :|

Bottom line is don't do it. If your strategy is profitable on its own and with the time frame you mentioned, trade it as such and don't use any strategies that will result in over-leveraging and an eventual blowup.

Good luck.

stephencrowley · Oct 22, 2005

I'd just like to add that the expected outcome of a martingale strategy is exactly the same as the whatever system without martingale over the longrun. The only thing you are ensuring by using this strategy is that you will blow up much more quickly. It will happen. I suggest picking up Fooled by Randomness by Taleb if you don't understand these concepts.

Quote from afedoro:

Well, your logic kind of makes sense for 99.9% of the time but the issue here is that you still have to take into consideration the remaining 0.1% of the time (please don't attach importance to the exact numbers), since it can and will happen.

When trading, you have to have a strategy that can handle the full distribution of possible outcomes, including tails. To top it off, the fact is that fat tails are common in finance so any strategy that cannot handle tail events will eventually blow up.

I hope this makes sense! :|

Bottom line is don't do it. If your strategy is profitable on its own and with the time frame you mentioned, trade it as such and don't use any strategies that will result in over-leveraging and an eventual blowup.

Good luck.
More...

DTK · Oct 22, 2005

Quote from LoosenUp:

If your success rate is only 40%, you can expect a much higher consecutive losers in the not so distant future.
More...

Very good point. With even distribution and 50% success rate it is possible to have 10+ losers (or winners) in a row.

There are plenty of books on money management (or with a good money management section) and I don't remember any of them mentioning martingale.

marketsurfer · Oct 22, 2005

as long as you don't "over martingale" yourself, its a very viable strategy. it can not ruin you if you stay within your max size/martingal level which is keyed to your account size.

best,

surfer

ssigma · Oct 22, 2005

Quote from DTK:

Very good point. With even distribution and 50% success rate it is possible to have 10+ losers (or winners) in a row.

There are plenty of books on money management (or with a good money management section) and I don't remember any of them mentioning martingale.
More...

A true martingale system doubles contract size after each loss. However, the winners with this system are 3x losers, allowing me to significantly slow the rate at which contract size needs to be increased.

Backtesting has shown max consecutive losers to be 5 (which only happened twice) I am prepared to expect 10 losers in a row. Anything more than 10 would probably mean the system itself has no edge.

If a system has an "edge" to begin with, does it make sense to employ aggressive money management to IMPROVE overall results..... by increasing exposure after losing trades?

Has anyone had success or experience with this?
Do successful hedge funds use this technique?

Choad · Oct 22, 2005

40yotrader apparently was able to start a new career and do well with variations of the martingale. And is still doing fine AFAIK.

Note that you do NOT have to double up until your account is blown! That would be kinda dumb, huh? If you have a good system, and you think of each trade as a discrete event, i.e. just another trade, buying more as a stock drops (for example) can increase winning potential. Years of testing and thousands of real trades have convinced me. BUT this is with my systems and trading style! Others will have different opinions, which is okay by me...

There are many threads on ET about "adding to losers" or the like. Do a search and you will find as many different opinions as there are ET traders.

NOTE: Of course it has to fit in with your overall WELL THOUGHT OUT and conservative risk management strategy. Or you WILL wipe out. JMHO

http://www.elitetrader.com/vb/showthread.php?s=&threadid=10237&perpage=6&pagenumber=9

psytrade · Oct 23, 2005

ssigma, Im going to show you an SP500 stock trading system based on 40yrs of backtested data.

One system I could show you uses a different stop than the other, but exhibits similar returns and trade #s. However, the 1st is only a slightly less profitable system and wins only 22% of the time.

System 1 loses up to 58 times in a row. Both systems trade so that no more than 5% of capital is in a trade (stock), and no more than .5% is risked.

System 2 loses up to 17 times in a row. The system is the same as 1 except for a looser stop.

You will see that there is something to be said for being wrong 17 or 58 times in a row. The drawdown was still very similar. In System 2's case, 51%. In the past 15 yrs the drawdown was under 40%. And the return decays because compounding 30% for 40yrs means that you are trading a fortune. The account went from $10 mil to $130 bil. Not very likely, but thats what happen if you compounded it tax free. If you want less of a drawdown, I can cut the leverage down to 75 or 100% margin from 50%.

Also the systems had over 3000 and 5000 trades respectively. 3007/17 = 177 You have a 1/177 chance of having more than 17 losers with this system. I didnt correlate the loser sequences to the drawdown, but its likely directly related. Systems get tested during these drawdowns, and this is the outcome. The thing that matters is what drawdown you can take, and that may be independent of the sequences of wins and losses.

System 1:

Statistics
All trades Long trades Short trades
Initial capital 10000000.00 10000000.00 10000000.00
Ending capital 89816095839.38 89816095839.38 21344232.79
Net Profit 89806095839.38 89806095839.38 11344232.79
Net Profit % 898060.96 % 898060.96 % 113.44 %
Exposure % 64.06 % 64.06 % 0.00 %
Net Risk Adjusted Return % 1401812.45 % 1401812.45 % N/A
Annual Return % 28.93 % 28.93 % 2.14 %
Risk Adjusted Return % 45.16 % 45.16 % N/A

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

All trades 5851 5851 (100.00 %) 0 (0.00 %)
Avg. Profit/Loss 14147219.78 14147219.78 N/A
Avg. Profit/Loss % 18.19 % 18.19 % N/A
Avg. Bars Held 85.82 85.82 N/A

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

Winners 1350 (23.07 %) 1350 (23.07 %) 0 (0.00 %)
Total Profit 91236468829.49 91236468829.49 0.00
Avg. Profit 67582569.50 67582569.50 N/A
Avg. Profit % 94.86 % 94.86 % N/A
Avg. Bars Held 307.79 307.79 N/A
Max. Consecutive 73 73 0
Largest win 7717030201.35 7717030201.35 0.00
# bars in largest win 403 403 0

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

Losers 4501 (76.93 %) 4501 (76.93 %) 0 (0.00 %)
Total Loss -8461085900.55 -8461085900.55 0.00
Avg. Loss -1879823.57 -1879823.57 N/A
Avg. Loss % -4.81 % -4.81 % N/A
Avg. Bars Held 19.24 19.24 N/A
Max. Consecutive 58 58 0
Largest loss -159066441.89 -159066441.89 0.00
# bars in largest loss 4 4 0

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

Max. trade drawdown -13435527003.71 -13435527003.71 0.00
Max. trade % drawdown -75.33 % -75.33 % 0.00 %
Max. system drawdown -24839089327.43 -24839089327.43 0.00
Max. system % drawdown -53.70 % -53.70 % 0.00 %
Recovery Factor 3.62 3.62 N/A
CAR/MaxDD 0.54 0.54 N/A
RAR/MaxDD 0.84 0.84 N/A
Profit Factor 10.78 10.78 N/A
Payoff Ratio 35.95 35.95 N/A
Standard Error 19041265552.82 19041265552.82 403851.00
Risk-Reward Ratio 0.09 0.09 0.80
Ulcer Index 14.45 14.45 0.00
Ulcer Performance Index 1.63 1.63 N/A
Sharpe Ratio of trades 0.21 0.21 0.00
K-Ratio 0.01 0.01 0.08

System 2:

Statistics
All trades Long trades Short trades
Initial capital 10000000.00 10000000.00 10000000.00
Ending capital 130473221782.16 130473221782.16 21344232.79
Net Profit 130463221782.16 130463221782.16 11344232.79
Net Profit % 1304632.22 % 1304632.22 % 113.44 %
Exposure % 63.41 % 63.41 % 0.00 %
Net Risk Adjusted Return % 2057555.80 % 2057555.80 % N/A
Annual Return % 30.28 % 30.28 % 2.14 %
Risk Adjusted Return % 47.76 % 47.76 % N/A

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

All trades 3007 3007 (100.00 %) 0 (0.00 %)
Avg. Profit/Loss 39356104.64 39356104.64 N/A
Avg. Profit/Loss % 43.59 % 43.59 % N/A
Avg. Bars Held 216.14 216.14 N/A

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

Winners 1571 (52.24 %) 1571 (52.24 %) 0 (0.00 %)
Total Profit 126178783469.90 126178783469.90 0.00
Avg. Profit 80317494.25 80317494.25 N/A
Avg. Profit % 93.69 % 93.69 % N/A
Avg. Bars Held 341.88 341.88 N/A
Max. Consecutive 56 56 0
Largest win 11146287097.40 11146287097.40 0.00
# bars in largest win 797 797 0

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

Losers 1436 (47.76 %) 1436 (47.76 %) 0 (0.00 %)
Total Loss -7834976824.83 -7834976824.83 0.00
Avg. Loss -5456112.00 -5456112.00 N/A
Avg. Loss % -11.22 % -11.22 % N/A
Avg. Bars Held 78.58 78.58 N/A
Max. Consecutive 17 17 0
Largest loss -199173530.50 -199173530.50 0.00
# bars in largest loss 162 162 0

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

Max. trade drawdown -13510068940.55 -13510068940.55 0.00
Max. trade % drawdown -90.91 % -90.91 % 0.00 %
Max. system drawdown -31727857761.18 -31727857761.18 0.00
Max. system % drawdown -51.10 % -51.10 % 0.00 %
Recovery Factor 4.11 4.11 N/A
CAR/MaxDD 0.59 0.59 N/A
RAR/MaxDD 0.93 0.93 N/A
Profit Factor 16.10 16.10 N/A
Payoff Ratio 14.72 14.72 N/A
Standard Error 28899515812.94 28899515812.94 403851.00
Risk-Reward Ratio 0.10 0.10 0.80
Ulcer Index 11.84 11.84 0.00
Ulcer Performance Index 2.10 2.10 N/A
Sharpe Ratio of trades 0.25 0.25 0.00
K-Ratio 0.01 0.01 0.08

CPTrader · Oct 23, 2005

Inetresting analysis, what software produced these reports?

Thx

Quote from psytrade:

ssigma, Im going to show you an SP500 stock trading system based on 40yrs of backtested data.

One system I could show you uses a different stop than the other, but exhibits similar returns and trade #s. However, the 1st is only a slightly less profitable system and wins only 22% of the time.

System 1 loses up to 58 times in a row. Both systems trade so that no more than 5% of capital is in a trade (stock), and no more than .5% is risked.

System 2 loses up to 17 times in a row. The system is the same as 1 except for a looser stop.

You will see that there is something to be said for being wrong 17 or 58 times in a row. The drawdown was still very similar. In System 2's case, 51%. In the past 15 yrs the drawdown was under 40%. And the return decays because compounding 30% for 40yrs means that you are trading a fortune. The account went from $10 mil to $130 bil. Not very likely, but thats what happen if you compounded it tax free. If you want less of a drawdown, I can cut the leverage down to 75 or 100% margin from 50%.

Also the systems had over 3000 and 5000 trades respectively. 3007/17 = 177 You have a 1/177 chance of having more than 17 losers with this system. I didnt correlate the loser sequences to the drawdown, but its likely directly related. Systems get tested during these drawdowns, and this is the outcome. The thing that matters is what drawdown you can take, and that may be independent of the sequences of wins and losses.

System 1:

Statistics
All trades Long trades Short trades
Initial capital 10000000.00 10000000.00 10000000.00
Ending capital 89816095839.38 89816095839.38 21344232.79
Net Profit 89806095839.38 89806095839.38 11344232.79
Net Profit % 898060.96 % 898060.96 % 113.44 %
Exposure % 64.06 % 64.06 % 0.00 %
Net Risk Adjusted Return % 1401812.45 % 1401812.45 % N/A
Annual Return % 28.93 % 28.93 % 2.14 %
Risk Adjusted Return % 45.16 % 45.16 % N/A

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

All trades 5851 5851 (100.00 %) 0 (0.00 %)
Avg. Profit/Loss 14147219.78 14147219.78 N/A
Avg. Profit/Loss % 18.19 % 18.19 % N/A
Avg. Bars Held 85.82 85.82 N/A

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

Winners 1350 (23.07 %) 1350 (23.07 %) 0 (0.00 %)
Total Profit 91236468829.49 91236468829.49 0.00
Avg. Profit 67582569.50 67582569.50 N/A
Avg. Profit % 94.86 % 94.86 % N/A
Avg. Bars Held 307.79 307.79 N/A
Max. Consecutive 73 73 0
Largest win 7717030201.35 7717030201.35 0.00
# bars in largest win 403 403 0

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

Losers 4501 (76.93 %) 4501 (76.93 %) 0 (0.00 %)
Total Loss -8461085900.55 -8461085900.55 0.00
Avg. Loss -1879823.57 -1879823.57 N/A
Avg. Loss % -4.81 % -4.81 % N/A
Avg. Bars Held 19.24 19.24 N/A
Max. Consecutive 58 58 0
Largest loss -159066441.89 -159066441.89 0.00
# bars in largest loss 4 4 0

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

Max. trade drawdown -13435527003.71 -13435527003.71 0.00
Max. trade % drawdown -75.33 % -75.33 % 0.00 %
Max. system drawdown -24839089327.43 -24839089327.43 0.00
Max. system % drawdown -53.70 % -53.70 % 0.00 %
Recovery Factor 3.62 3.62 N/A
CAR/MaxDD 0.54 0.54 N/A
RAR/MaxDD 0.84 0.84 N/A
Profit Factor 10.78 10.78 N/A
Payoff Ratio 35.95 35.95 N/A
Standard Error 19041265552.82 19041265552.82 403851.00
Risk-Reward Ratio 0.09 0.09 0.80
Ulcer Index 14.45 14.45 0.00
Ulcer Performance Index 1.63 1.63 N/A
Sharpe Ratio of trades 0.21 0.21 0.00
K-Ratio 0.01 0.01 0.08

System 2:

Statistics
All trades Long trades Short trades
Initial capital 10000000.00 10000000.00 10000000.00
Ending capital 130473221782.16 130473221782.16 21344232.79
Net Profit 130463221782.16 130463221782.16 11344232.79
Net Profit % 1304632.22 % 1304632.22 % 113.44 %
Exposure % 63.41 % 63.41 % 0.00 %
Net Risk Adjusted Return % 2057555.80 % 2057555.80 % N/A
Annual Return % 30.28 % 30.28 % 2.14 %
Risk Adjusted Return % 47.76 % 47.76 % N/A

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

All trades 3007 3007 (100.00 %) 0 (0.00 %)
Avg. Profit/Loss 39356104.64 39356104.64 N/A
Avg. Profit/Loss % 43.59 % 43.59 % N/A
Avg. Bars Held 216.14 216.14 N/A

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

Winners 1571 (52.24 %) 1571 (52.24 %) 0 (0.00 %)
Total Profit 126178783469.90 126178783469.90 0.00
Avg. Profit 80317494.25 80317494.25 N/A
Avg. Profit % 93.69 % 93.69 % N/A
Avg. Bars Held 341.88 341.88 N/A
Max. Consecutive 56 56 0
Largest win 11146287097.40 11146287097.40 0.00
# bars in largest win 797 797 0

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

Losers 1436 (47.76 %) 1436 (47.76 %) 0 (0.00 %)
Total Loss -7834976824.83 -7834976824.83 0.00
Avg. Loss -5456112.00 -5456112.00 N/A
Avg. Loss % -11.22 % -11.22 % N/A
Avg. Bars Held 78.58 78.58 N/A
Max. Consecutive 17 17 0
Largest loss -199173530.50 -199173530.50 0.00
# bars in largest loss 162 162 0

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

Max. trade drawdown -13510068940.55 -13510068940.55 0.00
Max. trade % drawdown -90.91 % -90.91 % 0.00 %
Max. system drawdown -31727857761.18 -31727857761.18 0.00
Max. system % drawdown -51.10 % -51.10 % 0.00 %
Recovery Factor 4.11 4.11 N/A
CAR/MaxDD 0.59 0.59 N/A
RAR/MaxDD 0.93 0.93 N/A
Profit Factor 16.10 16.10 N/A
Payoff Ratio 14.72 14.72 N/A
Standard Error 28899515812.94 28899515812.94 403851.00
Risk-Reward Ratio 0.10 0.10 0.80
Ulcer Index 11.84 11.84 0.00
Ulcer Performance Index 2.10 2.10 N/A
Sharpe Ratio of trades 0.25 0.25 0.00
K-Ratio 0.01 0.01 0.08
More...

psytrade · Oct 23, 2005

Amibroker CPTrader

40yotrader · Nov 5, 2005

I'm sure you guys realize if you're using Fixed Ratio MMgt then you're using a Martingale strategy (increasing size on losing trades relative to the account size).

43yotrader