D'Alembert-Based Trading Systems
Interesting
I like the looks of it. I am not sure where I would stop the progression and start over at 1 unit if I had a very long losing streak. Even at 50% accuracy, the potential losing streak could really add up, and then success would hinge on having enough wins in sequence to cover those losses. Any suggestions?
I used a similar system playing online Blackjack and it worked really well. In fact, it worked so well the site (no longer in business) closed down and took my money with it. I had turned $80 into $600. But those losing streaks were really hair raising! It wasn't so bad because I was really gambling with my winnings and not my original stack. I know I could not, would not take that kind of risk with my own money.
A few Betting Systems to think about
Hi Guys,
A few Betting Systems to think about:
The 1-3-2-6 System
A progressive system like this one makes assumptions about your winning streaks and rewards you accordingly. As an outcome these systems don't work well when your trading doesn't have winning streaks, but in reality consistent winning after losing after winning is unlikely. As usually, you must add an element of creativity and see for yourself what is more likely to happen in your trades.
The first bet is one unit, if you win, add another unit, making the second bet a total of three units. If you win the second bet, there are 6 units on the table. Remove four units, making your third bet two units. If it wins add two more units, making it a total of six units for your fourth bet. If the fourth bet wins you'll collect a total of 12 units, making 14 units in profit if you win at all stages! If you lose the first bet, the loss is one unit. Assuming you win the first, but lose the second bet, your net loss is two units. If you win the second, but lose the third bet, you have a profit of two units. If you win the first three bets but lose the fourth bet, you'll break even. If you lose the second bet five out of six times and win four consecutive bets once, you'll be right back to even. This system is not unbeatable, but it does give you a strategy to play to, which in the long run could save your bankroll more than a trader who has no strategy and whose staking remains variable.
Positive progression betting systems: You increase the bet when you win. Kind of 'let-it-ride'. Require less capital and usually employed to take advantage of winning streaks. (No sweat, easy ride.)
This is another simple system,once a bet is won you leave the winnings there to go on the next bet "let it ride". Therefore in the next bet if you lose you only has lost your first bet, if the bet is won you get 4 units back.
This bet does not have to be your entire bankroll in one.There is a method based on the player having 31 units to stake where you have a series which is 1 1 1 2 2 4 4 8 8 (for a total of 31 units). Your first bet is one unit. This could be any bet from a dollar to 1/16 of the table limit. If that bet wins, you bet the same amount again. If it loses you go up the series with a parlay on each bet. Notice that if you win any of the parlays you will be winning anyplace from one to nine units. In a parlay game, the player makes a successful bet and uses his winnings together with the initial stake in the next game for the same wager. If you win again, you have won a simple parlay game..
This system only works successfully if there is a long winning series of an even money bet like Black/Red etc. At the beginning of a parlay game, losses occur frequently until a multiple parlay bet wins and compensates for previous losses.
I will post more after we digest these.
DocB
Hi Guys,
I think we have the potential for a very interesting new thread here 
whilst there are a lot of martingale ea's which can be quite risky no discussion has been give to d'Alembert and other gambling/money management systems.
I strarted to code a d'Alembert EA but has a few bugs if anyone can tweak, will post it when I have access tio my other PC
Thank you all for your responses.
DocB's 1-3-2-6 progression is new to me, it looks interesting, too.
What I'd like to aim for personally right now is a simple trading technique(s) to begin to explore any potential this may have.
I've been experimenting with a simple breakout system that looks promising.
It's a simple strategy that tests very well on non-Forex testing platforms. Please see my original posts: https://www.mql5.com/en/forum/general
I have possibly been fixated on the notion that a system of this nature must require an equal T/P and S/L.
Maybe this is a false notion...?
Comments are welcome.
I have as yet been unable to obtain an EA for the above strategy; I've taken some test results, entered them into a spreadsheet and am applying some straight D'Alembert progessions to the data. I will try to post the results, hopefully with EQ curve pics.
Assuming an EA, I'd like to first test the profitablilty of the system on it's own, and then add the trend filter that Simba has posted. This may provide the "edge" that could put the win rate at or over 50%.
If a system produces an absolutely flat equity curve, this is great!
It would do the job. I'm trying to make the point that with a reasonable betting progression, what might have once been considered a crummy system may now have new potential... if you have any techniques, or better, EA's that "only" did 50% or just made a small profit, please trot them out and give them a second look.
Please share your thoughts and ideas.
Thanks for your response.
You are correct, a large drawdown can still occur at any time.
I have a couple of ideas on this, one was in the Binary Thread about delaying the beginning of the progession. That one has other problems inherent in it but please give it some consideration.
The other idea is to not increase the bet by full units. Instead of the progression 1, 2, 3, etc., use micro or mini lots. For example, 1.0, 1.1, 1.2, 1.3. etc. or .10, .11, .12, .13. This is still effective and would allow for very large losing streaks; granted, the resultant profit is smaller.
This is something you can't do in the casino...!
WNW,
Betting systems are designed for fixed odds play. You bet $100 on Red and you win $100 plus get your original $100 back. I think one must apply the same fixed odds approach to using betting systems in FX. In other words the TP and SL must be equal ( taking into account spread , commission etc.; The House) The advantage of FX is the ability to use fractional lots (mini or micro), large leverage, and to be simultaneously long and short without having to offset positions. We have to find ways to use this to our advantage. The higher the winning percentage the better but more important is the maximum number of wins and losses in a row. Most of these betting systems call for a winning streak to make you money. Please pay attention to Maximum Consecutive wins consecutive losses on the Meta Trader report and how many time it has those runs. I know Trade Station reports show the number of times you had 4 wins in a row, 8 wins ,etc but don't know if Meta Trader gives that info from backtesting. That might be a good place to start for a new indicator or counter. Maybe you guys have it already. A good system that can knock out a series of 4 consecutive wins is what we are looking for. It sure smells like a trend continuation or breakout system would fit the bill but I prefer counter trend or reversal systems as it is easier to define the risk or SL from a given point. Fact is just, about anything will work but it does need to be able to string the wins together.
DocB
Hi DocB,
Thanks for your reply. Your point is taken about fixed odds.
The Metatrader report does indicate max consecutive losses, but that is deceptive. Max Consec Losses may be only 8, but it could be followed by 1 win, another 8 losses, etc.. I usually try to take a full printout of trades and input the data into a spreadsheet and apply a progression to the entire data set. This gives a clear picture.
I agree with your idea of a countertrend or reversal system, but I don't have one. So far in backtesting the best results I've obtained generally have been simple breakout systems, but I'm open to anything that works.
If we could employ a trend filter I think this would help significantly no matter what the basic strategy. Don't know if you've plotted the fractal dimension indcator yet, it's pretty impressive. Could be it may work by itself as the only indicator necessary.
WNW
Ive been doing a coin flip expeirment on paper for a couple months now.
It works like this.
Head = long
Tails = short
enter market at UK session open
Profit 60 pips
stop 30 pips
trailing stop 31
thats it.
I have a excel file with the results, it has results for straight trading, as well as add one lot each loss then revert back to 1 after a win.
also double on each loss,
If the trailing stop is hit It makes a pip so thats a blah day and no add or take away.
I started this with a 30/30 TP and SL and it was going nowhere, it works better with the larger TP, It has been suggested that I use it on GBP/JPY, but I think that would need even larger stop and TPs.
Also I think a back test of something like this is totally worthless, because who know if that would have been the flip results on that day.
Ive been hinting around for someone to make a decent EA of this, I did have someone make one for the basic flip but it doesnt work right, and I would like to have the ability to change between straight, or double or add one lot.
the file starts with one mini
I am also going to start rolling dice, and see how that works out. Odds is short and even is long. Does any of you gamblers know if thats stilla 50/50 chance?
I think you can tell by this that if you had a decent system your idea would increase your normal profits. Disclaimer should be obvious but if you dont have a winning method dont even consider this.
Hi droesprarky,
I've looked at the spreadsheet, this is pretty interesting.
You had 5 more long trades than shorts, but you had 5 more short winners than long. The trailing stop seems to be effective too.
Your net at the end point is almost double your worst drawdown.
The only problem is that it's a short period of data, given the daily bars,
so it's impossible to know long term results.
I'm really surprised that the large TP works better, have you looked at the chart to see if there's been a trend in place?
My other question is the trailing stop.. are you using Metatrader? If so, will it allow you to put in a custom stop? I just would like to know how you're actually implementing this.
Somewhere on this forum there may be an EA that's doing this, maybe using a random number generator. I'm going to look around.
Maybe DocB has some comments on this, I'm not good with statistics or gaming theory.
Thanks!
Hi all,
Thanks for starting this thread. Money management is something that is important and that I've been missing on this forum.
I'm not really experienced at statistics, game theory, chaos theory or anything else, I just wanted to encourage you guys to keep this thread going, it's got great potential.
Regarding the EA -- I found something in my archive. I'm not sure where it came from, I don't know what it does (because I never got it to work), but maybe it could be of any help. If someone figures this out, please post your insights.
Thanks
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
You agree to website policy and terms of use
I'm starting this thread as a both an extension of the thread "Binary Equation Trading" located here: https://www.mql5.com/en/forum and because there seems to be very little on this forum about the D'Alembert numeric progression related to trading. The D'Alembert is a negative progression like the Martingale, but much more conservative and without the dramatic drawdowns.
Let me begin by stating that I am not by any means expert on this topic, there are many more knowledgeable members on this forum whose input is very much welcomed.
I'm only interested in trading and making a profit without undue risk.
The classic D'Alembert system in betting is this: add 1 unit when you lose, subtract 1 unit when you win. When (if) you return to 1 (equilibrium) you will have made a profit.
Bet: ____________ 1 2 3 4 3 2 1
(Sorry, I can't get the columns to line up)
Notice that your net profit is equal to the largest bet in the progession.
This assumes a return to 1 betting unit which might eventually be expected in a trading system that has an equal T/P and S/L and a 50% win rate.
HOWEVER- look at the above example again. There are 3 losses, but after 2 wins, one at 4 units and another a 3 units, you have a net profit of 1 unit.
This means that a system with a 40% or better win rate given equal T/P and S/L would make a profit in the long run.
This gist of this is that you can still make a profit without a great system.
I have been experimenting with simple breakout systems that appear to have much potential, but I do not have the time for realtime testing or the expertise in coding EA's.
As a primer to this thread, please go to the link above and pay special attention to Simba's posts.
Simba has been very generous with his knowledge on this topic and has included 2 indicators as well which
look to be very useful in this area. I am including them as attachments.
All input is welcome, I would hope this thread would continue and be helpful.