Probability, Pyramiding and Money Management

OK. Trading has been compared to gambling many times. Professional gamblers prefer to think in terms of probability and betting management.

So both have similarities. Both attract somewhat similar personalities. Although one is widely accepted internationally for trade reasons and can also be more of an intellectual pursuit as well.

Both entail "predicted expectation" (expectation of outcome) and "management of risk capital" to be consistently profitable over time.

Now that that preface is out of the way...

I discovered some results of a new EA I'm working on and it led me to think of another trading paradigm.

Profitable trading could be attained if one simply has higher statistical expectation of one of two possible events.

Some people say that the markets go up, down or sideways. To simply this, the markets go up or down. To make money on a single trade. You make money when the markets go up or down and keep your money if they stay exactly the same.

So the expected outcome of the market is either A or B.

And the expected result is either plus or minus.

So being profitable on a single trade simply means knowing where the market will go before the other.

For example, the market will go to A or B. You can simply this by saying that the market is either going to hit your stop level or your take profit. It's either/or. It will do one or the other, but not both on any given single trade.

A is your take profit, and B is your stop level. You just need to have confidence in which it will hit first.

Assuming you can afford for the market to hit your stop and still continue to trade, you still have a chance to continue to succeed as long as your A and B price points are statistically tradeable and financially manageable.

If you can predict with consistency, at which point A or B (your stop or your take profit) the market will hit first, and if you can still take a hit (a loss) or more than one consecutive hit and still have the capital to continue to trade the system, you will be profitable over time.

Still with me? Is that logic correct so far?

Knowing exactly which price the market will hit, either your stop or your take profit, all the time is difficult, yes.

However think of this scenario, I think you will agree the answers in these two very recent scenarios are "predictable".

The EURUSD closed on Friday at 1.3280

Where will the market go next? With that question you will have a thousand predictions and analysis.

Let's be more specific. Before we discuss predicting direction. Let's talk about price.

Which price will the market hit first: 132.40 OR 136.80 ?
Yes that last print is correct.

OK, obviously and statistically we would say that it will hit 132.40 before it hits 136.80.

We could be wrong, but we would be right more times by picking 132.40 in this scenario. Which is plus down 40 before it goes up to hit 400 pips up from here.

Without even talking about direction, by predicting that the market will hit 132.40 before 136.80 is saying that in the short to intermediate term based on this information, you would be short.

OK, what about the other way?

Friday's close was 132.80

Which do you think the market will hit first? 133.20 or 128.80 ?

Most would answer 133.20 which makes you a long position holder.

Both in this case would be statistically valid and have a positive expected outcome.

Those figures are arbitrary, but both are the opposite of the same ratio, based on recent historical EA testing. The SL to TP ratio can be anything to suit your style of trading.

Now, suppose you have a signal that gives you an accurate indication of direction over time (it doesn't matter how long it takes to go into your expected direction as long as it does before it goes to your expected stop out price).

Suppose your signal is generated 10 to 20 times a month with 80% accuracy. As long as you have wide enough stops and the capital to cover losses. You can expect a positive return value over X number of trades.

Finally with position pyramiding of profits, you could have a very respectable ROI.

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

And what is your image of? What are you trying to say here?

Actually, my SL TP was off on the long side. If the inclination is long, I should have said 133.20 or 128.80 which will be hit first? Which I see is close to the S/R lines on the chart you posted.

I'm guessing that by the thickness of your lines, you prefer the long trade side of the equation.

3 out of 4 of my EA's say the short side is in play now. But with Friday's massive buying, who knows.

Either way.... 133.20 / 128.8 or 132.4 / 136.8 either could work, and with the recent volatility, both profit levels should be hit before either stop level...probably within 2 or 3 trading days at most.

BTW, it just happens to be that the numbers my EA spit out just happen to be at the upper and lower extremes of a six month EURUSD daily chart. The EA doesn't use those types levels to indicate stop levels but now that you've shown it on a chart to visually illustrate my numbers, it would be another interesting thing to consider in creating an EA.

This particular line of thought on this thread has to do with "The probability of event A occurring before event B based on what could be construed as numerical distribution and statistical sampling." The method to derive an (accurate as possible) entry can be anything, but the important aspect is having as much statistical or reasonable certainty that your TP price will hit before your SL price. One way of determining that instead of momentum, support/resistance or trendlines is by simply having enough distance between your TP and your SL and having the equity to cover the difference for X number of consecutive losses. It's pretty elementary thinking but however in real market conditions, there are few traders who will trade with a stop 10 times their profit level. Easy to talk about but hard to do. Some would say "why not just trade without a stop?". However, trading without a stop requires monitoring the market 24/7 if you have serious capital on the line. It would require either price alerts that could wake you up in the middle of the night or a trading staff to monitor the markets and then a decision would have to be made in less than ideal conditions as to what to do with the position. I prefer the reasonably predictable trade outcome of knowing that my next trade is either going to be a winner of X dollars or a loser of X dollars and then manage the next position based on the outcome of and the capital levels of the the previously closed trade(s).

Manual traders (I used to be one for many many years) think in terms of best entry, best indicators, risk reward ratio, trendlines, support resistance, winners vs losers, etc, etc, etc,...I think most Manual traders are focused on looking for the next trade or managing the current trade. Many of the concepts of manual trading can be found online and in any Technical Analysis textbook. 20 years ago, I used to run the office of a TA software company that had wall street, chicago, and global trading superstars as our customers, some of them are huge names in the industry, so I've pretty much seen and heard it all.

Anyway after really developing EAs for the last couple of years, I found myself thinking about trading in terms of not the next trade and my next entry or big winner, but in terms of a self contained and self managed system, and in terms of my next several trades (the next X trades). In other words, I'm looking at the potential result over X number of trades or over the next several days, weeks, months or years. Of course I still think about ideal entry and exit signals and how to implement them, but the finding the PERFECT signal has taken 2nd place to finding an ideal position size, money management and risk reward.

By risk reward, I don't refer to the typical textbook idea of risk reward which is the amount of risk you are willing to put on a trade compared to the hoped for reward, such as novice and even experienced traders like to talk about like..."You must trade with at least a 3:1 reward to risk ratio". Some scaplers go for 1.5:1 risk reward, gaining 7.5 pips for every 5 risked on a trade. By risk reward, I look at the complete system over time. For example, how much money are you willing to risk for how much potential gain not on one trade but for the next X number of trades that your system will produce. How much are you willing to lose in total before you shut the system down? I look at risk reward like odds return at the track (ponies here). For example, I'm willing to risk $5000 on this EA in it's first real test run and I expect to gain $50,000 in X number of months (or years depending on your expected time frame). So that would be a 10x risk reward in X number of months or years.

It wouldn't matter to me if I risked $100 to gain $10, or $10 to gain 100 (on a single trade) as long as the overall profit outcome of X% return in a series of trades was expected. In the past, I wouldn't consider a trade unless it was at least 2.5:1 reward to risk. But now, I consider even a 1:10 reward to risk if the expected percentage of accuracy was high enough and the expected loss was manageable and still continued to trade the system with positive expectation even after a major loss. Most manual traders tend to break down and deviate from their system when such a break down occurs. Even system traders, even EA traders have a tendency to doubt their system or to change it after a major DD period or trade, although they have managed to expect a loss before hand. The key is to pre-manage the risk. Letting the EA do it and leaving it free to do it makes it technically easier to continue trading thesystem even after doubtful times.

By opening up my mind to accept higher risk and yet manage for losses at the same time, it has allowed me to develop more profitable EAs.

The key is to expect losses and to manage them in to the financial analysis over X number of trades. How many losses does your system historically produce? Can you manage 2x or 3x or 4x or 5x the amount of losses and still be profitable? When will you completely get out? This needs to be determined in advance. At what point do you quit because of losses? Or begin again fresh after a nice profit withdrawal?

The point of quitting because of losses has to be determined without change in ADVANCE. You must let your system play out to either your expected total loss determined in ADVANCE or your expected gain (withdrawal) in advance. This is the only way to know fully whether your system was a failure, partial success or complete success.

Yes, most of people believe that trading is gambling. Personally speaking I take this gambling as my profession.

With proper risk and money management it is not enough hard to make profitsss... And thanks to my money manager... Many time, I make a trade and close the lid of my laptop.. Rest my money manager works for me.. Only I set where I want take profit or stop loss after I find A/B or keep them trailing...
Seating on the screen changes my mind.. So I do only analysis, enter my orders and take position only rest money manager manages itself according to my command.

Without risk & money management trader will be thrown out of Forex market soon....

OK, so I've discussed a bit about Probability and Money Management.

What about pyramiding?

Pyramiding profits is the fastest way to accelerate gains with a system that produces consistent positive expectancy outcomes.

What is pyramiding?

In simple terms it's taking larger position sizes as your equity grows. Many traders trade consistent sizes. Such as 0.1 or 1.0 or 5.0 or 10.0 per trade. Consistent position sizes are easy to calculate and easy to manage and ideally suited for scalpers and those creating a daily income from trading.

The beauty of an EA or automated trading system is that it becomes very easy to mathematically scale your position size according to the amount of capital you have for EVERY trade and according to how much total risk in terms of percentage of available equity you have. Dynamic position sizing allows one to take full advantage of streaks and to scale down after losers and lower equity times. It allows one to manage risk proportionately to the immediately available equity and desired overall risk level.

Don't mistake this concept from a martingale system which doubles down after losers. Pyramiding or what I prefer to call "Dynamic Position Sizing" is not even close to a martingale system.

A martingale works like this....let's say you have a system that produced 10 trades...
W W W W L L W W W L (70%)
Let's say that you're at 100:1 margin and have $5000 to trade.
At one lot per trade, a marty will have traded (number of lots):
1 1 1 1 2 4 1 1 1 2
Depending on your profits, your system would have been over-leveraged and not able to trade on the 6th trade, in other words, just about out of business after 2 losers.

Of course in that example, you could say he or she should have at least started with $10,000 to trade a 1.0 lot. But the concept is the same...

The overall reward to risk of the system above (not the individual trades) is not worth the risk. Anyone trading a martingale especially in that case of a 70% hit ratio is simply setting oneself up for a huge loss to zero by doubling up on losers. Martingales have been proven time and time again to bust out the user whether it's forex or blackjack. It could be fun at a blackjack table but it just doesn't make sense in forex since you have the ability to manage your A and B points and the ability to scale position sizes to the penny.

Pyramiding is the controlled increase in size after a winning trade and the deliberate mathematical decrease in size after a losing trade. It increases the potential return of anything over 2 consecutive winners and decreases the overall risk of anything over 2 consecutive losers. So a system with a positive expectancy edge has an incremental increase in profits. Whereas, a martingale system does not have an increasing equity curve but a potential for exponential consecutive loss.

Of course, pyramid also increases net risk over time compared to consistent size trading, however under any positive expectancy system, it will generate a higher overall return, and with a system that produces consistent gains it will produce a significantly larger overall return than a consistent size system.

Sorry, but I gotta come in here to make some things clear.

A risk reward ratio by itself won't make any system profitable. Just because you use for example a 10 pip take profit versus a 100 pip stop loss, you would still lose in the long-term, if 91% of the time you win and 9% of the time you lose. => this corresponds to a win/loss ratio of 50% if you used a 1:1 risk/reward which makes you a loser if you count in the spread and comissions.

You need to think in terms of combined win-loss and risk-reward ratios. Only the right combination will make a system a winner. => Thus you still need an edge, e.g. for a 10 pip take profit and a 100 pip stop loss you need to be able to win above 91% of the time to be profitable, or with a 10 pip take profit and 10 pip stop loss you need to win above 50% of the time to be profitable.

Secondly, "pyramiding" refers to averaging in (scaling into a position the more it moves into your favor), what you mean with increasing lot-size per trade if your account increases is called "fixed fractional position sizing".

Thirdly, the amount of historical losers is totally irrelevant. Please google "long-term capital management": the losses those nobel-prize winners made were far greater than anything their historical studies suggested. That's the thing about the markets, they always change and will wipe everybody out in black swan events, if you don't see them coming.

Exactly. Couldn't have clarified my thinking out loud better myself. Wonderful reply Custos. Thanks.

Yes. One needs an "edge" to make any system profitable, and with dynamic position sizing (fixed fractional position sizing) one accelerates profits dramatically. I prefer to call "fixed fractional position sizing"...."dynamic position sizing" which moves up and down according to equity and is different for each trade in relation to available equity.

Lastly most importantly as you implied one needs to cash out periodically and re-evaluate potential risk vs. expected (desired) reward. One must compensate for black swan events, but even more sensible would be to compensate for consecutive black swan scenarios and accommodate for those events in their trading as much as possible.

The billion dollar question or answer would be how to maximize return while not only minimizing loss, but actually calculating it in. (Like a casino).

Been thinking a bit more about what Custos said...

"Thus you still need an edge, e.g. for a 10 pip take profit and a 100 pip stop loss you need to be able to win above 91% of the time to be profitable, or with a 10 pip take profit and 10 pip stop loss you need to win above 50% of the time to be profitable."

This is an easily verifiable fact statistically. However even with a highly accurate system, even with an edge, it's impossible to know in advance if your system will trade 90.5% accurately or better than 91.1% accurate over time. The nth trade which closes at 90.5% accuracy will put a wrench in your system. Even with the right combination of combined win-loss and risk-reward ratios one needs to know when to cash out, otherwise to continue would bring you one step closer to that devastating nth trade or nth trades.

Wasn't there a popular Country Western Song a few years ago that went like this....?

"Know when to hold em, know when to fold em...KNOW WHEN TO WALK AWAY, KNOW WHEN TO RUN. (emphasis added).

Well, you probably need a drawdown threshold: i.e. your system historically produced 15% drawdown. Allow for some deviation from that historical drawdown and stop trading when you are down 25%. Then the system most likely failed and hopefully until up to that point you were able to exploit the edge and make profits.

However, winning in the long-term on price action alone is an unnecessarily difficult endeavour. Incorporate fundamentals, some common sense and some market psychology and you are much more likely to stand a chance in the long run.

Hi Pipologist

Initially I thought, its an another attempt by someone who is unwilling to take the loss but I was wrong.

Interesting reading so far, subscribed.

In theory this would work. Yet, the key of all of this is figuring out which price to start from. You can break down the possibility of any one trade going in your direction. Yet, no one truly knows what will happen over the next few mins which will change the direction in a heartbeat.

Rather general conclusion to how to become successful over the long term. The problem is we want to become successful in the long term, without ever being successful in the short term.

Exactly my point.

Most traders are so intensely always looking for an EDGE (myself included). But even with an EDGE, such as fundamentals, some common sense and some market psychology as well as technical, ALONG with the inclusion of probability, money management and dynamic position sizing, one will have losing streaks. Long term profitable traders have factored in the losses before they trade and are capitalized enough to withstand the storms. Most unprofitable traders are simply over-leveraged or trade too frequently based on the amount of their capital.

Many traders especially newbies are also so focused on how to exploit the charting tools in every major software. Even those that trade on price action alone use some form of support and resistance levels and trendline projections, which are for the most part subjective depending on your time frame outlook.

You're a scalper so you're managing risk for the next X minutes of time in trade, so the reality that the market will move 1000 pips in either direction within the next 12 months is not managed into your trading plan. To figure out which price to start from, one looks at the longest term chart (long term monthly) and looks at the highest highs and lowest lows and figures out a standard deviation in pips. Of course, this is what scalpers do on a much short term scale.

The term OVER-TRADING is a rather selective term. Their is no rule which says one should trade X amount of times a day. Over-trading doesn't cause losing positions. Not understanding the dynamics of the market along with ignoring price action causes the losses we have. The reality is long term traders do factor their losses in before the trade is placed. Yet, for using small leverage with small equity. They become a slave to the position which seemed good just yesterday. Higher time frames shows you what has occurred, not what will occur. If it was that simple everyone would be millionaires with 1 year of trading.

So I'm letting the cat of the bag. I started this EA because I wanted to think about what could happen with an EA that I was experimenting with.

Without posting screen shots for those that are too lazy to read through this thread and think about the concepts, the results were as follows from starting May 28 of this year:

28 trades.
27 closed winners.
1 open trade (currently at loss, Friday close at 132.80).

Approximate open trade levels:
Trade open at: 1.3183
SL: 1.3618
TP: 1.3144

Max DD Percent: 17%
Max DD Projected on Single Loss: 43%
Max DD Projected on Two Consecutive Losses: 68%
Max DD Projected on Two Consecutive Losses: 83%

Starting capital: 1.00x
Balance so far: 2.63x
High watermark so far: 2.69x
Low watermark so far:

Of course the sample size is quite small at only 28 trades times 2 months, however it looks interesting.

The SL/TP levels are dynamic based on a percentage of price movement from entry price. They could just as well be static based on physical long term high/low plus or minus a buffer in X number of pips. The wide stop is deliberate. 1 loss over X number of profitable trades is factored in. Statistically, it would be difficult for the system to generate 3 losses in a row. In fact, if you think about how wide the stops are in percentage to the market 2 consecutive losses in the same direction within short time frame, would most likely be a major black swan. System would still be profitable with 2 consecutive losses.

The entry EDGE here is proprietary however, I'm thinking it could be just about anything, such as moving average cross, higher highs, lower lows, etc. This entry signal generates a trade or more within every couple of days.

So were not talking about astronomical returns here, although the risk is astronomical the probability of going to zero is pretty close to zero. But the reward is commensurate for the risk, IMO.

I'm not really sure how to respond so I'll just give some of my thoughts on pyramiding and risk reward.

The logic behind why pyramiding works,

If you have a target, and an entry in profit, why wouldn't you add a little more to the trade, if you know it is still very likely to head toward your target?

The risk is that if you are wrong, your break even is moved a little higher than original entries.




Often time pyramiding works best with breakouts. You catch the reversal with bigger entries, and then add a buy/sell stop for your pyramid trade on the breakout.


It's all about riskreward. I have realized the big players, they risk like 2-3 pips on a trade (not an investment, but a trade). Sometimes even just 1 pip.

If the swing high/low is 1.31225, they will put a buy limit 1.31230 and SL 1.31220 (yes they have true cost spreads of 0.1-0.5.. mine is 1.0-1.3)


THIS IS WHY you so often see price reverse within 1 pip of a swing high/low. It's the best risk reward and the big players take advantage of that... which is why the market reverses in the first place. Big players are taking advantage of a very high risk reward setup. If they weren't, the market wouldn't reverse.

I agree with you and disagree with you.

"If it was that simple everyone would be millionaires..."
Spot on.

"Higher time frames shows you what has occurred, not what will occur."
Do lower time frames show you what will occur? You must have an amazing charting system.

"There is no rule which says one should trade X amount of times a day."
Never said there was. I said "Most unprofitable traders are simply over-leveraged or trade too frequently based on the amount of their capital." Over trading in relation to capital costs money in terms of either commissions or spread. Being over leveraged at the same time is the kiss of death to a trader.

I see you have a nice return this month on your live acct. It's commendable. But I hate to rain on your parade. You are simply over-leveraged. Trading 2.0 lots while being maxed out on a 500:1 leverage acct. leaves you zero wiggle room. I'm taking a guess here but it also looks like you are trading manually. In any case, your leverage is not sustainable. One mis-cue and you're psychologically if not risk capital depleted.

So you may have factored Probability and even "pyramiding" (in the loose definition of the word), but your money management will devastate your system sooner or later (my guess is sooner).