Risk & Drawdown Analysis

You will notice a new column in the trade-by-trade analysis: Risk and Drawdown. Here, C2 looks at how far each trade moved adversely from the opening trade price (not the high-water mark) and compares this number to your eventual profit. It then draws a little graphical representation of how “risky” your trade was. The idea is to alert users if you are willing to let losses run into the hundreds of thousands of dollars in order to eke out an eventual $13 dollar profit.

Now, before all you system vendors have a conniption, let me alert you that I am in the process of tweaking the way the final assessment is determined (i.e. whether a trade “extremely” risky or just “modestly” risky.) Currently, the algorithm looks at the percentage of the adverse move versus the eventual profit, but makes exceptions for long option positions. (Since a long vega strategy will typically see drawdowns at 100%, with a few outlying but very profitable exceptions.)

I welcome feedback and suggestions for how to improve this new feature. As with the Realism Factor, it’s an algorithm undergoing rapid changes and improvements, so it’s not available as a sorting or searching criterion just yet.


I really appreciate your effort to provide us more helpful information. It looks very useful to me with a few exceptions of the data accuracy. I was very exciting about this new feature and went check the systems I subscribed (Tango). Right away on the first one, I found doubtful data base on whatever your data feed gets. The rating of ADSK is based on lowest price $27.72 occured at 2/2/06 9:42. I was very surprised since I have followed this one closely. Checking with my broker and Yahoo, they both show the lowest price sice 2/1/06 for ADSK was $35.17 occured at 2/3/06 2:15. Just like all the C2 “Best Fill” based on, there are quite a few unreal/ghost prices/fills I have trouble to consider them are all real.

Fantastic Matthew, the new feature is even better than I had hoped for (which was only the max drawdown $ amt). Great job!

Thanks. This is a good tool to analyze open positions.

For eg., I can reduce straight underlying positions (which is equivalent to raising the stop loss on the position), when I am wrong about the direction and may even reverse position if confirmed in the opposite direction, when the risk level indicator goes to high/extreme so as to reduce it to modest; conversely, I can add to positions when I am still correct about the direction when the risk level indicator goes to high/extreme, so as to reduce it to modest.

Also, I can increase positions when I am correct about the direction and the risk level is low/too low so as to bring it back to modest. Similarly for options, I can buy puts/calls in combinations to manage risk.

I would also like the indicator to go to “too low” level so as to balance out the whole range, so that the “modest” level is a true average level for the risk encountered by the trade.

But, it does not help much in analyzing closed positions, and in fact gives a misleading representation of the drawdown & risk, because the actual or true or real-ized drawdown is the “real” risk which are different from the unreal-ized drawdown which represents “unreal or imaginary” risk. That is why this is a good tool to analyze open positions, but not closed positions, because the “drawdown” for the closed positions is the “real-ized” drawdown not “unrealized” drawdown.

Any further analysis done to arrive at concepts based on this “imaginary” risk is “invalid”, because it cannot be reduced ultimately to actual existence of it, because it no longer exists. This is the proof that such concepts are invalid. This is called the fallacy of the "stolen concept.“

The fallacy consists in using a higher-level concept while denying or ignoring its hierachical roots, i.e., one or more of the earlier concepts/percepts on which it logically depends. This is the intellectual equivalent of standing on the 40th floor of a skyscraper while dynamiting the first 39 including the foundation. The higher level concept - “drawdown,” in the above example - is termed “stolen,” because the individual involved has no logical right to use it, because it does not exist anymore. He would be an epistemological dependent, who seizes, without understanding, a term created by other men who did observe the necessary hierarchical structure. The parallel to a dependent in matter, who seizes wealth created by others, is obvious.

The reason stolen concepts are so prevalent is that most of us have no idea of the “roots” of a concept. In practice, they treat every concept as a primary, i.e., as a first-level abstraction; thus they tear the concept from any place in a hierarchy and thereby detach it from reality. Thereafter, its use is governed by caprice or unthinking habit, with no objective guidelines for the mind to follow. The result is confusion, contradiction, and the conversion of language into verbiage.

The antidote is the process of reduction. In regard to higher-level concepts, reduction completes the job of definition. The purpose of a definition is to keep a concept connected to a specific group of concretes. The definition of a higher-level concept, however, counts on the relevant lower-level concepts, which must themselves be connected to concretes; otherwise, the definition is useless. Reduction is what takes a person from the initial definition through the definition of the next lower level and then of the next, until he reaches the direct perception of reality. This is the only means by which the initial definition can be made fully clear.

“Invalid concepts,” are words that represent attemts to integrate errors, contradictions, or false propositions, such as concepts originating in mysticism [eg., “ghost,”, “gremlin”] - or words without specific definitions, without referents, which can mean anything to anyone, such as modern “anti-concepts” [these are deliberately equivocal terms, such as “extremism,” “McCarthyism,” “isolationism”.” Any such term is detached from reality and invalidates every proposition or process of thought in which it is used as a cognitive assertion.

The test of an invalid concept is the fact that it cannot be reduced ultimately to the perceptual level anymore, because it does not exist anymore. This means that nothing in reality gives rise to the concept anymore. Any propositions based on such “invalid” concepts are also invalid, as concepts are the elements of propositions. Propositions too (if nonaxiomatic) must be brought back step by step to the perceptual level.

Logic requires a recognition of context and hierarchy. Logic is “the art of noncontradictory identification” - while observing the full context of knowledge, including its hierarchical structure. A logical conclusion is one which has been related without contradiction to the rest of a man’s conclusions (the task of integration) - and which has been related step by step to existing perceptual data (the task of reduction). Between the two processes, man achieves a double check on his accuracy. Every conclusion must stand the test of his other knowledge and (through the necessary intermediate chain) the test of direct existence.

It is futile to uphold true ideas while ignoring hierarchy as to uphold false ones. Contrary to today’s conservatives, for eg., it is not an axiom that man has the right to property. The right to property is a consequence of man’s right to life; which right we can establish only if we know the nature and value of man’s life; which conclusion presupposes, among other things, that objective value-judgments are possible; which presupposes that objective knowledge is possible; which depends on a certain relationship between man’s mind and reality, i.e., between consciousness and existence. If a thinker does not know and count on this kind of hierarchical structure, he can neither defend property rights nor define the concept nor apply it correctly. This is one of the reasons why today’s conservatives are ineffectual.

One should check their premises often. If one proposes to enter the field of concepts and propositions, they should check the premises - see what they depend on, all the way back to the base of the structure.

If one’s reduction is accurate, one will find that the base is the axiom: existence exists: rationality (commitment to reality) and its obverse, only existence exists: honesty (rejection of unreality).

I strongly disagree. Unrealized drawdown is very real as you will find if your broker ever gives a margin call on this “imaginary” concept.

Totally agree Pete. Take a look at vantage.com -a consistantly profitable forex provider. It makes pips every month almost without fail. However what you don’t realise is that the ‘cost’ of those pips can be a 2500 open pip trade drawdown meaning that you have to have at least $50K in your account to produce perhaps $5K a month -soemtimes more. You have to ask could this $50K be put to better use elswhere. For example knovamind.com average about 100 ++ pips per week BUT rarely allow a postion to move more than 10 pips away from their entry. SO I would require just $5K approx. to fund the latter account and achieve the same return as Vantage -ie a 10 X multiple difference!! If I used all the funds required to trade Vantage my possible potential return with knovamind could be multipied by a factor of 10 or $50K!!!

“Unreal-ized” drawdown becomes “real-ized” when the position is exited from an adverse move completely at any time whether a margin call from the broker is issued or not. Therefore to infer of a position’s drawdown from the fact that there are margin calls issued by the brokerage industry in every country, would not be a proper proof. The margin call from the brokerage is a consequence of our knowledge of a position’s drawdown, not a precondition of such a knowledge.

Concepts or propositions are based on antecedent cognitions - on the chain of evidence that led to them - going back ultimately to direct existence. To a mind that does not grasp this chain, a higher-level proposition is arbitrary, noncontextual, nonobjective; it is detached from reality and from the requirements of human cognition. This is why proof of an idea is necessary.

Proof is a form of reduction. The conclusion to be proved is a higher-level cognition, whose link to reality lies in the premises; these in turn eventually lead back to the perceptual level. Proof is thus a form of retracing the hierarchical steps of the learning process. (As with conceptual reduction, so with proof: the process identifies not the optional variants, but the essential links in the chain, the necessary logical structure relating a mental content to observational data.)

Proof is not a process of deriving a conclusion from arbitrary premises or even from arbitrarily selected true premesis. Proof is the process of establishing a conclusion by identifying the proper hierarchy of premisis. In proving a conclusion, one traces backward the order of logical dependence, terminating with the perceptually existing. It is only because of this requirement that logic is the means of validating a conclusion objectively.

For eg., if someone were to infer a given man’s mortality from the fact that there is a huge funeral industry in every country, this would not be a proper proof. The funeral industry is a consequence of our knowledge of human mortality, not a precondition of such a knowledge. The standard Socrates syllogism, by contrast, does validate its conclusion. It derives Socrates’ mortality from a truly antecedent generalization (which in turn integrates countless observations of men and of other living organisms).

Let me explain the principle of "Rand’s Razor."

A “razor” is a principle that slashes off a whole category of false and/or useless ideas. Rand’s Razor is addressed to anybody who who enters the field of philosophy. It states: name your primaries. Identify your starting points, including the concepts you take to be irreducible, and then establish that these are objective axioms. Do not begin with some derivative concept or issue, while ignoring its roots, however much such issues interests you. Philosophical knowledge, too, is hierarchical.

Today’s philosophers not only evade this point, but reverse it, just as they reverse the principle that knowledge is contextual. In regard to context-keeping, they not only fail to integrate their theories; they crusade for nonintegration, insisting that every question they study is independent of the others, that philosophy consists of “piecemeal analyses,” and that the cardinal sin is system-building. As an expression of this anti-contextual mentality, these same thinkers not only neglect the task of reduction; they brazenly invert the hierachical order of knowledge.

The rejection of hierarchy on so profound a scale amounts to the rejection of reason as such. It represents the attempt to enthrone naked whim as the ruler of cognition - which is not a mere error, but a form of willful irrationalism.

Pal: I beg of you – please stop long-winded philosophizing about matters that are only tangentially related (at best) to trading. You make the forums unreadable.

Also, no one in the world agrees with your position that open trade drawdown should not be counted until trades are closed. It’s a point you have made again and again, to no avail, and I can tell you flatly that my position on this matter will not change… at least until the day that the CFTC and the SEC reverse 100 years of precedent and announce that funds and CTAs don’t have to report open drawdown numbers. Until that day, please stop harping on it.


C2 is neither a Fund nor a CTA, so those SEC and CFTC rules that apply to them do not apply to C2. If one is operating a fund or acting as a CTA, then they can report the “unreal-ized” drawdowns to comply with those rules.

How is it possible to know such sophisticated facts, yet not know the difference between “unreal-ized” and “real-ized” drawdowns? Mr. Matthew Klein feels no need to raise such a question. He feels free to begin philosophising at random, treating advanced knowledge as a primary and using it to undercut the direct evidence of men’s eyes.

He does not merely use advanced knowledge while ignoring its roots; he uses the knowledge to destroy its own roots. And he does not merely misappropriate in this fashion a single term, but a complex body of conclusions. He is guilty not merely of one stolen concept, but of conceptual grand larceny. This is the kind of anti-hierarchical corruption that makes philsophies such as skepticism and hence compromises possible. It is this kind of philosophy that Rand’s Razor slashes off at the root.

ps: This will be my last post on this discussion.

Hi Matthew,

The new feature is interesting, but not totally relevant.

The new analysis makes no difference if one trade is very small or very big, the only thing that matters is the outcome of the trade.

Forex market being extremely volatile, almost every trade IS extremely risky, by nature. What makes a system more or less risky is not the pips at risk : money management is the proper tool in forex risk-management.

I find the maximal $ drawdown analysis far more relevant and far less complicated.

Back to the new feature, you must find a way to take size into account, because the risk and the drawdowns, measured in pips, simply means nothing, you must mesure it in $.

The other very negative point of the new feature : it erased other informations on the left of the page and I can’t see it anymore(?). Where have the $P/L, RF and commentary gone?

Anyway, thanks for the hardwork and constant improvement of the site.



I agree with your position on open drawdowns, but I think the longer term trading systems are getting treated unfairly with regard to equity graphs over daytrading systems.

Since the equity graphs are updated in the evenings, this result in closed trade equity graphs for daytrading systems, not showing any of their open drawdowns, but for long term systems, the open drawdowns are clearly displayed.

I think the addition of equity graphs based on weekly and monthly equity levels, might be a nice addition. Then, for daytrading systems, the default chart can be the daily equity graph, for systems with average holding times between 1 day and 2 weeks, the default can be the weekly graph, and for systems with average holding time over 2 weeks, the default can be the monthly graph and the viewer can have an option to switch the graph to lower time frames.


- Fanus


If you looked that the recent discussions about Hawk FX, this is very clear that unrealized drawdowns are very real to subscribers.


- Fanus

PS: After your postings as Kavin and now the changed name from Kavin to Futureware, I think you have lost any credibility you might have had.

Interesting discussion.

Unrealized drawdowns should be monitored by C2.

I feel in part, the issue arises because C2, and system clients, place undue emphasis on win% of trades. The systems with higher win% often get that way by riding out losers until they return to the profit column and then generating a profitable trade as an end result.

There are plenty of systems that have more winning trades than losers which lose money overall. And there are systems that have more losing trades than winners that are bottom-line profitable.

System consumers might be well served to look for systems with lower win% of trades and steady, upward-sloping equity curves, and they then might discover systems that methodically cut losses while thoses losses are still small.

It is not natural to consider a system ‘good’ where most trades are losers, but one quickly gains comfort knowing that the system operator is carefully managing monetary risks.

Couldn’t agree more, Christopher.

A stat that you don’t often see published is actual TIME in winning trades .vs. losing trades. Even tho a system may have only 35-40% wins, it probably will spend most of it’s time in winning trades, as losers will tend to be closed out quicker. Stats from development platforms like TradeStation and TradersStudio show average bars in winning and losing trades, and typically, for a position trading system, you’ll see many more bars in the winners.

Matthew - I have to agree with Jerome. You need to consider the big picture when assessing risk. This includes total account drawdown at the time the trade was entered, margin required, market volatility, diversification, etc.

I am a bit baffled why I have a closed stock trade with $620 maximum drawdown within a $110K account that was labelled extreme risk. I was not trading on margin.

The equity curves can also be misleading because of slippage, tight limit orders that often can’t be hit in a real account, etc. Two examples of nearly perfect equity curves on C2 are Extreme-os and Coin Collector 40 min, but try to find anyone trading either of these systems that has a positive slope in their real account for any length of time exceeding one month. I’ve fell victim to these types of equity curves in the past because that is exactly what I was looking for in a system. But I learned the hard way that these curves have very little meaning other than showing the system in a good light for advertising purposes.

The new trade-by-trade drawdown numbers are very useful I think because they do give an idea of the risk a system will take, whether stop orders are being used and how wide they may be, etc. I agree that the end result is all that counts and pay little attention to the W/L numbers myself, but also agree that some vendors will hold trades far too long in order to have it show a profit in the end (or a disaster, whichever comes first) so their equity curves and P/L tables look better. This seems to be much more common with systems that already have unusually high W/L numbers than for systems with more normal numbers like 2:1 or so.

As a subscriber to some of these systems, unrealized drawdown to me is VERY important. I’m getting irritated at (some) of these system providers contending it’s irrelevant. Yes, it’s irrelevant to those system providers who don’t put their money where their mouth is. It’s extremely easy to claim drawdown doesn’t matter when you’re not sweating bullets that your real account is getting screwed up. You can call me a chicken for stopping out on my own on those trades in the middle or tail end of a drawdown, but this is real money on the line; not play money like some of these system users fiddle with. Sure, if I made a system that had a trade drawdown of 50% of an account, I wouldn’t be worried in the least, because I have no real stake in it. I’m extremely happy that MK has now posted these "unrealized’ drawdown rigures, because they are very real to the subscribers. When you guys start actually putting 50K of your own money on the line with your systems with huge unrealized drawdowns, then post on here that it doesn’t matter! Now I’m sure some of you will say you do, but I seriously doubt it. Anyway, Paul Anad, you may think you’re some sort of philosopher or something; but I’m sick of your long winded posts that have zero to do with trading, and I imagine everyone else is too!

Randy: C2 is attempting to address some of the issues that you bring up. In the case of the two systems that you mention, go to their pages and click on the button below the equity graph labeled “Realism Factor, with commis.”. The graph, altho still ‘smooth’, isn’t quite as attractive anymore. One of them is dead flat - which means best case scenario it’s just treading water.

Perhaps if you expand your time horizon a touch and look for trades that are a little longer in duration. Position trading systems are not nearly as sensitive to slippage or timely order entry as those that try to scalp a few points here or there, plus they usually trade in smaller size which helps one sleep better at night.

Hans … I do have most of my money in much “safer” investments (real estate and diversified mutual funds) and these have a long time horizon of many years. But I have always enjoyed trying to trade the markets on much faster time scales and have not been very successful with my own ideas, although I never lost more than 10% of my speculating account in a given year which is much better than my system picking results on C2 so far!

So I am only looking for relatively fast systems (1-2 trades a day, or so) when shopping on C2 and not any sort of serious investment approach. I believe the people who can beat the S&P each year make their money for a good reason and its not easy to do. Despite the nice equity curves shown for many of the faster-trading C2 systems, I’ve tried about a dozen so far during the past 5-6 months and have yet to find a system that did not lose money at a faster rate than I could lose it myself. Reading comments from subscribers (which are far too few IMO) to these systems and others I’m not sure anyone else has found a truly profitable short-term system either, and maybe the conclusion is that they don’t exist. I did make a few dollars on one option trading system (Bris), but that was too wild for me and also not really a day trading system.

Hi Mathew,

Now that I see this feature applied to my system, I can give you some feedback.

I think to measure intraday drawdown per trade as % of total portfolio would be more useful. To label one of my trade’s drawdown risk as ‘extreme’ when as a % of total portfolio the dent was 1/2%. That doesn’t make sense.

There are also bugs - the @GCJ6 contract lists the worst price of the trade as 555 - I’m short from 562 - It’s not even a loss! But the software is showing a significant drawdown. I think the system can’t shift between the pit GC and the night @GC contract. I’ll reiterate my request for you to add the CBOT’s ZG electronic gold contract. The Comex can go to hell! :slight_smile:

Finally, many of my losing trades that were stopped out at their max drawdown show ‘low’ risk and ‘0’ as the drawdown amt. Another bug.

In any event - my hat’s off to you for tackling all these issues.