Chris,
Those are excellent points. 10 subscribers at $10,000 would not quite be the same as 1 $100,000 account, because of large block trades, but I understand what you are saying. This has not been an issue so far, but if I did get a larger number of subscribers, I would need to monitor the actual subscriber slippages.
I used to trade my own money with this system, and had good results. I now use a different system for my own trading, so I opened up this system to subscribers.
Brian
Brian
Thanks for the answer.
Chris
Julles,
I agree with your analysis and conclusion that technical analysis will do no good to evaluate systems. I had quite a few people telling me how much they liked my trading style but they are waiting for the first chance of a “dip” to take place before they join, as they feel “comfortable” that way. I just tell them best of luck in your “wait and see” approach, because you can never be able to measure the depth of the dip, let alone accurately time its turning point based on what happened in the past. It’s just impossible. That’s why if I were to join any system vendor (btw, I have no subscription to any system on C2), the first and most important measure of all is not the risk metrics that C2 uses, but how robust and experienced the trader is as in handling his losses before handling his winners because at the end, traders experience, longevity, and a true cutting edge, are the ultimate factors of how their system will perform now and the future. That’s my bottom up approach on how I differentiate between the deceptively good-looking systems and the truly awesome ones.
I’m not sure that I can agree with that 100%.
The thing is, although each trade is an independant event as far as probability theory goes, the system SHOULD be designed to provide a positive expectation on each and every trade, it should not be a 50/50 probability, even if the real-time results show 50% wins. That is if the system is well designed based on sound positive-expectation logic. If it is, it is not then a random game of chance, and those probability theories dont apply. “Reversion to the mean” is just a way of expressing that the results will over time average to the normal positive (or negative) expectation.
Also, in real time use, getting into a proven system during a drawdown is no different than averaging down into a sliding stock position to get a better price - it is a buying opportunity, rather than buying at the top!
But again, this discussion is academic because a) the C2 stats are not reliable enough (for me, anyway) and b) there arent many systems here with enough of a track record to be able to have confidence in a “mean” equity slope - although I think some like Athena and Brian’s BTS BottleRockets are good example candidates to keep an eye on. And I do agree Jules that a different set of rules applies for systems who hold positions overnight vs those that close EOD and therefore have a “daily” equity curve to use as a barometer.
W
Tarek:
I agree with your principle, but how do you know how robust and experienced a trader is? I try to determine this from what he writes at the systems page, but I cannot verify whether his claims there are true or just cheap ads. The advantage of the C2 statistics is that the trader cannot mislead you with them. The disadvantage is that you can mislead yourself if you misinterpret them, but at least that’s something under your own control.
Walter:
You wrote
" the system SHOULD be designed to provide a positive expectation on each and every trade, it should not be a 50/50 probability, even if the real-time results show 50% wins. That is if the system is well designed based on sound positive-expectation logic. If it is, it is not then a random game of chance, and those probability theories dont apply.“
Note that I assumed a win probability of 2/3 = 67% in my example. It could have been any other probability larger than 50% as well. Then the laws of probability still apply, contrary to what you suggest.
My assumptions were somewhat simplistic, but they were quite favorable for the system. In reality, I think the situation is even worse for most systems, because the market is not stable (like a slot machine) but is changing its patterns as other parties adopt different trade strategies, and because of unpredictable external events like war. Nevertheless, I think my assumptions form a nice and simple model to analyze some elementary pitfalls.
You wrote
” getting into a proven system during a drawdown is no different than averaging down into a sliding stock position to get a better price"
I disagree, it is totally different. The system is not a fund in which you buy a position. You just start following the system trades, but since you step in after the DD, these trades will ususally not anymore include the positions that caused the DD. This strategy is no different than expecting a 6 because the dice had so many 1s in a row, or expecting black because it has been red five times. The win probabilities are more favorable if it is a good system, but that has nothing to do with the moment that you step in.
Jules
Julles,
As I explained in my earlier post, you determine that by looking at how the trader handles him/her self when things are not going well. If the trader is able to make new highs in the equity graph WITHOUT taking more risks than he usually does when things are going well, then that trader has a unique edge that’s allowing him to succeed through good and most importantly bad periods. Secondly, The ability and speed of making new highs highlights that edge more clearly. Take my system for example, in '04, the system experienced two large dips yet every dip was met by an upward thrust, however it was failing to breakout for nearly 1.5 years. Why is that?? Well, because the system was missing critical components which were holding it back from catapulting higher. If you notice the results in my trade log, I take similar if not lower risks now than I did in '04 and the results are a whole lot better with continuous new highs, unlike the period of '04.
In addition, the discovery of an edge coupled with experiencing a prolonged bad period taught me about my self and allowed me to gain unprecedented experience of how I must engage the market. All of that is having a profound effect on the bottom line. In addition, a trader MUST adhere to risk parameters, part of a sound and thoroughly tested approach to insure long term success rather than experiencing “one day wonder” type of results. All of those are critical in judging the character of the trader as well as his/her ability to be a true survivor in this high stakes game we call trading (which in fact is just another form of gambling). I myself am not a gambler and never set foot in a casino, because I know that I have zero edge in that game. Sound calculated science is the driving force behind my trading methodology and that’s the reason I think why I’m still around after 2 years without restarting a new system. I think that is an accomplishment by it self.
In regards to deceptive ads to attract subscribers, that is a dishonest and cheap path to say the least. The trader must fully disclose his trading style and how he approaches the market with clearly stated defined risks that he takes. Any thing less is not acceptable.
.
OOPS, I think I must correct myself in my claim that the probability that the system will die is 1. I mixed up two things when I concluded this. So I withdraw this. I do not say that it is false, but I’ve to look at it closer. I remain with my other conclusions though.
Thanks for explaining Tarek. I think we have the same view upon trading.
About the ads: It need not to be dishonest, I was also thinking about an overconfident but inexperienced trader, who really believes what he says. But indeed it can be an evil person too, you cannot judge that from internet.
Thanks for explaining,
Jules
Forgive me for mispelling your name Jules.
Also, in real time use, getting into a proven system during a drawdown is no different than averaging down into a sliding stock position to get a better price - it is a buying opportunity, rather than buying at the top!
There are two sure ways to go broke trading a market: picking tops and bottoms based on valuations, and taking trades based solely on indicator readings. When did the Japanese stock market become overvalued when the PE ratio was 30, 40, 50, 60, 70, or 80 : 1? When did silver become over-valued? The previous high before 1978 was $6.40, yet subsequently it rose to over $50 per ounce.
Any trader can find an indicator to agree with an opinion or analysis, but it must be validated by price action. A market is never too expensive to buy despite the over bought indicator, but may be too inexpensive to sell based on intrinsic valuation.
>And I do agree Jules that a different set of rules applies for systems who hold positions overnight vs those that close EOD and therefore have a “daily” equity curve to use as a barometer.
That is precisely the reason why a “daily” equity cure should not be used as a barometer, rather, a “closed” equity curve should be used. The “closed” equity curve shows the same and accurate longer-term picture of any system. That is more important than showing the daily fluctuations. The daily fluctuations in the equity curve can be minimized by adjusting your position size. As the trader becomes more and more confident in his trades, it is natural to want to increase the position size to compete for higher profits. Warren Buffet, undoubtedly one of the 20th century’s best investors, says that smarts and talent are like a motor’s horsepower, but that the motor’s output depends on rationality. “A lot of people start out with a 400-horsepower motor but only get 100 horsepower of output; It’s way better to have a 200-horsepower motor and get it all into output” he said.
What the system vendor may give as a “stop” to control risk, is his opinion. However, that is an opinion, and you certainly do not ever want to trade from opinion: not mine and not even your own. The truth is no one can tell you where to place your stop. Your stop must be placed where it is mentally, emotionally, and financially comfortable for you. Only you have that kind of information.
Knowledge is not like a village of squat bungalows, with every room huddling down against the earth’s surface. Rather, it is like a city of towering skyscrapers, with the uppermost story of each building resting on the lower ones, and they on the still lower, until one reaches the foundation, where the builder started. The foundation supports the whole structure by virtue of being in contact with solid ground.
If every concept and conclusion were accessible directly from observation, then knowledge would involve no principle of order; it would be an accumulation of primaries. In fact, however, cognitive items differ in a crucial respect: in their distance from the preceptual level. Certain items can be learned from the simple sense experience. Others are not so easily available; they can be known only through a chain of increasingly complex cognitions, each level making possible the next one.
Knowledge, therefore, has a hierarchical structure. A hierarchy of knowledge means a body of concepts and conclusions ranked in order of logical dependence, one upon another, according to each item’s distance from the base of the structure. The base is the perceptual data with which cognition begins. If your perceptual data is inaccurate, so would all of your concepts based on it. An architecture is only as good as the foundation (framework) on which it is based on. If the framework is “flawed”, any architecture built on it merely an illusion, waiting to collapse on an accident. Proper human action can reduce the power of accident enormously. But this does not mean that accident can be eliminated.
Virtue is not automatically rewarded, but this does not chage the fact that it is rewarded, ultimately. Virtue minimizes the risks inhrerent in life and maximizes the chance of success. Concepts teaches one how to choose a system, how to organize systems so that the best among them rise to the top. It teaches one how to safegaured one’s interests in principle and therefore against every danger that can be foreseen. This does not give one omnipotentnce; what it given them is the means of preventing, mitigating, or counteracting innumerable evils that would otherwise be intractable.
There is no cosmic overseer, who takes note of virtue and crowns it with success. Nor is this an injustice on reality’s part; it is an expression of causality and identity - of causality, in that certain causes lead to certain effects, whether one desires them or not; of identity, in that man, like every other existent, is of limited knowledge. So success is not automatically guaranteed to the virtuous. Nothing can alter the given fact that one is not omniscient or omnipotent. Regardless of a person’s virtue, he may fail in an undertaking through simple error. The pilot “wrong-way Corrigan,” let us say, was conscientious and honest, but these qualities did not automatically point his plane in the right direction. If one does not acquire the necessary knowledge, then one cannot avoid suffering the consequences, even if one is no way morally deficient.
Besides errors of knowledge, one must also reckon with the factor of other men. If one’s goal in an undertaking involves the cooperation of others, his own virtue (or knowledge) cannot ensure success. The ideas, the motivation, the skills, character traits that he needs in others depend on their choices, not his. An individual in a free society is free to search the kind of men he wants or to try to persuade others to share his ideas. But no act of persuation, however skillful, can nullify human choices. One cannot change a man’s mind without his consent.
Hi,
averaging down on drawdowns or starting trading during a drawdown are common and useful techniques for systems with a good equity curve and a history of dips. I have successfully used it for some time and I know of at least two good guys (traders and system developers) that do the same: breakoutfutures.com and eminimethods.com. The former publishes a newsletter from which you can learn more about his ideas on this matter.
It’s important to know your system well, 6 months of equity curve is not enough for that.
Just 2 my cents.
Rob,
Averaging down on drawdowns or starting trading during a drawdown might be a good strategy if you apply it to the price of securities, but it makes no sense if you apply it to the return curve of a trading system. You won’t own a piece of the system when you start trading during a drawdawn, and you won’t be able to sell it later. I suppose the ‘good guys’ you are referring to apply it to things with some intrinsic value, like stocks or commodities, not to other trading systems.
Jules
Jules,
I don’t see what “intrinsic value” has to do with it. It is all about mean reversion. If the system has a positive expectancy with a proven equity curve, I’m not sure why it isn’t just common sense to buy the dips. The important (and difficult) task is to find systems with enough track record and reliable reporting to have a usable equity curve to game in this fashion. I agree 6 months may not be enough. But it depends on the system and how often it trades. You can postulate and theorize probability theory all you want, but in real life there ARE people doing just that - and I have to say it makes sense.
W
Here is a quote from Dr. Michael R. Bryant which might be of interest:
You would typically stop trading when the equity curve crosses below the moving average if your system or method tends to produce streaks of wins and losses, so that when it starts to lose, it’s best to stop trading until it starts winning again.
On the other hand, if your system or method tends to “revert to the mean” – after several wins, it starts to lose, and vice-versa – you would typically stop trading after the equity crosses above the moving average. Dependency analysis can be used to determine if your system has either of these tendencies with statistical significance.
Quote from article at: http://www.adaptrade.com/Articles/article-eq.htm
Walter,
What I meant with ‘intrinsic value’ is that when you buy stocks, you buy a piece of company that has some real value that may be underestimated temporarily. This underestimation will probably be corrected after some time, and then the price rises. When you bought this stock in the dip, you will make profit.
If on the other hand the equity curve of a system has a dip, this is because the system has earlier recommended to buy e.g. stocks and the price of these stocks decreased. Now, when you enter the system at this moment, you will not buy these stocks since the signal has already be given earlier. Instead, you will follow the new recommendations of the system, but there is no reason why these should be better or worse than the average of the system.
There are only two exceptions. One exception is if you buy the stocks that caused the losses, because they’re now cheap. But that is something different than following the system since you buy a long time after the buying signal has been given by the system. The other exception is if the system vendor deliberately changes his strategy in an attempt to compensate the losses, or if he deliberately adds to the loosing position (as you seem to recommend), or both. This seems to be what happened in Hawk-fx. But a system vendor who suddenly changes his strategy in the view of losses, that sounds like he is taking undesirable risks. Otherwise, if his new strategy is so much better, why didn’t he apply it earlier? Apart from the fact that many people are not happy with that, it depends on the system or the vendor whether he has this strategy or change of strategy. You can’t see that from the equity curve.
That many people do this, is not a sound argument. Many people are loosing, too. And many people pray, but that alone doesn’t prove that praying helps. The question is how successful the people are who apply this strategy. You may be right that they are successful when they apply it to stocks (I did not oppose that) but my point is that it doesn’t make sense if you apply it to systems.
You can sneer at probability theory and statistics, but this is where the term “regression to the mean” comes from. If I remember correctly, the term “regression” was used by Sir Francis Galton, who discovered that children of intelligent parents were on the average less intelligent than their parents. Hence the fear that the whole human population would “regress” in intelligence. However, as was later analyzed, this is just a statistical artefact - as I tried to explain above with the dices. Hence the term “regression to the mean”. There is no mystical force toward the mean, however. It just means that when things go extremely bad, they will probably go somewhat better (closer to average) the next time. It doesn’t mean that they will be better than average the next time.
Jules
Well,
I guess we can agree to disagree. As Jim Cramer famously says, stock is nothing but paper - it doesn’t really mean you own ANYTHING. What you get when you buy it long is the POSITIVE EXPECTATION of either appreciation of the perceived value or dividend income. Similarly, if you turn “on” a system with a proven track record during a
"dip", you own nothing but are hoping for a positive expectation, and the system’s logic and historical proof support (ostensibly) that expectation.
I don’t sneer at probability theory - i have lost more money than I care to remember in Las Vegas. I’m just saying that probability theory relates only to random chance, and by design the INTENT of these trading systems is not to bet on random events, but rather attempts to foretell events (prices) that are shaped by their relationship to other events (volume, past prices, news, etc). So to apply game theory/random chance is just not applicable.
W
What I meant with ‘intrinsic value’ is that when you buy stocks, you buy a piece of company that has some real value that may be underestimated temporarily. This underestimation will probably be corrected after some time, and then the price rises..
Exactly the premise of the C2 system Extreme-os, although their time frame is about one day. Not sure I can intelligently relate this to the present discussion, but that system shows a beautiful equity curve on C2 over a relatively long time period and can’t be traded in a real account to show anything like the same performance (based on my own experience and other subsrciber posts here). It has no dips to fade so Walter’s scheme could not be tested.
But I agree with Walter that if there was a system showing a long-term positive slope on the equity curve … long enough and with enough trades to conclude, statistically, that there is a “good” chance the slope value will remain roughly the same over time, then “buying the dips” might be a sound strategy. The system itself has intrinsic value in that it has a long term track record of maintaining a positive equity curve slope, and so could be expected to maintain that slope going forward.
Or do I suddenly sound like Pal!
Walter,
I agree, we disagree. I would agree with your quote about the value of stocks, though…
Sorry that I was a little oversensitive with respect to probability theory.
A large part of statistics is concerned with prediction on basis of patterns, but that doesn’t rule out probability theory. I don’t say that trading is the same as gambling, but since no one can predict future stock prices with 100% certainty, it has in part probabilistic features like the regression effect. You can easily be fooled by them.
Randy,
Entering a system with a long-term positive slope is a sound strategy, but I still don’t see why this would be better at a dip. I understand it for stocks or funds, but I fail to see why that would also apply to systems (apart from the two exceptions I discussed above). I think that it is no more than a weak association on basis that they both have curves.
You don’t sound like him. But I admit, I too was afraid that others would think that of me, given the length of my posts.
Jules
I think that it is no more than a weak association on basis that they both have curves.
But the most significant information available for evaluating a trading system offered on C2 is the historical data contained in the equity curve (and the associated details underneath it that can be gleaned from the P/L table columns). If the assumption going in is that the equity curve past history has no value in predicting future performance of the system (essentially the standard disclaimer on any offering to the public) then everything is a coin toss and it is a waste of bytes to even provide all this information for potential subscribers to injest.
My own C2 system picking results would certainly support the conclusion that future performance is in no way related to past history! But I would never consider subscribing to a system with a negative equity curve slope over the long haul.
If the assumption going in is that the equity curve past history has no value in predicting future performance of the system
No no, that’s not what I mean! Rather the opposite. Of course the long term trend is very important! And the frequency and size of dragdowns are important too. These things are important to decide whether you want to follow the system or not. But the discussion between Walter and me was: Given that you want to follow a system, what is the best time to start with it. Walters opinion (as I understand it) is that the best time is to start when the system has a dragdown. My opinion is that it doesn’t matter whether it just had a dragdown; if it is a profitable system you can better start as soon as possible.
Jules
