Jason from the Angry Investor blog has proposed a semi-mechanical volatility system in both my chatbox and his blog here - http://theangryinvestor.blogspot.com/2008/12/blog-capsule-my-dow-jones-futures-for.html
I thought it would be a good idea to comment on his system here than in the chatbox, since it is more enduring than a chatbox comment, and as he correctly noted, there is practically zero / minimal risk of moral hazard for Malaysians to comment on a global and highly liquid trading platforms such as the Dow Jones Futures.
The first thing to note is that the mechanical volatility system proposed by jason is appealing in its simplicity. Every night, note the DJ closing value (say C). Then, enter both a Limit Buy and a Limit Sell order which is C-75 points and C+75 points. Then, go to sleep and pray. The next day, hopefully, both trades - Buy and Sell are executed because of the high Dow Jones volatility. That's basically what jason described in his blog.
However, jason appears uncertain, because subsequently, he altered the parameters from C+/- 75 to C+/-100. This begs several questions:
1. Is this the final amendment, or will there be future amendments?
2. If there are future parameter amendments, under what conditions would such amendments be made? Is jason able to describe such conditions that would trigger such amendments?
3. Is this a genuine mechanical system, or is this a semi-mechanical system, "semi" since in reality, the system parameters requires jason's own discretion as time passes? For example, jason has not described how he came about with the new +/-100 points change from +/- 75 points change. For example, will future changes be larger or smaller? What conditions would trigger such a review?
On the system itself, jason did not describe what would happen if only one of the trade was executed but not both. For example, suppose Buy at C-75 points was executed, and the market subsequently tanks. A few questions arises:
4. At what point would jason set the stop loss? C-175? C-200? No stop loss?
5. When would jason excute his stop loss, since he has gone to sleep after entering the limit orders? The next trading day?
6. If jason has an open (losing) order such as the one described here, will he continue to trade with a second contract (i.e. requires extra capital) or put everything on hold until the first contract is closed? If he opens a 2nd contract, is there a limit to the maximum number of open contract should he encounters a string of losses, and if so, what is that maximum limit? Furthermore, since he uses Futures with margin calls, where is that final limit?
Another comment I like to make is that when jason's proposal is back-tested over the past couple of months, it looks impressive because it works and it is so simple! Just a quick glance at the DJ charts in the past 2 months, and jason's system works. However, the critical question is will this continue to work in the future? And does it really work when back-tested further back in time?
For example, imagine we are at the close of 19 May, 2008.
On 19 May 2008, DJ closed 13,028.
According to jason's proposal, set a Buy Limit at 13,028 - 75 = 12,953, which would have been triggered the next day.
At the same time, the Sell Limit at 13,028 + 75 = 13,103 would NOT have been triggered, because the High on 20 May, 2008 is 13,026 only.
The question now is a critical one, because the hypothetical trade on 20 May 2008 is one example of a trade that has clearly gone wrong. Jason had expected both Buy and Sell to be executed, but it didn't. Instead, only one trade got executed. The question is would he then liquidate this trade the next day regardless of the size of the loss (or profit) on 21 May, 2008? Or will he just hold and "hope" that the DJ would climb back to 13,103 to close his position at a profit? Is there an equivalent mechanical rule for trades that have gone wrong? Or when trades go wrong, discretion is required?
This is a critical and extremely important question, because if he continues to "pray and hope", then, he is guaranteed to be killed one day (it's just a matter of time), because as it turns out, 13,103 is not reached again, even as of today nearly 7 months later. And let's not forget that he would be pretty much wiped out if he continues to buy and hold until today, as the DJ has fallen to around 8k, or even lower at the lows of 21 Nov 2008.
In other words, if he only keeps one contract open at all times, jason would be out of trade for nearly 7 months. Furthermore, he would not have participated in subsequent volatility gains which are extremely significant.
On the other hand, if he cuts loss, he must set his stop loss level extremely carefully because on other days when volatility is high, he would also be stopped out as well, even if the trades turn out to be profitable later in the day. There is no free lunch - damned if you're stopped, damned if you're not.
In other words, markets are too smart to give someone a free lunch for too long a time. It doesn't mean that markets cannot be beaten - it can, but the truth is, by only a very small % of players consistently like Buffett. Jason's initial guess that this might work for the rest of December month might turn out to be correct, however, it still requires Jason's discretion to decide whether such a system will work at the start, mid or end of 2009.
Another thing that jason has to watch out for - when trading mechanical system - is that there is No Holy Grail. There is no system that gives a 100% win rate. You simply will have a loss / a string of losses one day. That statement is pretty much a certainty and a fact over the long term.
The critical question is whether jason will have the discipline to continue or stop trading his system after he has a string of such losses, or whether he will be tempted to "fine-tune" and alter his system. His article indicates that he is tempted to "fine-tune", and he needs to be aware that "fine-tuning" will then turn his mechanical system into something else - a "discretionary" system - in other words, the success/otherwise of his "evolving" system then becomes dependent on his skill / otherwise in fine-tuning his system. Markets have shown that only a very small % of traders are successful in such an approach, but the majority will probably get it wrong. (In other words, his originally simply described system can be thrown out of the window when it evolves).
Another suggestion which I would like to give jason, since he strikes me as a hardworking young man, is to keep a daily record of his profit and loss from trading this system. Identify the starting capital now, rather than look back and then find the ideal capital size that would give the highest % return retrospectively (this is called "back-hand"). By identifying the capital NOW, he can then calculate the % returns appropriately at a future date.
A real trading portfolio is almost certain to give different results than paper trading. On paper, it is very simple to show good results. But in day to day real life trading, emotions, doubts, fear, greed, trading costs, slippage in execution, lack of discipline, lack of emotional control (widely documented in trading literature), unexpected sickness (e.g. grabber's recent dengue :-)), personal problems (e.g. divorce, etc.), holidays (e.g. Seng's Taiwan holiday), etc. etc. etc.comes into play and actual results is certain to be different from paper trades. I have yet to see actual trades turns out to be identical to paper trades, especially after the system have been traded for an extended period of time. Hence, I believe jason should test it with an actual portfolio since he can afford it. *grin*
Anyway, hopefully, this is enough food for thought. I would like to take this opportunity to thank jason for sharing his ideas with us, and his open and positive attitude in welcoming comments to his thought provoking article in my cbox and his blog.