Is it possible to create an automated system that will yield an average of 100% gain every year under any all market condition? The reason I’m asking the question is because I haven’t found one yet. But I think it is possible to create such a system and it has to be an self-adapting system that will learn the market and will be able to distinguish a trending market or range bound market and trade accordingly. I found that most systems on C2 work for a period of time and then either go into draw down or disappeared from existence. Since these systems are NOT adaptable to the market and therefore can not work all the time. An adaptable system is doable in my opinion but is difficult to implement. I’m asking this question since I would like to stimulate some developers into trying to create such a system. If this system exists, it will benefit both developer and his clients.
Thanks for your response.
Anh Nguyen
Anh,
The systems that I have developed are essentially fully adaptive, and this requires 5 hours of cpu time each evening to continuously optimize the various models for the past 9 years (and growing each day) of back-testing data. Now, attaining 100% per year would be extremely hard for stock/ETF-based systems, but my IWM (Russell 2000) signal has successfully yielded about 30% per year over the past 3 years that it has been tracked by TimerTrac. I could see doubling that by using the 2x ETFs (UWM/TWM) … and in fact I am attempting this with one of my Collective2 systems … but it would be quite hazardous to push for 100+% annualized gains by using the 3x ETFs (TNA/TZA). Anyway, I have found that my models start to degrade if I do not continually optimize them – and from time-to-time add new algorithms (and associated empirical constants) to extend the models to handle new kinds of observed market behavior.
Hi Anh,
Good question.
I don’t think that’s possible.
Average of 20 to 40%, even 60% maybe, but no higher. Even then it is very likely that there will be the occasional large drawdown - changes in market dynamics (trending to ranging etc.) are only really obvious after the fact and any adaptation must necessarily be based on past observations from which the future is estimated.
So gunning for 100% necessitates excessive leverage, whatever the system employed, which increases the magnitude of drawdowns to such an extent that sooner or later any real account will blow up.
I believe it is no coincidence that you have not found such a long term system on C2, or anywhere else for that matter. The ultimate adaptive system, though fraught with emotion, is the human brain - the track records of people like Warren Buffet, George Soros and Anthony Bolton (up to 20-30% per year over the long term) are telling. If anyone knows of an example of a long term 100% average investor/trader/speculator/system please do share it with us.
Dean.
Dean,
I think you’re right. Without using leverage I think that it would be very hard to do better than about 40% per year when trading model-friendly ETFs like IWM with well-tuned adaptive models. I have learned that it is hard to obtain more than 1/10th of back-tested annualized gains in actual trading, and with IWM back-testing at 435% per year this would work out to 43.5% max. And, of course, drawdowns will always be much higher than the back-tested numbers show – perhaps by a factor of 3 or more.
If I were to try to achieve 100% without leverage, I would turn to a small group of select highly volatile stocks like AKAM, ILMN, NVDA, JOYG, and RIMM. Using the 1/10th ‘rule’, a portfolio made up of these stocks ‘might’ make a bit over 100% annually, but even back-tested drawdowns are in the 23% range – so actual drawdowns could be 40-50% or so. I am actually trying this in a Collective2 system so in a year or two we can see if this works or not…
Anyway, I will be quite happy if I can reliably produce 40% annualized gains with acceptably low drawdowns – and that will be quite good enough!
Not to disagree with Dean’s reply above (which people should read at least twice - it is that good), but Yes, 100% annual returns are possible.
I actually traded a self-developed system from 2004-2007, and it averaged over 100% per year, each year. 2005, 2006 and 2007 results (148%, 107%, 112%) were for a worldwide trading contest, and the results were fully audited and traded with a $15,000 starting account - real money (you can probably verify all this with google).
BUT, that being said, Dean is 1000% correct: “gunning for 100% necessitates excessive leverage.” And that is why I scaled back my leverage with that system after those years (I still trade it today). The drawdowns were too extreme, and I got tired of them.
My rule of thumb: assume your drawdown % will equal your annual return %. So, if you want 25% annual return, be prepared for a 25% drawdown.
100% return? Sure, just be prepared to lose 100%. I was prepared for it, and I got a bit lucky, and that is why I was able to do it.
I would not recommend such an approach for anyone trying to make a living trading. Eventually you will get wiped out.
Aim for 20-30%, max of 50%, and you’ll be in top-tier investment company.
Kevin
I’m looking for realistic opinion for me and for other traders as well. I’m working full time and try to look for stable systems to trade. Got burned once by Enchante system. Now when looking for systems to trade, I scan the $DD column and if $DD is comparable to P/L, then it is acceptable for me to trade. Systems with excessive draw down will eventually crashed and unrecoverable. I saw quite a few new systems with large accumulated profit for a short time on C2 but also incur HUGE draw down. These systems will eventually crashed. I saw quite a few of them already crashed and I predict the new ones will also be crashed when the huge draw down turn into loosers. It’s good to hear from developers and traders and it will re-enforce a realistic expectation when choosing systems for trading. I want to make money and be able to keep them too. Thanks everybody for the contribution.
Anh Nguyen
Well here was what I was taught in high school.
In ancient India a mathematician relieved the boredom of a king by inventing the game Chess. The king asked what reward the mathematician would like. Being a humble man, the mathematician asked only for rice to feed his family. Only one grain on the first square, two grains on the second, 4 grains, 8 grains, etc., until the end of the chess board.
The king gladly consented. Unfortunately, according to my math teacher, that amount of rice has not been grown since the world began.
Also as you probably have heard, $10k in Berkshire Hathaway in 1960 would have made you 1 billion by 2000, after compounding for 20% per year.
I.e., anything above 20% gain tends to cause serious problems, such as figuring out who owns the planet, or whether or not money has any value.
Incidentally, if the stock market were illegal, bonds would compete with higher rates, inflation would subside, billionaires would have less unequal market leverage, there would be fewer wars and recessions. We would all be generally better off making only 5% interest without the ups-and-downs created by the market system.
Incidentally, the most popular god in Bali is called "Chatra Muka" meaning 4 faces. (Something like the Roman god Janus = January who meets incoming and outgoing years with 2 faces x 2.) Chess is popular in Indonesia and its name is "Chatur." This probably comes from Hindi and is probably related to the name for the Audi Quattro. I believe that the original name for "Chess" if translated literally probably means "Squares."
In summary, there is a good chance that it is not mathematically possible to make 40% consistently and it is definitely not possible to make 100% consistently.
{}
Your conclusion is based on the assumption that compounding of returns is required.
Good point. If, instead of compounding, you only put in 1 thousand or one million, and took out all the gain each year… then you have a better chance of sustaining the 100% without owning every company on the stock market and gambling against yourself… Also please note, this is third party quote but, supposedly Warren Buffett has claimed that if were not a billionaire, he could make 50% gain annually, by doing small-caps instead of being forced to buy entire companies.
On the other hand, instead of attempting to make 100% annually non-compounded… why not find a few c2 systems that have a good chance of consistently making 40% annually, take out 1/3 the gain each year, and compound 2/3 the gain…? If in addition you remove rebalance every 1-2 months to equalize the gain between 3 or more systems, you will cut drawdowns in half, and increase net gain by at least 10% annually. I.e. by targetting 40%, diversified among 3 systems, you have at least 5 times better chance of making the same money with 1/5 the risk, compared to finding one system that makes 100% annually non-compounded. My current top picks:
#1. [LINKSYSTEM_46439436]. Stocks. CAGR 42%. Max DD -14%. Monthly fee about $43. Base value 60%.
#2. B1Global. Forex. CAGR 89%. Max DD -12%. Monthly fee about $50. Base value 25%.
#3. sp500-trading.com. Futures. CAGR 66%. Max DD -31%. Monthly fee $50. Base value 40%.
#4. Shorting Options. Options. CAGR 37%. Max DD -21%. Monthly fee $10. Base value 60%.
#5. Rebound Solo. Stocks. CAGR 39%. Max DD -20%. Monthly fee about $20. Base value 80%.
P.S. Craig Schulenberg who joined this discussion above is a real mathematician. His systems are all thoroughly backtested and some include direct third-party verification over several years. The Schulenberg systems may have a better chance of sustained 30% to 40% gains than anything on this list. I suggest you read the descriptions which include links to his website where you can study the histories carefully. The Schulenberg systems are not on this list because his fees are $99, whereas I try to find systems costing $50 or less. However perhaps you get what you pay for.
P.S. I just noticed that Kevin Davey is also a serious trader and I also suggest anyone take the time to browse his website.
""""Kevin Davey is an award winning private futures, forex and stock index trader based in Cleveland, OH, USA. Kevin has been trading for over 20 years.""""
http://www.kjtradingsystems.com/
And there may be others of this quality at c2. My "top 5" list is relatively superficial, prepared quickly based on readily available Grid information.
Yes, compounding is the only way to compare apples to apples.
Compounding example:
You put 100 into a system making 20% per year.
So it starts off by making 20 in the first year; i.e. 20% of the capital. After 10 years the account is at about 620 and that year it makes 124; i.e. 20% of the capital.
The percentage return on current capital stays the same as time goes on.
Non-compounding example:
Same system except you keep the gains in cash instead of reinvesting.
So it starts off by making 20 in the first year; i.e. 20% of the capital. After 10 years the account is at 300 but that year it still makes only 20 on the invested capital of 100; i.e. 6.7% of the total capital.
The percentage return on current capital drops as time goes on.
If you take the gains as income then there is no increase in capital at all so the two examples are the same.
Hi Anh,
I think your question…
"Is it possible to create an automated system that will yield an average of 100% gain every year under any all market condition?
…is the wrong question.
From what I found out during the last years of research your expected gain depends on exactly three things:
1. Predictability: Are you able to find situations in the securities market which give you a chance better than random for a profitable trade?
2. Exposure: How much such situations are you able to find?
3. Risk: How much of your capital can you risk in a single trade without going broke sooner or later?
All three points are not very well understood by most (traders and investors).
Let me illustrate this concepts a bit. Generally the securities markets are very noisy. Most movements are purely random. It is very hard to find situations with a somewhat better probability for a favorable future price movement. The relevant measurement here is “average profit per trade relative to account value”. And guess what, this measurement is not available form C2’s stats for a system. (Win rate alone is misleading as are most other stats).
[Example: Topaz has an average profit of abut 0.04% per trade. Seems very low, but even this leads to a hefty risk of about 40% expected drawdown.]
If you have a method to find “High-Probability-Situations” the next question is: How much of these situations can you find and use with your available capital.
[Example: Topaz finds on average about 2.3 such situations per trading day in the Nasdaq-100 Universe. It would be possible to do similar things in the Rusel-3000, in Europe, in Asia. Of course you need much more effort and capital to exploit all these opportunities]
Risk. This is the most important component. Think about “fat tails” i.e. unexpected clustering of losing trades. You need plenty of “headroom” for the “big crashes” which used to happen once in 20 years but seem to accelerate.
Now do the math: 0.04% gain per trade are possible (see Topaz). Probably 5 times per day…
Thank you Dean, another math teacher. I suggest everyone always take Dean seriously folks. Here is my result following Dean’s suggestions. It’s like watching three turtles catch a hare. And keep in mind, the slower the turtle, the more likely it exists. The hare probably does not exist through investing. There are however a few countries in the middle east in need of a new dictator.
Gain & Allocation… Principle… 3-yr cum… 6-yr cum… 9-yr cum
40% compounded… 10,000… 27,440… 75,295… 206,610
30% compounded… 10,000… 21,970… 48,268… 106,045
20% compounded… 10,000… 17,280… 29,860… 51,598
100% no-compound… 10,000… 40,000… 70,000… 100,000
Mr. Koch has one of the most promising c2 systems, and his Topaz is also a good example for my points. Topaz has turned $10,000 into $40,000 over the past 5 years, rather consistently. This seems about equivalent to my results above for 30% compounded.
Please note that c2 lists the CAGR for Topaz as 44%. I am told the c2 CAGR does not include trading fees. Therefore Topaz is really probably doing about 30%. This may seem disappointing, but it is good news. The slower the turtle, the more likely it exists. In my research on c2 systems, there seems to be a “ceiling” of about 45% CAGR. There are numerous systems which show 40-45% CAGR over 2 years, and very few above that, and these few have huge drawdowns.
This means that the dozen or so c2 systems which are reporting about 44% consistent annual growth, are actually achieving about 30% growth. This in turn suggests that 30% consistently may be possible. This turtle may exist! I.e., c2 results might be viewed as clinical trials that indicate probability of existence.
However, keep in mind FEW of these systems have seen 2007 and 2008. Topaz is one of the few. It broke even for 2008. That was very good. It did very well for 2009, not unusual. It did well for 2010. Now for 2011 it is having problems. I.e., there are only 2008 and 2010 that I can say Topaz performance was outstanding. This is very promising, but time will tell.
You can go to TD Ameritrade now and screen for ETF’s and mutual funds that did 30-40% over the past 2 years–you will find many. But then looking at their 10 year histories, you will be very disappointed in their results!
Also, in addition to system fee and software fee, don’t forget the short term capital gains taxes. If you don’t have the correct account size and the correct tax structure, you might have to deduct another -10% from the annual gain.
