The opinions expressed in these forums do not represent those of C2, and any discussion of profit/loss
is not indicative of future performance or success.
There is a substantial risk of loss in trading. You should therefore carefully consider
whether such trading is suitable for you in light of your financial condition. You should read,
understand, and consider the Risk Disclosure Statement that is provided by your broker
before you consider trading. Most people who trade lose money.
Just wanted to stir up a hornets’ nest a little bit and ask you about your experience with algo systems.
I am investing with algo systems only as for me it is more efficient and generate better results in long term.
Never missing an investment opportunity,
Strictly following the strategy
Having clear and effective money management and risk management systems for algos
*Backtesting and results analysis of 1000+ trade samples before going live
are the main but not the only advantages that I can think of.
This is my point of view as I know every detail of my strategy. For the investor however I imagine it might look like a pandora box.
What is your perception of alogs?
What would you like to know about algo so that you would be willing to follow it?
What advantages / disadvantages of algos would you do you see?
Happy to discuss with any of you and confront my views.
Just haven’t found an algo system that is consistent and stays within the rules of what they claim. They all claim as algo driven, then when shit hits the fan everything changes. Frequency of trade, size, double or tripling down, all the signs not algo trading. Or some just don’t perform.
Both Algo trading and traditional trading have strengths and weaknesses. Just stick with algo trading if it gives you greater efficiency and gives better results in the long run.
I think the adaptability to suddenly changing markets is the best strength of traditional trading and the weakspot of algos. Also with traditional trading you usually have clearly defined setups as well as risk- and moneymanagement so I wouldn´t count that exclusively for algos. However, there are more loosely defined trading styles aswell.
It´s just that human nature gets in the way of rules more often than we like. But if you as human work on doing fewer mistakes you will outperform most algos in the long run because at some point they just don´t fit the market anymore. This assumes that you are one of the 5% of people who can adjust yourself consistently and adhere to a strict risk- and moneymanagement.
By the way, the market is always a blackbox going forward and Index investing is also an algo. Arguably one of the more successful ones though.
You touched an important topic here regarding changes on the market.
The question is if the markets are really changing?
My take on this is- yes but in a repetitive way. The question is how much time does it take between repetitions. 1, 2, 5 ,10, 20, 50, 100 years?
From my experience with data, statistics it seems that 15+ years which include the full cycle of high low and volatility is a breakpoint which allow for some conclusions about the strategy performance. In this period strategy should place at least +1000 trades.
Less than this - usually not enough to say anything. Had many strategies that seemed brilliant on shorter periods, but went quickly bankrupt after extending test period.
If we assume that this 15 years period is more or less accurate how can any investor trust an asset manager who has less than 15 years experience?
How can I know that he/she knows how to operate when things go south?
With algo it is much “simpler” as you can test it in any conditions, and tell him in advance what your algo should do in that case.
And if you as an algo developer don’t know that - just tell him not to trade. 0% return in some situations is an excellent result.
How in the world do the rising influences of 401k (etc) investments, passive index funds, HFT and algorithms have any kind of “repetitive” predecessor? I can see that the market goes up and down, and I’ll grant you that is repetitive, so if that’s what you mean to ask when you say “how much time… between repetitions”, ok. But the influences of other changes (a few mentioned above) are monotonic or even episodic, and are not readily backtested. Noise and signal are not easily separated. I prefer identifying a rationale for my signals before I look for them, so that I have some sense of whether I’m finding signal or just noise. If I can’t explain why my algorithm works, it’s not mine, and I won’t trade it.
You are right that the reason is unique, but the reaction and price developments that are result of it might not be. Question is how often such similarities occur and how important they are for our investment plan. If you make long term trades then rising 401k influence will be a factor you have to consider. It would be also probably not possible to find a similar example in a past because the history is too short and has not generated similar situation.
However if your trading horizon is within next few hours to few days then this 401k thing is probably a noise for you. You are also in much better situation, because there are many repetitions of this “few hours” period in a past to analyze many more possible price action behaviors.
My conclusion from this - for algo development you have to choose an investment horizon that will match the historical data period that you have for your analysis, so that number of repetitions of your investment horizon is high enough. For my trading style 15 years history seem to be a breaking point.
To give you an example
Let’s say that I have developed 10 different algos with similar investment horizon and tested them on 10 years period and they all performed well. Then I extend testing horizon to 15 years in total, and 7 of those are not good anymore. 3 are still performing well. Then when I add another 5,10,15 years, those 3 decent ones are still performing well.
…
