I have back tested my system extensively however that is no guarantee of success going forward. If the markets are like they have been in the past i will have a chance of making money, if they change too much i wont. Hopefully they change slowly enough i can still eek out some good profits before the systems degrade to point they not worth trading anymore. I don't believe there is any holy grail system that is robust enough it will continue to work forever.
I can feel a real doze coming onto this journal. There are systems that do work year in year out, but that's not month in or month out. Just annually they should always be profitable or you shouldn't be trading it.
For something systematic to work every year it has to generate lots of trades. A low trade system will just have too much variance in returns. Something that generates lots of trades will also generate lots of transaction costs. A retail trader doesn't exactly have the lowest transaction costs. These transactions costs will take the system negative in a bad year. Anyone can design a system that seems to work very well on past data. Has 100% profitable years. But going forward it is much harder to maintain as market conditions change.
Kind of my thoughts, too. How does one every really know if they have an 'edge'? Isn't it always a matter of current market conditions 'suiting' your methodology?
Yes. I agree. You only know you had an edge over the back data or in the past. And you might have an edge going forward (if the market continues to suit your strategy). Trading systems are not like robust roulette wheels, which are in effect holy grail trading systems for the casino, as the roulette wheel edge will always hold up.
It's only roulette wheel if time is short, as in the spin of a roulette wheel and day trading. Swing Trading with greater than 2:1 win ratios and 1:1 win/loss ratios are profitable, and there's quite a few I've written that have this characteristic but not all are profitable year in year out. You're basically admitting that your system hasn't had profitable simulations in every year so I don't think you should keep using that and I'm not talking about how you're really doing.
A system (lets call it system A) might average 50% a year, with a variance of 50%, so returns fall between 0% and 100% a year (with 99% confidence if market conditions holdup) System B might average 50% a year with a variance of 25%. So returns fall between 25% and 75% for System B. My back test results are closer to system A. (Assuming 1% risk per trade, with 2% risk in theory i will double returns but with double the drawdowns). What kind of results are you getting? especially from a Monte Carlo simulation which can give a better idea of future variance of returns. Now the point im making is lets say the worst case in my back test is 10% profit in a year. Now that assumes the back test is a good match of future market conditions. As is more likely the future will not be exactly like the past, so the chances of having a negative year are probably higher than the back testing and even the monte carlo would suggest.
It takes like 25 years to prove you have an edge on performance alone. Otherwise, you know if you have an edge. It will be obvious - like access to inside information or better funding than fed funds, etc.