Thank you themickey, I respect your comments but I disagree with you. Maybe I am just inexperienced (only be developing my own strategy for about 3 months. Learning to program and my platform) If I do a backtest and a 5 year report shows a year after year profits to my liking, then I will click start on that algo and never click stop. Otherwise, why are doing this for? Why are we backtesting and not planning to run algo as is. Maybe is don't have have the experience yet but for now, i will keep that simple unless proven other wise.
We never stop learning with trading, it just never stops. You will design an algo and continue to modify and improve. There is a very high probability you will scrap it entirely for a different model. That's just the reality of the evolution of trading. I would imagine in institutions they would be running a number of different algos, some related, many unrelated to each other, all producing varying results, profits and losses coming and going in waves.
https://www.jdaltontrading.com/discretionary-trading-vs-rules-based-trading/ This is a nice article I found about the difference.
Thanks themickey, I can understand what you mean, that the algo may need updating, but not soon as I click start on it. I will let the algo hit drawdown before stopping it. Again, this is just my thoughts.
Imagine this scenario. Start algo, churning along, make money, lose money, lose more money, clock is ticking, lose more money, "wow, apply the brakes, let me think about this a bit longer, if I lose all my money I'm done for". No point committing slow suicide because of a vow "not to touch my running algo".
Thanks themickey, Good comment. When I click start on algo, i will have capital im that algo that's meets the backtested drawdown requirements. After clicking start, the algo will either hit that drawdown, losing all the money invested that algo or rise to higher expected profit like it did in shower in back test. In summary, it will either lose all the money or alot of make money. There is no in between. Once the algo makes dounce the drawdown, i remote my initial Investment in the algo, or increase contract size. I placing all my bets in what the back testing and other testing methods show me.
Me and a partner have tried automating and backtesting my DOM read for years and have not been able to recreate the bias I come up with from just watching order flow. I'd love to do it but not sure if it's possible... in the meantime, the less information I try to digest during the trading session, the better I trade. I'm down to one chart that I reference a few times a day, otherwise I'm just watching depth of market. Maybe that's old school, but it makes money!
I appreciate your comments, intent and sincerity, but my guess is that once you do this you will probably not attempt to do it again. It is nigh on impossible imo to find a 'one packet' solution to trading or a one stop solution to an algo. Why not think like this; "I expect my algo will perform perfectly to my expectations, this is calculated the rate of return I should expect..... Should said algo underperform by greater than 10% (or whatever) over a period of longer than x days, I will consider preserving my capital" Because that is the crux of trading - preserving capital, not 'the need to be right'
Unfortunately that's the reality of it. I live it everyday. It's the aggregated performance of all my algos that smoothes the curve. Lots of capital is needed once you move into such a multi-dimensional approach.