I have great passion for automated trading, the pictures below reflect one of my strategies for S&P emini future ES from 1/1/2017 till 29/6/2018 (based on back test) Profit: 60% Max. Drawdown: -2.8% Sharp Ratio: 1.55 I love developing trading strategies and good at programming. I am seeking cooperation. If you are also interested in this field and have any idea of potential cooperation with me, don't hesitate to PM me! Thank you!
I think you are missing the real picture. First of all, automated trading (or quant trading of any sort) is equally (if not more) labour intensive as discretionary trading is. There is a lot of "gardening" required to keep your strategies going on daily basis, a lot of work reviewing the results and deciding on changing the strategy scaling and finally, lot's of research work involved to keep coming up with new alphas as the old ones slowly die. It's Sunday night and I am at the office supervising my strategies in Asia, so I know what I am talking about. Second of all, most of the real life research on quant trading is very different from the ones you see in "automated trading for beginners". On one hand, books spend a lot of time on things like win rates and profit factors are not usable in real life. Your thought process should be more along the lines of Sharpe/Sortino/max DD for understanding the quality of the strategy and things like PnL per trade value for understanding the execution concerns. On the other hands, the usual issues about overfitting (such as setting data aside for forward testing) are less important if you have a genuine understanding of what phenomena you're capturing and why it exists. There are better ways of understanding the stability of your strategy than forward testing. If you are still passionate about quant trading after consuming the red pill, sure, it's a worthy enterprise and a lot of money can be made.
With that profit:drawdoown ratio 60:3 your sharpe doesnt seem right. On the low side. How are you calculating it?
Ya, good catch. Lowest possible SR should be on the order of 2.5. Assume that his worst DD was a single day and that was also his daily return volatility while his mean return is 60%/120 days. 16 * 50bps/300bps =~ 2.6. That would also make the returns severely not normal, so that's a very unlikely scenario.
I see what he has done. His quoted sharp of 1.55 is his (monthly average) / (monthly stdev), 2.65/1.72 But he needs to multiply that by SQRT(12) to get an equivalent annual sharpe ratio so its 1.55*SQRT(12) = 5.36
How many total trades in this report and win-loss percentage on these trades. Most of the trading has been taken over by automated computers following strategies, so those following automated path are "with the trend" in the industry.
I am highly curious about this as well but unable to program anything for it sadly. Ill aim to keep watching this
AlgTrd2018, List what programming languages you know, any APIs you have written into, data feeds you are familiar with, etc. and I think you will get a decent response. The image and info of a more than likely curve fit system is meaningless.