I have spent a year building strategies, and I have come a long way since the beginning. I have built a complete ATS from scratch for TWS, and a complete independent back testing and plotting framework; I built an SQL database and data cleaning routines to automatically download and store all my minute data; I have found a few algorithms that make money consistently with full accounting of slippage and commission; I have forward tested those and confirmed they do agree with my back test on a trade by trade basis at least for the last 4-5 months. But so far, the best ones show a PF of about 1.6. I have many ideas that I would like to implement that I am certain should improve my PF. However, in the last month, I have hit a brick wall and my progress have become slowed to almost a halt. I can't describe it well but it feels like a version of writer's block for programming. Everyday, I sit down, and I spec out my new signal and I start writing subroutines that implement it in my back tester, and then I get stuck. I am unable to make the decisions that I used to make quickly and test the idea. I try to force a decision and start writing code, and then I later scrap all of it and start over because of some problem -- e.g., the routine gets too long and messy, or the logic would require too much rewriting of old stuff, or I start to doubt the design of some of the lower level classes and start thinking about re-implementing some of those, or I even start doubting the signal and I start searching for a better definition of my signal, etc. etc. So for the past month, I have been spinning in circles, not able to make progress and code in the new ideas that I am almost certain will improve my PF. It's gotten to the point where every time I sit down, I get a sense of dread that I will end up not being able to implement anything again that day. I guess I am kind of burnt out but I am not sure why. I am almost certain I can make significant improvements, but somehow deep inside, I am tired of it or I have huge doubts or something. Anyone else ever experienced something like this?
You do sound burned out. I'm not sure what your finances are like, but you need to go somewhere, where you don't code, don't follow the market, just clear your mind and have fun. Come back, and try again. This can happen to anyone. Take a break....relax.
Everyday, I sit down, and I spec out my new signal and I start writing subroutines that implement it in my back tester, and then I get stuck........ .................................................. a new grail writer pissing in the wind. curve fitting dead data until the string is so tight it blows out in real time. just another trader lying at the side of the road with the backtesting needle hanging out of his arm. if you cannot measure probability and trade the hard right edge, walk away from the game. markets are simple, irrational and stupid all at the same time. this will continue until further notice. gotta luv retail. if you ever get around to real trading..... when we puke it out, make sure you are short. cheers, s
Travel somewhere, take a break. People are not machines. I've been where you are now, a change of scenery helped.
well, lets put it in perspective. 90% buy and hold, 10% trading, of the trading 9% trading the system, 1% completely discretionary. I don't know what a vacation ever did for you, but all it ever did for me was make me a nervous wreck. How's a guy supposed to enjoy his life if he doesn't even know what the market is doing? I don't do anything in a small way. When I took my break it was complete. No computer, no TV, the only thing I ever read at the coffee shop was the sports page. I was long the S&P and never even knew it crashed. When I finally came to, everything was about the same except for corn and crude.
.............................. so whats your problem. trade it. let us all know how it is going in 12 months. turn the flat off, keep the box running and walk away.... the algo will handle it , right......? shop out
my problem is simple: the stats are still not good enough for me to be comfortable going live. I am close though and I have been steadily improving those stats in the last year. the only problem is that the algorithm is a little too selective and the number of trades are low so I am not confident about those stats yet. I have a few concrete ideas to implement - once I put those in, I expect to see increased trades, and hopefully the same or better stats. then I will go live. if I can't improve my stats, I could be here for a long time without going live -- better than taking unnecessary risk and losing.
the stats are still not good enough for me to be comfortable going live. .................... news flash. the stats will never be where you want them. ever. just the way it is. been there, done that. not the way 2 go..... s