No one is amazed at anything. All that was stated is that you have a very small chance after factoring in all other things previously referenced to be successful. Again, anyone with a legitimate two year plus track record of automated trading can chime in at any time to slap me in the face. Real money btw, not paper trading your way into this track record. It isn't the same, promise. Your best bet, if you are good at programming, is to sign on with another firm and make your small 6 figure salary. Let them take all the risk. There is a better than not chance they will be a net loser after factoring all costs. There are very few firms I am guessing ( without having the proof in front of me ) after factoring in all other costs that consistently are profitable. I have ran into plenty of people in my short time in this business with the money to burn. If you enjoy losing money and enjoy wasting years of your life on research in exchange for meager results you should get into this business. It will suit your personality type well. The self destructive type - I found out I might be one of those people. Trying to change my game up currently.
I've got to chime in Algo Kid. The process of creating an algo requires skill. Trading requires skill. The majority of traders fail. I would also guess that the majority of autotraders fail. Why do auto traders fail? Probably because they trade indicators not markets. Or because they try and trade a black box. Being a quant is the wrong skill for directional trading. Building an expert system firstly requires an expert who understands why he is making money, and who can write his process down, step by step. Converting that to an algo requires a skill but the necessary skill is being able to trade. Many people think that something can't be done because they can't do it. It's very easy to give up before you start. As Edison said: Genius is 99% perspiration and 1% inspiration. there lots of us out there making money with algos. You just need to learn how. ElectronicLocal
I 100% agree with you. This still though does not overcome the obstacles that I and others have previously described. There is no perspiration or inspiration that can overcome the inherent flaws of current software architecture.
I wanted to edit, but I was over the time limit. I believe in your example you would never need to backtest this idea, is this correct? You have already validated it from years of discretionary trading and are just looking to put that into some sort of programmatic design? Then sure I understand this. On a side note it would be interesting how the backtested results would compare if you could somehow have exactly the same timing and all other elements aligned. I think that is getting away from the OP a bit though.
I love your analogy comparing trading to playing a guitar. Many traders make a lot of profitable sounds but few make the sweet sound of making consistent profits that is music to our ears...