Jtrader, do you have a rolling data base with all the stocks you track that constantly updates at session close everyday?
Yes, markets are constantly changing, however they are not going to email you next time they change - as a result, the best is to ensure your strategy can survive or better adapt to varying market conditions, and for this you need a lot of historical data. I am an automated trader, so I can't speak to what's good to discretionary traders. But for me, backtest only has value if it executes the same code-path as live trading (read this, many times, until you are sure you understand what it means and all of the implications). Because there is no other way to test & verify the system you are putting live (I consider market replay as a form of backtesting). So at some point (if you are going automated) you will have to develop & backtest on whatever platform you'll execute live. I find it much more effective not to duplicate efforts, which is my point about not testing an idea in Excel only to have to develop it again in Ninja/MC/whatever. Of course, I use Excel to do statistical analysis (and I know I would be better off using R, just getting started there), but there is a world between statistical analysis and creating a fully functional trading system. 1/2 decade is my own experience. At least after that I have traded profitably for the last couple of years, fully automated. Most fail at ever reaching consistent profitability, most probably because they give-up at some point along the way - or run out of money.
Dom, really appreciate this! "But for me, backtest only has value if it executes the same code-path as live trading (read this, many times, until you are sure you understand what it means and all of the implications)." This really stuck out like the only highlighted line in a sheet of paper
Luckily for me I am a experienced software developer in .NET and have developed systems for banks etc. I decided it was better to write my own back testing/trading software from scratch. I know this may not be an option for most people but I think it gives you the best flexibility. It also allows you, as was mentioned above, to have the same code execute on the historic prices as well as live (when you finally go live). All you need is a quote feed and a API to a brokers order system. I can also include charts, graphs, statistics exactly the way I want, which might be constrained in a off the shelf application.
I attest to it. As a rule of thumb, anyone starting out in this business will take 3-5 years before he/she becomes consistently profitable. Doesn't matter whether he/she went to yale or mit or community college.
Although I don't think you can put an exact figure on it, I agree it takes a few years at least, just to get up to speed on the whole area of trading and to go down the many dead ends. I think this is where the majority of traders give up, they see its not going to happen in a few months or a year or two and give up.
around 12 years for me.....before becoming profitable on a consistent basis and being able to determine whether I have an edge or fools good and developing a methodology to trade with the trend, limit my risk, and run my winners
Exactly. The "Holy Grail" is quite elusive in the financial markets. The best of the best traders have keen intuition and insight...and don't really need software. Read the story on Michael Farmer of Red Kite Metals fame. No software in the world could have accomplished what he had done. He knew he was right...lost a ton of money...and stuck it out till the end. Software trading is only for high frequency situations.
How do these "best" traders view charts/quotes or place orders without software? Also there are many traders with automated trading systems, whether is be partial or complete, making fucktons of cash, which are not considered high frequency.