take a look at my portfolia curve.

[trend2009]

How do you go about doing this?

use ib scanner and yahoo calendar and social media

 

[Craig66]

use ib to scanner

So you're looking for big pre-market gaps?

 

[trend2009]

How do you backtest a beast of a strategy such as that? Tick data? Daily bars?

daily bars

 

[trend2009]

So you're looking for big pre-market gaps?

Filter out any stock with news

 

[trend2009]

News is hard to play it can go either way though my model expectancy is positive with news but with wild swings which I dislike. I hate loss.

 

[R1234]

daily bars

I've tried backtesting intraday strategies using daily bars in the past. For simple rules the daily bars are OK. But once I introduce some complex rules such as profit taking or stop loss I had to start making some significant assumptions on the day's price path - like which occurred first, the High or the Low? The backtest could yield significantly different results depending on my assumptions. How do you get around that issue?

 

[trend2009]

I've tried backtesting intraday strategies using daily bars in the past. For simple rules the daily bars are OK. But once I introduce some complex rules such as profit taking or stop loss I had to start making some significant assumptions on the day's price path - like which occurred first, the High or the Low? The backtest could yield significantly different results depending on my assumptions. How do you get around that issue?

for up day, i always assume low comes first, then go up; vice verse for down day. that is because I entered at low. I admitted by that assumption, the back-test result is significant different from live trading, but backtest results are positive, that is a green light for me to go live anyway.

 

[fan27]

I developed this strategy in 2007 and have used it since then. it produced three digits return in first several years, and then the performance went down since 2014. I guess that was due to the low market volatility. in the last 12 years till now, the maximum dd is below 2%, the sharpe ratio is around 7. below is the performance chart for this year.
View attachment 213958

Those stats are fantastic if they are real! How much capital do you estimate you could deploy to this strategy without significantly degrading performance?

 

[trend2009]

i never compound, because i feel most comfortable at the current amount. it can handle much much more though.

 

[trend2009]

a little about my background. bachelor and phd from top universities. landed a research scientist position in US government institute, working on big data. tweaked my expertise to finance. luckily hit a strategy that has been working so far, so quitted the job in 2008. In the last 10 years, focusing on developing a new strategy that has capacity in billions and sharpe ratio around 4. I have not completed the task yet, but quite close. if the new project is successful, I will compound to the last bit of its potential.