You are using an out of date browser. It may not display this or other websites correctly.
You should upgrade or use an alternative browser.
You should upgrade or use an alternative browser.
Stops - do you want them?
- Thread starter mizhael
- Start date
What you have then is a stop implemented out of the market.
In that case you need to look VERY closely at your slippage assumptions in you backtest. Because all stops slip, but stops that are implemented out of the market slip worse.
In general, stops should be implemented in the market unless either
1) the position the stop has to close is too big to handle in a single market order
or
2) the exchange in question exposes the "dark" side of the order book to some traders either for a fee or as part of special market making privileges.
In that case you need to look VERY closely at your slippage assumptions in you backtest. Because all stops slip, but stops that are implemented out of the market slip worse.
In general, stops should be implemented in the market unless either
1) the position the stop has to close is too big to handle in a single market order
or
2) the exchange in question exposes the "dark" side of the order book to some traders either for a fee or as part of special market making privileges.
Quote from mizhael:
HI all,
Lets say I have a diversified multi-markets system.
When I backtested it on 15 years historical data with no stops, it has a Sharpe Ratio of 1.50.
But when I backtested the same system with trailing-stops added in, it has a Sharpe Ratio of 1.51. And I then optimize the trailing-stop percentage number using a exhuastive search, the best Sharpe Ratio came out to be 1.52 ( I used the same trailing-stop for all securities, thus only adding one parameter to the system).
Of course, if I allow different trailing-stop parameters for different securities/markets, I will obtain good Sharpe ratio but then I add a lot more parameters.
I would like to consult with the more experienced people here:
From 0 stop to 1 stop, the Sharpe Ratio went from 1.50 to 1.51, is it worthwhile to add that percentage parameter and optimize it?
What do you think of allowing different percentage parameters for different securities/markets?
Thank you!
Before you start worrying about stops you should ask yourself if a Sharpe Ratio of 1.52 is different from 1.50 from a statistical perspective. Combine that with the fact that most systematic strategies produce Sharpes during actual trading that are much lower than backtested Sharpes (0.2-0.5 lower in my experience) and I think you are not focusing on the right issues.
Stops only work if you are can pick winning trades, they are sort of insurance that if something unexpected happens, you are protected.
If you are not consistently picking winning stocks, stops will loose you money.
Try to practice http://www.turtrades.com (free) historical trading with stops and without, you can fast forward after your order has been executed to see if your stop helped your or hurt you.
If you are not consistently picking winning stocks, stops will loose you money.
Try to practice http://www.turtrades.com (free) historical trading with stops and without, you can fast forward after your order has been executed to see if your stop helped your or hurt you.
Quote from jajuanm2:
Most of the consistently profitable trades don't use stops. Lescor and Don Miller both don't use stops. They manage the trade with size. The also average down aka scale into their trades. In trading what seems right is usually wrong and what seems wrong is usually right.
This seems like a strategy that works until it doesn't just that one time. Pleas correct me if I'm wrong. lescor also has 50% DDs. If this were to happen in the beginning, it would be pretty tough to recover from.
Quote from short&naked:
Pleas correct me if I'm wrong. lescor also has 50% DDs. If this were to happen in the beginning, it would be pretty tough to recover from.
You're wrong. I've never had a drawdown even close to 50%. And I do use stops, just not all the time, and it varies how they are implemented and managed.