Re: Questions and Discussion About Multi-Level Trailing Stop

Progster's Multi-Level Trailing Stops Package for TradeStation - Trailing stops that change dynamically!
User avatar
progster
Posts: 948
Joined: Thu Jul 26, 2007 7:20 pm
Contact:

Questions and Discussion About Multi-Level Trailing Stops

Postby progster » Mon Jun 01, 2009 10:11 am

This topic is for questions and comments about Multi-Level Trailing Stops for TradeStation.
User's helping users, CodeForTraders helping users, it's all good!
Please stay on-topic, use an informative and accurate subject line, and respect the editorial policies of this Forum.
Thanks!

kwchau
Posts: 2
Joined: Mon Jun 01, 2009 10:38 am
Contact:

Re: Questions and Discussion About Multi-Level Trailing Stop

Postby kwchau » Mon Jun 01, 2009 10:11 am

I wonder if your MLTS product suffers the same shortcoming as the TS provided strategies : Percent Trailing LX and SX.
The problem with them is that in real time trading, my trade is stopped at real time in accordance with the floor amount dollar and percentage trailing that I have set. Let say it is a long position which is stopped, if the market continue to move up in the same bar and if I re-fresh the strategy after the entry bar, the chart will show a much higher stopped price than the actual traded stopped price. Therefore, the equity curve shown in the performance report is totally overstated. I also believe that the optimization report based on historical data is totally overstated as well.

Appreciate your feedback.

User avatar
progster
Posts: 948
Joined: Thu Jul 26, 2007 7:20 pm
Contact:

Re: Questions and Discussion About Multi-Level Trailing Stops

Postby progster » Tue Jun 02, 2009 1:07 pm

Hi, and thanks for your question.

Did your historical testing have look-inside-bar (LIB) enabled to a fine resolution?

If LIB was not on, then the historical testing results are "on close", which is clearly much different than tick-by-tick actual trading in realtime.

I will be back with more info regarding a comparison between the built-in keywords and CFT's MLTS product.

kwchau
Posts: 2
Joined: Mon Jun 01, 2009 10:38 am
Contact:

Re: Questions and Discussion About Multi-Level Trailing Stops

Postby kwchau » Wed Jun 03, 2009 1:25 am

No, I did not enable the look inside bar feature for back testing because it slows down the optimisation too much.
Do you mean that if the look inside bar feature is enabled, when I re-fresh the strategies during realtime trading, the strategy filled orders will match the realtime filled order? At the moment, it does not match. This is the main reason I switch to fixed profit target from trailing stop in my trading.

User avatar
progster
Posts: 948
Joined: Thu Jul 26, 2007 7:20 pm
Contact:

Re: Questions and Discussion About Multi-Level Trailing Stops

Postby progster » Wed Jun 03, 2009 1:33 pm

Actual fills in the real world are subject to slippage. IOW, the price at which your strategy issues a market order will typically not be the price at which you get a fill in the real world.

However, the price at which your strategy issues a market order will be the price at which you are marked filled in simulation/backtesting.

Therefore, one difference between testing/simulation and the real world is slippage. However, that is not the only possible difference.

Do you mean that if the look inside bar feature is enabled, when I re-fresh the strategies during realtime trading, the strategy filled orders will match the realtime filled order?

It's not as dead-simple as that, but that's the general idea.

There are multiple aspects involved in simulation of fills vs. realtime fills. These would include the LIB setting, the behavior of TS keywords in simulation and live trading, the particular way your code is written, and the "bar path assumptions" that TS uses whenever an OHLC bar is under consideration (as compared to a single tick).

There's been alot written about these subjects in the TradeStation Forums. IMO, the bottom line is, it is possible to trade with trailing stops in TS, and it is possible to model such trading, and you can make the model agree "pretty close" with actual live results (but in general, not exactly). Doing this though, is more complex than just using the SetPctTrailing keyword!

Again, I will be back with some further specific info re: the MLTS product.

User avatar
progster
Posts: 948
Joined: Thu Jul 26, 2007 7:20 pm
Contact:

Re: Questions and Discussion About Multi-Level Trailing Stop

Postby progster » Wed Jun 03, 2009 6:42 pm

CFT's Multi-Level Trailing Stop (MLTS) Package for TradeStation is implemented completely without the use of the SetDollarTrailing and SetPercentTrailing reserved words.

This is by design to ensure that the behavior of those keywords is bypassed.

The MLTS package does all it's own tracking calculations and places normal limit and stop orders. Thus, the package provides "as expected" operation in all cases, meaning that all trailing stop order placement is under EasyLanguage control and will be timed just like any other entry or exit order placed under EasyLanguage control.

You still need to understand though that orders displayed as filled on a TS chart (historical or realtime) are themselves a simulation, and not identical with, orders actually placed in a TS brokerage account.

Slippage still applies, and simulated slippage is not the same as actual slippage. Also, in a historical test, LIB resolution applies. The package provides both non-IOG and IOG versions of the strategies. Thus, properly used, the package provides the best possible MLTS modeling and trading, subject only to the deep fundamental design constraints of the TS charting and automation system itself.


Return to “Multi-Level Trailing Stops”

Who is online

Users browsing this forum: No registered users and 1 guest