Positive Slippage Ideas

[MarkBrown]

how to take a negative and make it a positive by breaking from typical accepted theory that you must experience slippage on a trade.

say a trader expects to get a fill on his position, he is hoping to get a price that he sees or that his model says he should get. when he gets that fill back and it's better than he expected by enough to more than cover all cost why should he use any negative cost in his testing?

1st image). trading in single lot mode - in the first image is a system using slippage and commissions "only" for the satisfaction that some "believe" they can not be avoided. i call this the original order - it's what i expect worst case - it's what you expect.

2nd image). trading in single lot mode with up to 2 better additional entries enabled - free money trades no commission no slippage only additional profits i would not have otherwise received. normally you would need of course the margin to add on additional trades. but with a couple of hundred thousand extra dollars per market i suspect that will not be a problem as the model will pay it's way.

3rd image). trading in single lot mode with 2 better additional entries with fixed ration enabled - note that the risk is never more than the original risk would have been in the first place. when we flip on "fixed ratio" the little 100k system now makes 1.5+ million which still includes slippage and commission on the initial contract but not all the free trades i didn't expect to take with the original order. but i could add the cost it really would not matter still if you have a model which in capable of "Positive Slippage"!

mark brown

 

[Fractals 'R Us]

Tune me in here: You are talking about price improvement on entries? Sure, you can slow your strategy so that there is noise after your signal then use a limit order for entry. The upside is a better entry price. The downside is you slowed your strategy and you don't always get an entry.

 

[TheMagican]

Wtf is this!Happy New Year!!!

 

[MarkBrown]

i explain above but to clarify. i make a model that that identifies a trend a real long trend and then i wait for moves against the trend and enter.

example: long term trend is up it rips down one day i am a buyer at mid day and eventually it goes back my way. but before it does it continues to rip lower and so given an opportunity to buy better i add onto my position.

my original position is what i expected to get, the add on positions are free less the margin i have to put up. takes some statistical confidence i know but it works. once you feel confident you can do this then you implement things like fixed ratio etc.

m

 

[traitor786]

Quote from MarkBrown:

i explain above but to clarify. i make a model that that identifies a trend a real long trend and then i wait for moves against the trend and enter.

example: long term trend is up it rips down one day i am a buyer at mid day and eventually it goes back my way. but before it does it continues to rip lower and so given an opportunity to buy better i add onto my position.

my original position is what i expected to get, the add on positions are free less the margin i have to put up. takes some statistical confidence i know but it works. once you feel confident you can do this then you implement things like fixed ratio etc.

m

Good stuff. You actually made it on my none disproved ideas list.

Its nice to see some work being done ! I wish I could help you in any way, But it is not my field of expertise. don't hesitate to ask though, I'm curious about how you can measure slippage in your example. slippage I would think is only true after an actual trade is placed.

 

[maler]

I would not call this slippage.
Simply a better model, long term persistence enhanced with
short term mean reversion.
For me slippage is the pound of flesh you feed the market infrastructure.
What does it cost to enter and immediately unwind the same position.
It is never positive unless you are part of the infrastructure,
that is you have customers to skim off of.

 

[Traveler]

He's talking about something like a long term trend system that buys a 200 day MA or high/low, but then he puts on a 1/3 position for the initial signal, then adds 1/3 and another 1/3 as it goes against him. If it keeps going against him, he gets stopped out on all 3 but for the price of the risk of a full position entered all at once on the signal.

Mark, what happens if the market breaks out and never looks back? Do you eventually keep adding?

 

[MarkBrown]

Quote from Traveler:


Mark, what happens if the market breaks out and never looks back? Do you eventually keep adding?

there is more than just a moving average and buying positions actually that is like scale trading i think your thinking about.

this is a little diffrent you can see the red box where the 200 ma is now in a down trend. the system now rethinks the position and starts selling on little bump ups. it is limited to only adding 2 additional positions after the 1st one. so there is no danger of getting hung for too long. see the positions are in and out pretty quick then looking to repeat or reverse.