Mar 222011
 

There is an interesting discussion about stops & targets going on in the forums.  I will copy what I wrote there but I invite you to read the thread and give your own input.

…why are stops & targets based on a set number of x ticks without taking into account context?

I have studied my trades and I’ve found a few things:

1 – The best trades do not go against me at all. Here is an example from yesterday.

2 – If it goes against me by 4 ticks then it’s likely to go against me 8-10 ticks. Why? Because if we believe the average stop in ES is 8 ticks, which I do believe simply because the average rotation in ES is around 10 ticks which is just enough to stop out a lot of traders.. so if it goes against me 4, all the other traders who entered when I do will stop out at -8 ticks and then if there isn’t size on the order book there then the market will run another 4 ticks or so as they stop out and flood the time & sales. So now we’re looking at -8 or -12. Interesting how the OP came up with 13 ticks as that’s just outside my estimate of 8-12.

The problem with a 13 tick stop is that you’re in a bad position. A rotation in ES is on average 10 ticks, so it will take 2 rotations for you to scratch. That’s not good odds just to scratch a trade.

My stop varies from 3 ticks to 8 ticks. It depends on the behavior around my entry. I enter at levels so the logical stop for me is the high/low just before I entered. Either the level holds or it doesn’t. I try my best to get this to be 4 ticks. The reason is if I stop out at 4 ticks that’s $50. If I need to, I can re-enter and risk another 4 ticks and still come out risking less than the 8 ticks.

Go back over your charts and see how many times a level holds to the tick. Why give it 13? Either it holds or it doesn’t. If the level doesn’t hold then my reason for entering the trade is gone and I want to be out.

That was about stops, targets are similar. I usually go for 6 ticks to scale half as I know the average rotation is 8 ticks (on both ES & Bund). I never enter to the tick at the high/low so I figure there are 6 ticks left. That will cover my risk. Then my next target is at the next level, usually a CHVN.

I’m debating about having a first target of 6 ticks. Today I got a 14 tick move on the bund, so I got +6 +14. My hypothesis was for the bund to reach 121.91 which was 16 ticks. I front run that 2 ticks. So why take +6 on the first half? The theory is because that covers my risk and the average rotation is 8 ticks so I can’t count on more than 8. But I’m not sure. When I take half off at +6 then it’s like I’m betting against my hypothesis!

When consistent with 2 I will add a 3rd and will be able to capture more of a move. I feel the becoming profitable with 1 contract is extremely hard. With 2 it’s still pretty darn hard. But if one can do it with 2 then they should make even more as they add units.

That pretty much sums up my thoughts. Most importantly this is all a personal choice and everyone must discover what works for them. This is a trial & error process and it takes quite a while. But when you discover what works for you then you own it and it’s yours.

PS: The MAE of the first trade says 0, I believe that is a ninja bug and I’ve reported it in their support forums. I checked the rest of the MAE’s and trades 3&4 should be 3 tick MAE, the rest are correct – they really were only 1 tick which is pretty amazing.

And another post about scaling out of losing trades:

This is an interesting theory. I do not feel it is a good match for the way I trade and here’s why: One of the criteria for my entry is, for example if I want to get long, I want to see support on the order book. So if that support on the order book disappears, then my reason for entering there is no longer valid. So I want to be out of my trade and look for another entry. So for my method of entering, it wouldn’t make sense for me to scale out because that would leave me in a trade without a valid reason for being in the trade!

I think if one uses static x ticks stop/target, including scales, then it’s attempting to apply a “one size fits all” solution. I don’t know about you but whenever I see a hat or shirt that is “unisize” or “one size fits all” it never fits me properly.

Instead, the stop & target should be unique for each situation. That can still be studied an optimized. For example,

let’s say I enter on a HH on a 1min chart. I can put my stop under the previous bar’s low. I can put it 1 tick under, or 2, or 4, etc. So I can do a study on what is the optimal number of ticks below the bar. But the total stop size will be variable.

Compare that to say 4 ticks. If the previous bar is 6 ticks tall then my stop is not away from the action and is not safe. Say the bar is 4 ticks tall and my stop is 8 ticks. That could be too far.

I’m not saying any of the ideas in this thread won’t work, I’m just saying I don’t believe they work for me and presenting another point of view. If dynamic stop loss works for someone then I’m happy for them that they found what works for them. Everyone has to find what works for them.

  2 Responses to “Thoughts on stops & targets”

  1. Hi,

    “average rotation in ES is around 10 ticks”, did you check it historically? I keep hearing different numbers all the time but couldnt find any study on it.

    Thx.

  2. ValYouW – That’s a question I’ve asked myself as well. I actually did measure this on the bund. The way I did it was to use the fractal indicator and I exported all the turning points to excel and threw out the very small & very large values. I ended up with 8 ticks. I was expecting it to be higher than that since Bund is more volatile.

    I didn’t do the exercise for ES but I plan to.

 Leave a Reply

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>

(required)

(required)