Mean reversion based strategy

robster970

Veteren member
Messages
4,567
Likes
1,390
Folks,

I have been messing around during the quiet periods with a mean reverting strategy which is showing some interesting results.

The question I have is whether you take trades in both directions, particularly if there is good evidence of a trend on a higher timeframe.

Answers on a postcard.

Love

robster
 
I've got to assume you mean more than just taking the counter-trend trades on the sole instrument. Or you wouldn't have posted. So is this a pairs strategy? Or are you on the trail of something higher?
 
I've got to assume you mean more than just taking the counter-trend trades on the sole instrument. Or you wouldn't have posted. So is this a pairs strategy? Or are you on the trail of something higher?

It does basically distill down to taking counter-trend trades but what I have managed to do is de-trend trends even when they are not really going in a straight line (i.e. parabolic - usually as the trend fades and starts to lose momentum).

No pairs. Single instrument.
 
"de-trend trends even when they are not really going in a straight line", sounds like you are trying to swing trade...
Yes I trade both sides, "price goes up and down", if this is what you defined as "both directions".
 
Folks,

I have been messing around during the quiet periods with a mean reverting strategy which is showing some interesting results.

The question I have is whether you take trades in both directions, particularly if there is good evidence of a trend on a higher timeframe.

Answers on a postcard.

Love

robster

Are you able to backtest and compare results of trading in both directions, with trading only in the direction of the trend (if there is one)? Do you think if you were able to collate some data on that, it would help you answer the question?
 
OK, I'll try and be a bit more specific instead of you guys trying to guess what is going on in my pea-head.

Take a look at the attachment.

On the left, we have a little bit of basic maths stuff which plots the best estimate of what is happening with the trend. This was $ES_F yesterday around the beginning of the European session (8am to about 10am). OK - a trend. So what.

On the right we have the plot of the price deviations with the trend line removed. It is this part I am interested in. You can see it is a distribution that is slightly skewed towards the left indicating that the most common price deviation is a below the trend line which would mean the most common mean reversion would be a contra-trend trade.

Typically I would trade the big pullback in line with the trend (in fact I did and pulled 10.5pts off 4.5pts initial risk) getting in around 1905.00 and getting out at 1894.50.

What I was asking is whether people typically would take a counter-trend trade if you are using a mean reversion strategy around a short term trend.

Make more sense?
 

Attachments

  • 20160202-ES-R.png
    20160202-ES-R.png
    18.3 KB · Views: 806
Are you able to backtest and compare results of trading in both directions, with trading only in the direction of the trend (if there is one)? Do you think if you were able to collate some data on that, it would help you answer the question?

Not really George. At least not currently. I was trying to get a sense of preferences from here before making a call on what to do next.
 
OK, I'll try and be a bit more specific instead of you guys trying to guess what is going on in my pea-head.

Take a look at the attachment.

On the left, we have a little bit of basic maths stuff which plots the best estimate of what is happening with the trend. This was $ES_F yesterday around the beginning of the European session (8am to about 10am). OK - a trend. So what.

On the right we have the plot of the price deviations with the trend line removed. It is this part I am interested in. You can see it is a distribution that is slightly skewed towards the left indicating that the most common price deviation is a below the trend line which would mean the most common mean reversion would be a contra-trend trade.

Typically I would trade the big pullback in line with the trend (in fact I did and pulled 10.5pts off 4.5pts initial risk) getting in around 1905.00 and getting out at 1894.50.

What I was asking is whether people typically would take a counter-trend trade if you are using a mean reversion strategy around a short term trend.

Make more sense?

So at stages you will be long and short the same instrument at the same time?
But on different time/tick moves?

And looking for prices to return to the norm from the time/ticks

There is full time guy based in London/Asia who works on an average of time.
Only trades one direction.
Does very well out of this, mind you has 3 ex wives and 6 kids so need's to.
Let me know if this what you are looking for.
Got a forum on the F Factory.
 
Last edited:
So at stages you will be long and short the same instrument at the same time?
But on different time/tick moves?

And looking for prices to return to the norm from the time/ticks

There is full time guy based in London/Asia who works on an average of time.
Only trades one direction.
Does very well out of this, mind you has 3 ex wives and 6 kids so need's to.
Let me know if this what you are looking for.
Got a forum on the F Factory.

Yes that's broadly it. Point me at the forum?
 
Rob

Presumably you will always be going counter to the short trend that's gone so far as to merit consideration of a reversal back to the mean of the wider trend/range.

I'd say that the stronger the wider trend the more likely it is that pullbacks don't go much beyond the mean of the trend channel which means that counter trend opportunities will predominate.
 
What I was asking is whether people typically would take a counter-trend trade if you are using a mean reversion strategy around a short term trend.

Make more sense?

There's no reason not to, keeping in mind that a counter-trend trade has less profit potential than a trend trade given that the trend is diagonal. There is also the possibility/probability that price will reverse off the median of the trend channel rather than travel all the way to the opposite extreme.

Db
 
There's no reason not to, keeping in mind that a counter-trend trade has less profit potential than a trend trade given that the trend is diagonal. There is also the possibility/probability that price will reverse off the median of the trend channel rather than travel all the way to the opposite extreme.

Db

Indeed. Generally speaking I would avoid entering unless price had move maybe a couple of standard deviations away. That said your observation regarding the pro-trend based move having a better profit potential is exactly the reason I am interested in the question; Is the principle of mean reversion in itself enough to warrant taking a counter-trend trade?
 
Indeed. Generally speaking I would avoid entering unless price had move maybe a couple of standard deviations away. That said your observation regarding the pro-trend based move having a better profit potential is exactly the reason I am interested in the question; Is the principle of mean reversion in itself enough to warrant taking a counter-trend trade?

Sure. That's pretty much the point of trading mean-reverting instruments in the first place.

Aside from the difficulties so many beginners have trading reversals, the chief problem with mean reversion is knowing what to do as price approaches the median. If it powers its way through, no problem. If it doesn't, one might take the option of exiting the trade and banking the profit, then reconsidering the situation as it stands at the median, whether to take a reversal off the median back in the direction of the trend or attempt a continuation trade if there are signals that price will move on to the opposite side of the channel. If one hasn't investigated these signals, there is also the option of doing nothing unless and until price reaches the opposite side and trading the reversal, which will put him back into the major trend.

Db
 
Top