Search
× Search
Tuesday, November 19, 2024

Archived Discussions

Recent member discussions

The Algorithmic Traders' Association prides itself on providing a forum for the publication and dissemination of its members' white papers, research, reflections, works in progress, and other contributions. Please Note that archive searches and some of our members' publications are reserved for members only, so please log in or sign up to gain the most from our members' contributions.

Stop function ideas

photo

 Kirill Pankratiev, CEO and founder of Rumine Asset Management

 Thursday, February 26, 2015

Dear members, I was wondering if anybody was eager to discuss the subject of the initial stop and subsequent stop function in the context of a trend following strategy. I will start first; we assume that occasionally a market develops a full Elliot sequence which could be predicted as early as wave 2. At that point we assume that it is also possible to project a target for wave 5 and therefore to decide if a position could be initiated. Two of the key success factors of such an approach would be: 1) the initial stop price which determines your position size and 2) the ensuing stop function So far we have been quite successful with an RSI(14)-implied envelope [40-delta . .60+delta]. Explanation: given relative strength average loss 1 bar ago (numRSAvgLoss[1]), given relative strength average gain 1 bar ago (numRSAvgGain [1]), given the close 1 bar ago (Close[1]), The corresponding envelope in price points for a fixed RSI value can be calculated as follows: numLowerExitBand = Close[1] + 13*( numRSAvgLoss[1] - numRSAvgGain[1]*MODEL_STOP_CONSTANT ) numHigherExitBand = Close[1] + 13*( MODEL_STOP_CONSTANT*numRSAvgLoss[1] - numRSAvgGain[1] ) where MODEL_STOP_CONSTANT=1.50 for RSI values 40 and 60 (but could be anything else) If one assumes that price is in a strong uptrend it should not break the lower band even during retracements and vice versa. A possible derivative which yields slightly better results is to ‘trail’ the stop price (ie stop(i) = max( stop(i-1), stop_function(i) )). Any comments suggestions or completely fresh ideas are very welcome Thanks for reading!


Print

3 comments on article "Stop function ideas"

photo

 Salah Elmorry, Founding Partner at Systemathics

 Friday, February 27, 2015



Hi Kirill,

Maybe you need to do things separately.

- First you generate a trend following signal

- Then you setup an execution tactic (using smart orders for example) : send an arbitrage order, hedge with both a Stop and Trailing Stop order for the stop loss and take profit.

The limit prices of both the stop and trailing orders depend on the volatility and the expected profit or loss.

Hope this is helpful.

Salah


photo

 Johan Kretz, Senior Investment Manager at Deutsche Bank

 Friday, March 20, 2015



Hi Kirill,

By using a trend based indicator, as you suggest, would a SL not be unnecessary?

I certainly would advocate a SL strategy if the trading signal was based on valuation or an exogenous series like macro series, since these are not derived from the price and correlations as well as causality with the price movement in the security that you have invested in my not be stable.

But if the indicator is based on trends in the price of the security that you are trading, surely the trade should be kept until the indicator "turns"? '

In my trend based (autocorrelated) models, I have no other SL than the signals from the model it self. Hence, I keep a long position as long as the trend that originated the position is positive.

Cheers,

Johan


photo

 Kirill Pankratiev, CEO and founder of Rumine Asset Management

 Sunday, March 22, 2015



Johan, we have tried a similar approach but the whipsawing outweighs all the benefits or at least it is unconclusive.

In our approach we are not 100% in the market but rather 2-5% in a given market but we tarde 350 markets at the same time. The rationale is that you can have only so many instances when a market presents to you an acute opportunity where you can apply a model to describe future behaviour with somewhat precision. In our case it is a model of the Elliot wave. So our research has shown that yes you need to have a physical for many reasons both tehcnical and conceptual. Now the question remains what is the best stop loss because it determines your leverage and you d rather be as leverage as possible on a given winning trade.

If you assume that you try to jump in a possibly new trend at the beginning of Elliot wave 3 ( which is the end of retracement from wave 1 ) then in our example the best that we found empiricaly and statistically is to start with Max(RSI-Implied-Stop, Min(Wave2, TargetWave2))

where RSI-implied-stop: is the price which would equal RSI to 40

Wave2: is the lowest closing price of the retracement

TargetWave2 : is the .618 retracement off (wave1-wave0)

thereafter you apply the same approach as desribed in my original post.

Before we came with this conclusion god know how many things we have tried including what is called a "statistical" stop which was desribed in scietific publication. Nothing worked better and the reason is that because no other approach allowed for mild retracements because all other approached tried to squeeze the price out as much as possible and therefore very often passed up on very large trades.

Thanks for you comments and thanks for reading mine

Please login or register to post comments.

TRADING FUTURES AND OPTIONS INVOLVES SUBSTANTIAL RISK OF LOSS AND IS NOT SUITABLE FOR ALL INVESTORS
Terms Of UsePrivacy StatementCopyright 2018 Algorithmic Traders Association