In this post we present the finding of another example of a good automated trading. The method is that of grid trading, on a grid of levels 50 pips apart. There is no trend detection. Only trend-neutral grid trading. Grid trading is essentially the trading the the market makers and liquidity providers utilize in fast frequency, and small grid-levels around the bid and ask.
The market behavior behind such a trading system is double
a) The histogram of draw-down and draw-ups in a chart of prices is a Pareto or Power distribution and not a normal distribution
( It is a quite well known property and has been proved by many researchers. For example see this one here (from Scotland )
and also this scientist here who is proving this distribution for all the companies, and.......cities. See also post 25.
b) The smaller the time scale in the markets the longer the time intervals that the market is stationary with trend-less neutral volatility (see also post 12 the inter-scale law of the markets)
This grid trading algorithm is not based on any type of pattern like continuations or reversal, or channels, or on any pattern recognition method or system of indicators, or candle stick patterns etc. That is why it is simple in its programming logic, and the simpler the algorithm the more robust and stable its performance. It is based on the simple statistical assumption that there are horizontal equilibrium levels of the prices (both support and resistances) , around which the markets fluctuate with a constant (random variable) distribution.
The particular grid trader we present here , applies group-take profit on trades (not individual trades take profit). It applies pyramiding and escalations at 50-pips levels of a grid , which are the optimal trading for trends or micro-trends that have a histogram which follows the Pareto distribution (post 25). It has a timer of acting every of 15 minutes,
The grid-trader is of less profitability compared to the other examples of automated trading. E.g. here it is about 18% annual at 23% 5-year maximum draw-down. In other words for a 30% 5-year maximum (floating) draw-down the average monthly rate of return is about 2%-2.5%. This is as far as I know the score of the most successful grid-traders in the web too. This performance of 2.5% monthly , or 25% annual , at at a 5-year max Draw-down of 30% is low compared to other automated non-grid trading systems (e.g. those n posts 50, 51, 51 etc) that may go to 10%, 15% even 20% monthly. The difference here is that most of the automated systems have a forward life-cycle (that is , forward duration that are profitable) for 1 at most 2 years. And at the end of their lives they may lose you considerable part (some times all) of the hither too profits. BUT the grid traders have a much longer life-span that may go far beyond 5 years. In addition the have a uniform flow of profits so that every month is profitable. This compensates for their lower profitability performance. In particular it is seen from the detailed statements below, that EVERY month is profitable, in this grid-trader!
But the advantage is the statistical stability of it which is based on definitely verifiable statistical properties of the markets as above a) and b).
But the advantage is the statistical stability of it which is based on definitely verifiable statistical properties of the markets as above a) and b).
Detailed report of the 5-years back-test 2009-2014 is here
https://dl.dropboxusercontent.com/u/107295772/Old_hedge_weaver_likerealestate_grid500_FBack_spraed12_alparioptimal_m15open_2009_2014.htm
And an analysis by the MT4i or Fx-blue site here
http://www.fxblue.com/users/neutral_grid_trader/stats
