Grid Trading Potential

As I briefly mentioned in the previous post, Grid Trading can be done a number of ways and can either lead to great success, or complete failure. So with any implementation there is a major point that needs to be considered – when do we pull the plug on the grid, if at all?

Let the first implementation be called the Restricted Grid. In this implementation, the entire grid has a pre-defined stop loss to minimize losses. If the stop loss is reached, meaning the price has trended too far instead of reverting, all open positions are closed at a loss. If the price did revert and the overall PnL is positive, then we can also close all positions and restart the grid.

Here is an example of a Restricted 100pip Grid on EURGBP for 2010-2011. As you can see, there is a positive trending equity curve – great.

Restricted Grid 2010

Restricted Grid 2010

What about the other method – the Unrestricted Grid? In the Unrestricted Grid, the grid is never turned off. This means that orders are re-placed for trades that are closed in profit, and for trades in loss – well they are never closed such that there is no loss ever realized. So how does it look? Below is an example of an Unrestricted 100pip Grid on EURGBP for 2010-2011. It actually has a much healthier equity curve providing consistent returns with no losses.

Unrestricted Grid 2010

Unrestricted Grid 2010

The two implementations in this case however ended up with roughly the same final equity. Well, that is great but obviously a system with no stop losses like the Unrestricted Grid cannot work forever. So, let’s look at that.

Unrestricted Grid 2008

Unrestricted Grid 2008

Above shows the testing results for the Unrestricted Grid over the 2008/2009 GFC period. Don’t let that lovely balance curve fool you, the equity curve is lower… a lot lower. Final Balance is $29,756 but Final Equity is $26,478. That’s a $3278 loss we are carrying around but we have secured $1478 profit.

Well that large loss we are carrying around is no good of course – what happened to “Cut your losses early”… Let’s check out the Restricted Grid implementation over the same period.

Restricted Grid 2008

Restricted Grid 2008

Well, that’s no good either. The Restricted Grid, with a final equity of $22,831 lost $ 2169! That isn’t really a surprise though is it? After all, we are closing the trades in loss and accepting that loss. The Unrestricted Grid however never closes a loss trade – ever. That’s why we are carrying around this loss. Even still, we ended up on top.

So what’s the catch then? Well, in case you missed it, there was a huge dip (20.7%) in equity in the Unrestricted Grid during the GFC as the market trended strongly. As mentioned in the previous post, Grid Trading’s weakness is trending markets.

So, in this simulation, with a $25,000 account, betting only 1 micro-lot (that’s right, just $1000), we took a 20% drop in equity. Had we been using 5 micro-lots, we would have been margin called.

Now we are in a tough spot. By only betting with 1 micro-lot, we probably aren’t going to make enough money to live off. However, by betting with more we would need a much larger account balance, at which case we probably wouldn’t be doing this. We have to find some way to hedge ourselves…. And so the research continues.

Posted in Grid Trading, Research, Strategies and tagged , , .

2 Comments

  1. Thanks for your research, especially with the equity curves. Since what you presented is fully automated, have you ever thought about discretionary methods to predicting whether the instrument traded would be trending-range bound? (e.g. keeping an eye on central banks?)

    I’ve seen many of these do so very well, but once the drawdown comes it just decimates so much progress. I think that’s the main achilles heel of grid trading, that (personally) keeps me away from it.

    From your perspective – would it be something you’d implement as your main trading strategy or something supplementary to an existing strategy?

    Thanks!

    • Hi Buena, thanks for your comment.

      Yes, I have thought about it when researching Trending strategies but it doesn’t align with my goal of complete automation and thus I am not entertaining the idea for any strategies currently, though it is on my to-research list.

      I have considered putting in parameters that represent market sentiment and based off those parameters the strategy may change its bias or confidence in signals. Those parameters should only be tweaked with when there is a significant change in market sentiment or a currency’s perceived value essentially ending up with semi-automated strategies which may not be that bad. I have noticed that introducing such parameters would make back-testing significantly more complex and time-consuming though.

      However, to be honest, I think the combination of long term fundamentals with technical analysis could be very successful – at least for getting into trends early – but you really have to know your fundamentals well. These kind of long term changes in the market usually lead to long term trends (for example: Abenomics) and getting in at the start of the trend is not necessary to make good pips.

      For my Trending strategies, if it can detect and enter into a trend a little late but still get a reasonable ride of out of it, then my job is done.

      For the Grid strategies, I’m currently looking into detecting trends and hedging open positions or having a directional bias. More on this later…

      To answer your last question, I think Grid Trading alone might be sufficient if you had a big enough bank account to absorb the total equity loss of the possible range in prices. Otherwise, a fleet of different strategies that either make money or stay flat is the best, then hopefully as the market changes you’ll always have a strategy that is making money.

Leave a Reply

Your email address will not be published. Required fields are marked *