# Grid Trading: Choosing a currency pair

Firstly, apologies for the lateness of this post! I’ve been working on a side-strategy (a Bollinger based reversion strategy) that was looking quite promising with less effort. More on this later – for now, let’s talk about Grids!

The success of Grid Trading very much depends on the currency pair it is executed on. As we already know, trending markets are bad for grid strategies, and some currency pairs are more likely to trend than others. Typically trending pairs are when there are large interest rate differentials between the currencies, i.e. carry trades… just think AUDJPY and NZDJPY. AUD: 2.50 % JPY: 0.10 % NZD: 2.50 %

In addition to a currency pair’s likelihood of trending, we also need to take account its possible trading range. For example, AUDJPY highest high in the last 20 years was 107.83 and the lowest low was 54.92 – that’s almost 5300 pips.

If we wanted to make to run an Unrestricted Grid that would cover that entire range, we would need to be prepared to hold on to some very large losses. Actually, let’s calculate that. To calculate the open losses, we use the Triangular equation: …where n represents the number of trades, and the number of trades will be determined by the grid size in pips.

So assuming we have to cover the price range of 55.00 to 108.00, that’s 5300 pips, and let’s make it a 100 pip grid, so divide 5300 by 100 to get 53. That means, if we did one trade at 55, and then another at 56, right up until 108 (not inclusive), we could have up to 53 open positions in loss.

So going back to the Triangular equation: n(n+1)*0.5 = 53 * (53 + 1) * 0.5 = 1431 units of loss.

A unit of loss is how much a single trade would be in loss if the price moved 100 pips against it. For my Mini account, a standard lot is 10k. If I bet 10k, 1 pip is equal to 1 pip-cost. For AUDJPY, the pip cost is 1.12.

So that means, for every 10k trade, a 1 pip loss equates to \$1.12AUD loss. So, multiply that by 100 pips, and then by 1431 units of loss … we get \$160,272. Ok, so that is our maximum loss in equity excluding swap rates, though realistically it shouldn’t be that bad as we are doing a hedged grid, and all our opposite trades would have won. So let’s take 53 winning trades into account too. \$160,272 – 53 * 100 * 1.12 = \$154,336.

Wow. I need a huge account to trade just one lot, if I want to be able to cover the huge range covered above for AUDJPY. So what about other currency pairs? Well, I’ve put together this table to show you the difference between them.

The table below shows the pip ranges for various pairs, and the grid sizes for 50, 100 and 200 pip grids. More importantly, it covers the maximum loss for each grid as well (as calculated above for AUDJPY 100pip). I’ve then sorted it descending so we can see which pairs we can run with the lowest investment.

So that’s good. Now we have some profiles about currency pairs and can identify which ones might be more appealing to us given our bank roll.

The next burning question is, which one makes us more money?! This depends on how likely the currency pair is to revert back our grid size. Does it trend a lot or not?

At first I was going to find some clever way to calculate this (using a similar concept to Renko charts or using the Hurst Exponent) but I decided it was easier just to back-test my already implemented strategy since I know it works!

EURGBP backtesting for the last 10 years with a 25,000 AUD account showed me the following results. Grid pips is the variable factor here. Or plotting the the Equity as well So looking at the two charts we can see obviously 25 pips is too low, we got margin called. 50 pips however made the most money (30%) but had a 70% drop in equity at some point. That’s pretty rough. What’s interesting is the larger grids still made a bit of money. Now Profit and Loss in Forex is obviously proportional to the amount you bet. A strategy that only wins 5 pips per trade can make millions if you have larger trade sizes. So, I wonder if we can calculate a ratio to determine if the larger grids can make more money than the smaller grids with less risk?

To calculate, let’s take their profit (Final Equity minus Account Size) and then divide over the Max fall in Equity. No idea what to call this, but I guess it is kind of like a reward/risk thing – let’s just call this RR-Ratio. Plotting that we get: From this we can see that they are not equal. While the 50 pip made the most money, the 225 pip grid had a better ratio. So let’s see if we can make more money on the 225 pip grid with less risk just by increasing the trade size. Where:

Profit = Final Equity – Account Balance
Risk = Account Balance – Max Fall Equity
RR-Ratio = Profit / Risk
Size Ratio = Grid Pips / 50Adjusted Profit = SR * P / P(50_Pips)
Adjusted Risk = SR * R / R(50_Pips)

Calculating this in a fun table And plotting the last three columns We can easily see that there are a number of grid sizes that make more money with less risk (Ratio > 1) compared to the 50 pip grid (225 for example). It’s also kind of concerning that the next grid size, 250, was quite the opposite. This doesn’t seem very robust so if I was to apply this, I would probably go with a number between 150 and 225 (200?) since all these values were better than the 50 pip. Just to confirm my findings, I’ll back-test the larger amounts right now and post the results later.

I think we have made some good progress on this research and I’m getting excited about this strategy. I think we have proved the feasibility of the Grid however the large draw-downs are certainly something we want to avoid. Also, because of the scary draw-down we haven’t been able to make real money. 30% over 10 years is not what I call real money. We need to find some way to hedge ourselves against the draw-down so we can increase our trade size and make more money! Time to start looking into that.

Update: The back-testing for 200 and 250 finished with the increased amount sizes and the values matched the calculated values above – as expected.

 Grid Pips Amount (lots) Calced Final Equity Actual Final Equity Calced Max fall equity Actual Max fall equity 50 1 32505 17508 200 4 33300 33303 17284 17285 250 5 31120 31121 17375 17380
Posted in Backtesting, Forex, Grid Trading, Research, Strategies and tagged , , , .