Last week, a reader known as Matt sent me a comment pointing me to a Grid on MQL5 that might be of interest. I thought it would be fun if we did this analysis together so I’m going to bump up this post.
So let’s see, the Arin grid can be found here: http://www.mql5.com/en/signals/35388
Its equity profile looks very good, as seen in the featured image above. So, it obviously makes money – check. That’s good. However, immediately some alarm bells are going off – 6000% growth!! That’s just too good to be true and simply not sustainable.
So, how long has this grid been around? Ah, only one month.
That should be the first caution sign. Strategies need to be run for at least 6 months (perhaps longer) before you can gain any confidence in them. If we only run them for short periods of time, we do not expose our strategy to enough market conditions to know how it will behave when the market turns heavily against us.
However, it doesn’t mean it isn’t a good strategy, just we have to be careful.
Let’s dig deeper. The next point of interest is this large equity drop around the 17th. The balance dropped from about 1100 to 730 … that’s 33% of the account or 1850% of the initial deposit – seems pretty risky. Also, according to the MQL5 stats on the side, its max equity drawdown is 79%!! Actually, if you look there are many places where the account would have gone bust, had the start of the strategy not been timed just right.
Let’s try to figure out what is actually doing. Let’s pick a point where in the equity curve that is landmark, like the equity drop on the 17th followed by a huge jump in profit. Let’s find the trade history (large profit trades) for that and take a look.
So locating the large profit trades on the 17th, we see the following in the trade history. We can usually group together trades in the same Grid instance by looking at the times they are closed.
Doing some analysis
– Let’s calculate at the change in price in pips (absolute value) – called above Pips Delta
– Let’s calculate the difference in entry prices for trades in sequence – called above Entry Delta
We can immediately see
- it is a 3 pip Grid with some kind of custom incremental trade size sequence (at least I do not recognize that pattern)
- the last entry was 6 pips away while the rest were 3 pips.
- the final profit trade took profit after 10 pips. This may be significant, because we need to determine what the profit target the system uses
- all open trades had their trade profit updated as well.
But, is this a single or bi-directional grid? Well, going back to the Trade History – what other trades were done in the above Grid’s lifetime … i.e. from the time the first trade opened @ 01:02:15 to the time the grid closed @ 15:30:01… Well it turns out there were a lot of trades, the below is a sample.
So, now that the trades are sorted by open time, we can see that we are indeed trading in both directions. However, there are far more buy trades than we have sell trades. It looks like we actually have two separate directional grids in play here, one for long positions and one for short positions. Additionally, the entry prices do not appear to be specific round number, so we are not sure how the strategy decides when to enter or start a grid instance.
In this particular period we are examining, the long-grid is working nicely, indicating the price was going up. Meanwhile, the opposite short-grid kept re-entering every 3 pips with the increased amount size.
Let’s look at the price chart data for this period as well to confirm this (note: my timezone is not the same as the trade history but I matched the data).
Yep, we can see that we were right. The price kept going up, hence the long positions doing very well.
Actually, the long trades are interesting. By looking at the trade history, the long trades usually had a 2 pip profit target. Sometimes, it was around 4 pips, sometimes 6. It looks like the more times it had to re-enter, the further away the profit target. This is interesting, as not only are we increasing the trade size but also the profit target. However, it isn’t as simple as just doubling it I think as because the final short trade after about 11 re-entries only had a 10 pip profit target….
While the buy trades kept entering even as the price reached 1.38611, the short trades peaked at 1.38545. Even though the short-grid was a 3 pip grid (and the final one 6-pips), and the price moved at least another 10 pips after it peaked, there were no more short trades entered. This is interesting … If the last entry was after 6 pips, then perhaps the next one needed 12 pips. Or perhaps we have reached some limit and there will be no more re-entries. Either way, the behavior clearly changes once the price moves against us a certain amount.
Another interesting observation from the price is the Grid instances started when the market was very quiet. Perhaps this is the trigger to start the Grids? Actually, if we look at the history of April, there were many periods where the price ran in a single direction. However, the strategy only seemed to suffer one large draw down. It seemed it managed to not trade during the other times… how did it do that?
In the end, lucky for this user, the price did retrace and the grid closed out in profit. However, the price only moved about 45 pips in its extremes, and we had already suffered a 33% drawdown … what if the price had moved more than 45 pips (which happens often).
So this has been an interesting Grid to analyze, though there are many unanswered questions like
- What is the relationship between the increasing amount size and the spaced out profit targets?
- What is the story with the space out entries after 10 or so re-entries…?
- What is the sequence used to determine trade sizes?
- What is the trigger to start trading/start the grids when we have no open positions? How does it avoid the periods where it would have lost big-time?
- What is the trigger for the winning directional grid when the other has outstanding losing positions?
There is a lot to think about here, and more research required on this grid to try to answer the questions above. I’ll try to keep an eye on it to see if anything else comes to mind. Without knowing the answers to some of the above, I would not trade this grid. It has huge potential to wipe itself out, so I am very interested in finding out how it decides when to start trading. If anyone has any ideas, please let me know!
Anyway, thanks to Matt for sending me the link to this Grid. Remember, comments and suggestions always welcome.