I would say that they use some inefficencies of the backtester.

For instance if you look at trades 667 to 683 you will see that in something like 10 minutes thay will have many times the same successfull trade as if the price was oscillating during 10 minutes in the 30 pips range covered by the order.
I have not analysed the price in detail during that period (I think it is not worth the time spent) but in my opinion it is very unlikely that the price did actualy that.