EURNZD 4hr

Here’s my EURNZD strategy.  Crosses like this I look for larger timeframes like 4hr and daily.  Here I found what I feel is a good 4hr strategy and is based off of .01 lots on MT4.  This is a range and volatility strategy that is fairly complex.  I’d appreciate any feedback or tips.

Thanks,

Todd

I see that whole WFM test took you around 778 seconds. It’s quite fast as for such long term strategy. Can You tell, what method for WFM You use? Is it Minute data , tick or real tick? 

I did this a few days ago so I had to reload the strategy and Apply Settings to it.  If that mirrors the original settings then I tested it with “1 Minute data (slow)”.

I tested it on MT4 and it still showed a nice profit and smooth chart, but not nearly the gain showed in SQ.

I´d always verify and optimize with tick simulation in the end – or real tick data if you have. The differences between 1M-mode and tick simulation or real ticks can be REALLY huge by my findings…

Hi GeekTrader,

I was under the impression that tick data was really only needed for scalping, 5min and below TFs. Where do you get your tick data from?

Todd

Go at least with tick simulation mode, some strategies already fall out completely in this mode. I generate mine in “selected time frame only” mode and then re-test with tick simulation mode, which already eliminates the intra-bar-sensitive ones.

What data source do you use?  I use Dukascopy but I thought that I read where SQ doesn’t support tick data from Dukascopy.

Just using 1M bars in SQ but tick simulation mode afterwards. Thats usually sufficient.

In SQ, M1 data uses just the open price of the bars for testing I believe, so one tick per minute.

Tick simulation in SQ uses the OHLC points of the M1 bar, so four times the number of ticks.  Of course real tick data may be 10 or 100 times as accurate as this, but is more difficult to source and takes much longer to process.

For me, tick simulation has been sufficient to match real tick data in MT4 (loaded using TickStory).  But if your strategies average trade is less than the average size of a M1 bar, then you will need true tick data in SQ I would say. 

Yes, as long as you are not scalping intra-1min-bar, it´s fine. The tick simulation mode, by the way, does not just take OHLC, but interpolates synthetic ticks via fractals within that 1min bar (and of course also randomly hitting the OHLC that really occured), which comes close to how the market behaves.

Thanks for the feedback I will look into this.  This particular strategy uses 4hr bars, so 1m data should suffice.  However, I do have two strategies that use 5min data within a specific timeframe so tick testing might be more useful in those cases.

Thanks again!

Todd

Yes, as long as you are not scalping intra-1min-bar, it´s fine. The tick simulation mode, by the way, does not just take OHLC, but interpolates synthetic ticks via fractals within that 1min bar (and of course also randomly hitting the OHLC that really occured), which comes close to how the market behaves.

I asked Mark this before and SQ does not interpolate in tick simulation mode, it merely uses the four OHLC points. MT4 however does this fractal interpolated ticks generation, so they behave differently in that respect.

I would wonder if it doesn´t, since SQ backtests with tick simulation mode and intra-bar stop/limit orders match down to the pip to MT4 every tick mode backtests which use fractal interpolation. However, when using MT4s control point mode (which does just use M1 OHLC), backtests don´t match. So I can´t imagine SQ doesn´t use interpolated ticks. But may Mark can clarify this please?