• Registered: 2020-12-10
  • Posts: 19

Topic: Multi-market testing - FX vs Index/Commodity/Crypto

Hi All

I was just wondering if people consider a multimarket test relevant if it is in a differnt type of market? So if you have an FX strategy that doesn't work with many/any other FX pairs, but does well in Crude or SPX500or BTC, do you consider this to be a relevant multi-market test?  Or do you only include FX pairs with FX strategies when testing multi-market?

  • From: Buenos Aires
  • Registered: 2019-09-30
  • Posts: 44

Re: Multi-market testing - FX vs Index/Commodity/Crypto

JRod78 wrote:

Hi All

I was just wondering if people consider a multimarket test relevant if it is in a differnt type of market? So if you have an FX strategy that doesn't work with many/any other FX pairs, but does well in Crude or SPX500or BTC, do you consider this to be a relevant multi-market test?  Or do you only include FX pairs with FX strategies when testing multi-market?

In my opinion, no. The only way a multi-market test like that would be applicable (in my eyes) is if you are focusing on diversified trend trading where you are looking for a trending condition that is present across all markets. A system like this would probably trade on higher timeframes, D1 or above and using simple trend identification indicators such as SMA's or Donchian breakouts with large look-back periods. In this way you would simply treat other markets as "more data" to increase confidence and sample size as these types of strategies generally have a low trade count for a single market.

I personally do not use multi-market testing even within the FX universe as I find most strategies generated from EA Studio are mean-reverting in nature and are only applicable to the pair they are generated on. For me it is a waste of CPU power. Instead I prefer WFA and Monte Carlo.

My 2 cents!