I think that holding overnight is feared by many when it shouldn’t be. I trade using spread betting rather than CFD and short trading can be expensive on indices as you have to pay any dividends back but when long you get paid the dividends which helps considerably – even if technically the price should fall in line with dividend payments…. and I don’t short indices anyway as you are fighting an unbeatable long term trend.

Quick question… is it your own momentum indicator that you use or a well known one?

If you are making more profit then you are paying out in fees then it is fine to be all in all the time. You just have to be very confident in your strategy. The opposite is that every minute that you are NOT in the market is an extra minute where your risk is zero. The best way to never lose money is to never put it on the markets!

If we think about it then every time you place a bet then it has 50% chance of winning and 50% chance of losing. We can adjust this % up and down by moving stop loss and take profit levels but the odds never change in reality. If we then have to pay fees then we can see that we have less % chance of winning and more % chance of losing. So all the time we are just looking for that one idea or one indicator or combination of indicators that will give us a few % edge to swing the odds slightly further our way so as to overcome the fees. So if you are confident that your strategy has a big enough edge to overcome the imbalance then it really does not matter how long you are in the market – what matters is the size of your ‘edge’.

The dream we all have:

One holy grail strategy that opens positions often that all win and that return 1000 pips every time for a £1 bet.

The reality:

Many diversified systems running on many different markets and different time frames that all together combined hopefully make slightly more than they lose once trading fees have been taken off.

We need to forget the dream and concentrate on the reality as un-sexy as it is.

I am a big fan of momentum based strategies and have few of my own under test at the moment. If you fancy sharing your strategy then maybe someone on here might be able to add an even better edge to it. The results you show look very positive for such a long time period but it sounds like your method of construction is spot on (in sample then out of sample) but maybe another set of eyes will spot some over-fitting that might lead to future failure.

Personally I have so far found that the simpler a strategy is the better it is. If you pile indicator on top of indicator and have variable after variable then at each step you adding more and more curve fitting. KISS is best. If you have to throw everything at a strategy to make it work then it is best to put it in the bin and start again. All you want to be doing is to look for a market condition to be occurring, check that you are not going against trend and use sound money management and then all should be fine. If you have to check that you meet 1ooo different conditions then not only will you find few trades but they are likely to be based on a curve fitted idea.

I find it best to write a seemingly working strategy and then start taking things out to see what true value they add. If it adds little value then leave it out. It is also a good idea to run an optimization with one variable changing and then plot a graph in excel of the result to check out the working optimized range and if all looks OK then select a mid table number for the variable to reduce the effect of curve fitting.

Sorry if I’ve gone on a bit – but sometimes the discussion of theory can be more fun and educational for me than the sitting down and testing code!

Despair wrote:

If your strategies hold many days it might be worth trying to run them on futures instead. This way you have no overnight fees.

But as I have found there is no historical data if using spread betting futures (on my IG account at least) so you are restricted to CFD and all the tax implications (for UK) users that that involves. I’ve always found it pretty difficult to work out which is best DFB’s or Futures and normally find that if I open a trade on a DFB contract expecting that it will close within the day that it defies me and runs for several days and that if I open a bet on a futures contract it closes within the same day – I think they call it sod’s law. I also think that I read somewhere that if you expect to hold a position for over five days then normally Future contracts are better than DFB’s but then if I knew how long a position was going to be open for in advance I’d be a millionaire!