Is there an advantage to demo trading really?

There is no slippage in demo trading, which is the real enemy most likely in live trading, so why do this “trial” period?

Since there is no slippage in demo trading whats the point?
If you want to give a strategy a “waiting period” before live trading it, why not just wait X amount of months then backtest on that those months after?

You are probably better off just live trading it on micro lots as a trial period honesty. Atleast you’ll get the slippage.

I still like to use demo as a live forward test to see if is strategy is potentially profitable on x broker where I will eventually be running on a live account at my set risk levels.    I have definitely seen that same strategy results between brokers are wide and varied.

 

After a while If it shows is still profitable on both demo account and an updated back test with slippage then I stage to a live trading account, however the staging process I set my risk to a low value usually around 0.5% and again watch for a while.   I may have a number of other strategies running on same live account so I use MyFXBook custom analysis on the magic numbers to track progress. 

 

If all going well and not getting close to my WC scenario as per my MC analysis then will up it to my full risk as per my calculated levels for my account size which I did in SQ analyser and let it loose while continue to monitor the WC DD levels.   I’d love for an easy way to suck trades back into SQ analyser and watch their performance against my backtest in the MC predict/section

 

So I think demo accounts still have their uses and to me its one of the steps I perform in strategy verification.

 

Up until not so long ago I was a firm believer of not using them and to only test on live account at reduced risk settings, but I learnt the hard way when everything looked good on paper/backtests only to have it fail on live account.  However real money is real money so I realise I need to protect my capital as best as I can so have added the additional step of a using a demo account as part of my verification workflow.

 

As Howard Bandy has said in one of his books something along the lines of no matter what testing you do or don’t do, the real forward test starts when you put it on a live account.  

 

I would prefer it if my broker had some micro accounts so I could test on real money easier, but they don’t, so for the time being I will use this as part of my workflow until I find a better way.

 

Stephen…

I just generate strategies with 2.5 spread with 0 slippage, then add it up to the spread and slippage in MC,

i use 1~4 spread & 0~4 slippage, it useuly standard values with most of the brokers..

 

and yeah i’m also checking the updated data with the same MC tests on my .str projects

 

demo is a manual trade’s best friend,

it got nothing to do with algo trading.., no slippage…, pffff easy…, false results.

If the EA is completely unprofitable on a demo, with much bigger drawdowns than seen during backtesting/development, then you can discard that strategy as curved fitted or somehow wrong in reality, without costing a live account.

 

By contrast, if the EA is profitable on the demo account, it does not mean it is a workable strategy, only a live forward test will show that.

 

So unless you have lots of $1000 accounts to test out the strategies on, I think it makes sense to see if the strategy wipes out a demo account.  I’d be surprised if a strategy that proved unprofitable on a demo account was profitable on a live account, so I see it as another filter in the development process.

 

Maybe others disagree?

yes i disagree 🙂 i have strategy that was losing on demo and is earning on real account.

Threshold wrote:

I disagree in the sense that you have the option to not install, and run, yet another instance of MT4 for several months. You can simply wait, then back test it after X amount of months. So why demo it? Just wait, then backtest it. There’s no advantage to running on demo for 6 months over waiting 6 months then simply backtesting it.

But then by that logic, why not just leave 6 months out of sample, don’t touch it till your done the walk-forward and are at the point you normally would go to demo and then just back test that last oos instead and pretend it’s a live forward test?  =)

yes demo testing imo is useless. I prefer live test with micro lots 0.01.

Backtest (even real tick data) gives you test on market data only, but real forward test gives you trading conditions test – system, latency, slippage etc test.

Go live with micro lots.

It is okey to Demo if you do not have money to run 400 strategies on micro accounts. I wished i had most of them on demo and only the winners  live . with IC markets ECN demo accounts i must say the result is very simular on live and demo strategies i have runned them on both for comparison.

Further more it seems the only way to figure out if strategies is correlated or not is to trade them. Launching strategies live that you think is uncorrelated and then You find out they are not can bring down your live account 40 % before you fixed it.