Back test and live different

I wrote a robot. The back test has an excellent result, but it is very bad in live. Where is the problem?

@FXEZ

can you help me or other people?

You have to be aware that past performance is no guarantee of future results, especially during times of market uncertainty. Also make sure to simulate the commissions and fees charged by your broker when backtesting. Don’t forget to use quality historical trading data - this will bring your trading results closer to reality (live-trading).

1 Like

overfitting, bad quality data, commissions and fees, to short backtest period, change market, backtest quality

1 Like

in back test the graph is very good.but in live ,is very bad graph.
always sl
not go to tp

but in back test always tp.
i don`t now why??

my broker is alpary.

where can i use the obos indicator code??

I tried to work with different robots, but to be honest, I got the best result by working manually. Alas, this is often the case.

My robot does not take the whole process. I want to use the obos indicator. I need the code.but i didn`t find it

From what I’ve read it’s best to test it on different assets to understand where it works most accurately, robots are not usually durable, so it could be that it no longer functions fully…
In general, I wish you good luck in figuring it out.

What is the strategy? How big is SL and TP? What time frame?

Back testing on very short time frames doesn’t work well, the back test doesn’t simulate every tick even though that’s what it suggests it does.

Did you use the optimize function to calculate things? This is a bad strategy because you’re perfectly fitting your strategy to work on past data, but that doesn’t mean it works in the future. If you want to use that, the best chance is to optimize over a short period and then back test over a much larger period, that will tell you if the parameters are likely to work over a long time.

simply because someone is watching your SL and hit it purposely,that’s why forex is hard

There are a lot of spikes in the market that are caused by news - as far as I know a robot cannot control this, even if it is an excellent system in terms of quality.