Backtest with different trades than forward test

I just ran a forward test on a demo account over the last week. Today, I downloaded the price files from my broker so that I had the latest set of data for testing in zorro.

When I ran a backtest over the last week's data, I would have expected this to tie up with the forward test results I had just seen. However while it was largely the same, I noticed a trade that was taken in the forward test was not generated in the backtest.

Any reason why?

What is a forward test? Do you mean a trading session?

Having developed an algo that had generated good results when testing, I wanted to undertake an addional check before going live with real money. So, I connected to a demo account at the start of last week and just let the algo run.

This allowed me to see the actual trades that would be generated on a new set of data, and to see if the performance was similar to that generated when testing.

As to the why, you should normally see in the log why a trade was triggered or not, and what was different, like price, time, or an accumulative indicator.

Thought I would offer my 2 cents on possible difference between backtest and paper trading.

1. Bars vs. ticks in backtest. If you backtest off bars expect difference to paper trading off live market real time stack.

2. Limit vs Market orders. If your algo uses limit orders these can be less likely to execute in livepaper tradng than market orders.

3. Big size. The full size you are trying to trade in backtest may not be there in livepaper market.

4. In backtest your trading does not affect the market. But in live/demo your trades are part of the market and will change what is happening and what other people are doing.

All this is called slippage. So paper/demo trading does not add up to backtest most of the time.