first of all I find it great that NT has Renko Charts in version 7 ( I use beta21). But I guess I´ve seen 2 bugs which result in false backtest results when using NinjaScripts on them:
1. Chart from 06/04/2010:
When there is a change in direction of the Renko Chart (f.e. down brick to up brick) then the OpenPrice of the new brick is wrong in my opinion. It should be equal to the ClosePrice of the previous brick. Of course this would generate a brick with 2x bricksize but at the moment NinjaScript uses a wrong EntryPrice for a new position at this brick.
In the chart: In my opinion the OpenPrice of the 14:59-brick should be 1,5146. The Sell-Signal is generated at the end of the last downbrick (14:46 @ 1,5146) when SMA-Cross has happenend but the fill is at 1,5151.
2. Chart from 29/03/2010:
For past dates NT generates a "closing brick" (without body) at the end of a session. This does not happen on the current date (in real trading). On the current day the open of the closing tick is equal to the close of the previous brick (which should be right) and the close is the last price before SessionEnd. This results in a closing brick which is smaller than the regular bricksize. In my opinion this behaviour is right and should be applied also for past dates.
In the chart: The open of the "closing brick" (and the fill of the Buy Signal) should be at the same price level than the close of the previous brick (~ 4 pips higher). The signal is generated at the close of the green brick before and should be filled imediately at the same price.
To change Renko BarType in this may should solve 2 problems:
- Differences between backtesting and realtime behaviour (at session end) &
- wrong results for backtesting (at direction changes & session end)
Thanks a lot.
Joerg

Comment