Curious if anyone here knows how MC replay works or has tons of experience. Programming a strategy that relies on renkos and uses limit entries. I know backtesting is worthless...that said I was assuming that tick by tick market replay would be a useful backtesting tool; however, my replays in MC have come very close to backtest. I then watched tick by tick and upon bar completion the limit order is 'triggered' when it otherwise shouldnt have.

Anyone have experience with this in MC?

Thanks

greenroomhoo 

Programming a strategy that relies on renkos and uses limit entries. I know backtesting is worthless...that said I was assuming that tick by tick market replay would be a useful backtesting tool; however, my replays in MC have come very close to backtest.

@greenroomhoo,

Your replay matching backtest isn't a coincidence -- it's by design. MultiCharts uses the same fill engine for both replay and backtest on Renko charts. There's no differentiation in order generation between Renko and regular bars in either mode, so replay is basically re-running the backtest with the same logic.

Here's the core problem with your limit orders: when a Renko brick completes, the close is always exactly at the brick boundary. The fill engine evaluates your pending limit against that completed bar's price range. If your limit price falls within the range, it fills. But the actual tick that completed the brick may have already blown through your limit level -- meaning in a live environment, that fill likely never happens. The bar-completion trigger is creating phantom fills.

MC's own documentation actually flags this directly -- they state that resting orders (limits and stops) on Renko are "much more complicated and definitely not recommended" and suggest using on-close market orders as the reliable approach.

A few workarounds worth exploring:

• Bid/Ask data series -- Add secondary data series and reference BarsOfData(2).Close for Ask, BarsOfData(3).Close for Bid. Gives you more realistic fill simulation than last-trade-based Renko bricks.
• Tick buffer on limits -- Offset your limit entry by 1-2 ticks from the theoretical level. Won't eliminate the problem but reduces the phantom fill rate.
• FlexRenko with 1-7 second resolution -- Higher granularity reduces the gap between simulated and actual fill behavior.
• Convert to market-on-close logic -- If your edge survives the slippage, this is the most honest way to test it.

One more thing to watch: if you change your data start date, the entire Renko chart redraws from a new anchor point. Different starting tick = different brick boundaries = potentially different trades. That alone makes Renko backtest results fragile regardless of fill logic.

Bottom line -- your instinct that backtesting Renko is worthless isn't far off. Replay doesn't fix it because it's running the same flawed fill simulation. The real test is forward-testing with live data and watching how your limits actually get hit.

Feel free to tag @Fi if you want to dig deeper into any of this.

-- Fi
"The backtest that confirms your edge is the one you should trust least -- the one that challenges it teaches you the most."