Tester visual mode lies about speed
Posted: Tue Nov 03, 2026 5:37 am
Tester visual mode lies about speed
The visual mode of a strategy tester shows trades unfolding on a chart, which is useful for debugging. It also gives a false sense of how the strategy behaves in real time. The playback speed, the absence of latency, and the lack of other market participants make it feel smoother than reality.
In the tester, orders fill instantly and indicators update perfectly. Real conditions include delays between tick arrival and indicator calculation, order transmission time, and queueing.
Watching the visual mode can also induce overconfidence. Seeing neat entries and exits repeated many times feels convincing, but it is a sample of curated stories rather than statistics.
Use visual mode for verifying logic, not for judging performance. After debugging, rely on numerical metrics, distributions, and out-of-sample results.
Also compare the tester's timing with live behavior. Running the same strategy on a demo account for a few days and comparing trades with the backtest for the same period reveals discrepancies caused by latency and data differences.
How closely do demo runs match the backtest for the same period and settings?
The visual mode of a strategy tester shows trades unfolding on a chart, which is useful for debugging. It also gives a false sense of how the strategy behaves in real time. The playback speed, the absence of latency, and the lack of other market participants make it feel smoother than reality.
In the tester, orders fill instantly and indicators update perfectly. Real conditions include delays between tick arrival and indicator calculation, order transmission time, and queueing.
Watching the visual mode can also induce overconfidence. Seeing neat entries and exits repeated many times feels convincing, but it is a sample of curated stories rather than statistics.
Use visual mode for verifying logic, not for judging performance. After debugging, rely on numerical metrics, distributions, and out-of-sample results.
Also compare the tester's timing with live behavior. Running the same strategy on a demo account for a few days and comparing trades with the backtest for the same period reveals discrepancies caused by latency and data differences.
How closely do demo runs match the backtest for the same period and settings?