The Chart Tab is located at the Reports Area of the “Backtest Strategy” page.
It shows the performance graph of the strategy against a benchmark. The “performance” here is the percentage return over time (in relation to the initial capital). With this graph, you can quickly eyeball how the strategy performs: not just changes in the investment value, but the greatest drawdown it has suffered.
1. The light-green graph shows the strategy's performance, while the blue graph (or darker shade of green) shows the benchmark's performance.
2. The horizontal and vertical axes represent time (daily) and compounded gain (in percent of the initial capital), respectively. These axes will change their interval values once you zoom-in the chart.
These equity curves are shown in percentage terms, not the dollar value. And they start at 100% (the vertical axis), because that's the original amount of your account balance. So a value of 200% means it's currently at 2x or twice the original amount, 50% means it's half of that, etc.
3. The red shaded area shows the deepest drawdown the strategy experienced during the whole backtest period. It shows the size of such drawdown, as well as the duration it takes to go from peak to trough.
This red area doesn't show the entire duration of such drawdown (from the last highest peak to the new highest peak). Note, the benchmark doesn't have its greatest drawdown shown.
4. The grey shaded area represents the backtest period reserved for out-of-sample testing. You can set the in-sample and out-of-sample ratios with the “In Sample Periods” section (at the Backtest Panel).
5. To look around the chart: Pan the chart around by click-dragging it left or right. Zoom in and out with the mouse scroll-wheel. To revert to the default zoom level (general overview), double-click the chart.
You can also right click the chart to show a tooltip on your cursor, which dynamically shows the date and percentage return of both the strategy and the benchmark. Alternatively, you can use the bottom scroll bar to manipulate the views: drag it left or right to pan the chart, and shrink/expand it (by dragging its tail/head) to zoom-in and out.
6. On the “Chart Type” parameter, you can set how the chart appears. “Line” is the preferred type here, as it shows the two charts (strategy and benchmark) distinctly from each other, without any overlapping shades.
“Mountain” shows the line charts filled with their respective colors; this is good if the strategy and benchmark performs quite differently, thus the difference in performance seems even more apparent. But the colors may overlap and can get confusing to distinguish.
The “Column” type shows a bar chart instead of a continuous graph line. If you zoom in closely, you can see each bar represents a single day, with each bar containing both the strategy and benchmark's values.
This is good if you want to get precision reading on the chart (the exact values contained in each day). The tooltip also “snaps” to that single day, instead of wandering around.
Now, if you look closely on any chart types, there are two dots and a line connecting them. You can use these as visual aid when comparing the curvature of the two equity graphs.
7. On the “Y Axis” parameter, you can define how the chart's vertical axis (percentage return) is mapped out. “Logarithmic” is the default mode: it uses logarithmic scale, perfect for showing more clearly the changes in percentage-return.
For example, a change from 100% to 1,000% would appear in the same distance as the change from 100,000% to 1,000,000%. Both changes represent a ten-fold increase of the investment value, hence they're displayed in the same distance for better clarity.
In a “Linear” scale, the change from 100,000% to 1,000,000% would cover 1,000 times the distance of 100% to 1,000%, which means smaller changes can't be seen clearly on the chart as they're too subtle (remember the chart may range from zero percent to millions of percent). But the “Linear” mode is good for understanding the real or actual distances between values; for example, a benchmark that tops out at 1,000% return would appear completely low compared to the backtest that returns 5,000,000%.
“Logarithmic” scale is more suited for equity graphs. It shows the changes accurately: for example, losing $50 on a $100 balance hurts bad, ditto losing $50,000 on a $100,000 balance feels about the same, as it's the same 50% reduction in balance. If you lost $50 on a $100,000 balance, it's a pinprick.