Skip to main content
Open Volatile range stress from the demo app’s home screen, or open /demo/range-stress in the running app. The synthetic feed targets 30 updates per second and keeps 6,000 rolling line points. Prices swing sharply. One-second candles change in place between commits.
  • Run Y-axis outlier test: reset to prices near 100 and a 12-second window. At 3 seconds, send 1,000 for one second. At 7 seconds, send −500 for one second. Watch the actual Y-range readout expand in both directions, then recover after those outliers leave the window. The sequence resumes the feed automatically. Run it in either line or candle mode. Negative values are enabled in this demo to exercise the lower bound.
  • Volatile feed: leave the outlier sequence and restore the random-looking swings.
  • Spike to 400 / Crash to 8: force the next live price, then let the feed recover.
  • Pause / Resume: stop and restart the feed and chart’s time movement.
  • Show candles / Show line: switch between the two views of the same feed.
  • Zoom out / Zoom in: change the visible window between 12 and 30 seconds.
  • Correct visible history in place: alternate an interior history value between 500 and 40 using .modify(), without changing the array length. Candle corrections also retain their open/close values inside the high/low range.
Try pausing before a history correction. The fitted range should respond even though the feed is stopped. Resume and check that live prices and candles keep moving. Spike/crash buttons queue a value until the next running feed update. History range caching is automatic. It reuses committed-history low/high results between data notifications while live values, live candles, and range animation continue updating. See Historical data. This screen uses synthetic data and is a visual correctness stress test. A smooth simulator recording alone does not measure device CPU, battery, or release FPS.

​
What to expect from the outlier test

The fitted Y-range starts near 90–110, grows to about −680–1,180 with both outliers present, then returns near 90–110 by 27 seconds. The extra space around the data is the normal chart margin. Upper and lower bounds recover separately as each outlier leaves the visible history. The X-axis remains a 12-second time window; the price outliers change the Y-axis. For a repeatable run, leave the other controls alone until recovery completes. The visible range readout samples the chart scale four times per second and is part of the demo, not library instrumentation.