Three-finger tap reliability when macOS three-finger drag is enabled: notes from comparing BTT with Jitouch

Feature Requests
1

Hi,

I wanted to share a small but very real gesture-recognition case study. This is not meant as a complaint. I use BetterTouchTool because it is extremely powerful, but after a long comparison with Jitouch I think this may be a useful edge case for improving BTT's trackpad recognizers and diagnostics.

The setup:

  • macOS three-finger drag is enabled.
  • BTT three-finger drag compatibility mode is enabled.
  • A BTT three-finger tap trigger is mapped to Command+W.
  • The trigger exists and sometimes fires, so this does not look like an Accessibility permission or shortcut-sending issue.

The problem:

Even after tuning and enabling the compatibility mode, three-finger tap in BTT remains much less forgiving than Jitouch. In repeated manual tests, about 20 light three-finger taps would only trigger a handful of times. The first few attempts might work, then the success rate drops. The subjective feel is that BTT is much stricter about posture and timing, especially when three-finger drag is enabled.

By comparison, Jitouch's Three-Finger Tap feels much more tolerant. The fingers can be slightly rotated or imperfectly placed and it still fires. Its One-Fix Left-Tap and One-Fix Right-Tap gestures also feel more natural than trying to approximate them through generic triggers.

Why I think this happens:

We looked at the Jitouch source/build and it appears to be a much narrower recognizer built around Apple's private MultitouchSupport.framework. It has first-class gestures such as Three-Finger Tap, One-Fix Left-Tap, and One-Fix Right-Tap, and seems to operate closer to raw contact frames with short gesture-specific state machines.

BTT, understandably, is a much broader trigger/action platform. That is its strength. But for a tiny high-frequency gesture such as a light three-finger tap, especially with macOS three-finger drag also active, the broader recognition path appears less tolerant.

One related finding:

While debugging BTT drawing gestures, we also found old disabled-preset drawing gestures that still appeared to compete with current drawing gesture recognition. After backing up and deleting 54 old disabled-preset drawing triggers via BTT's AppleScript API, drawing gestures became easier to reason about. I am not sure whether this is intended caching behavior or stale local state, but a recognizer-audit/cleanup view would be very helpful.

Some suggestions:

  1. Show a clear warning on three-finger tap triggers when macOS three-finger drag is enabled.
  2. Provide a gesture debug panel that logs rejected candidates, not only successful triggers.
  3. Add a more tolerant/raw-contact mode for three-finger tap, even if it comes with a higher false-positive risk.
  4. Consider first-class support for Jitouch-style One-Fix gestures.
  5. Provide a cleanup/audit tool for drawing gesture candidates across active and inactive presets.

I wrote a longer Chinese write-up of the investigation here:

AIYA link: https://aiya.de5.net/t/topic/203

Again, I am not saying BTT is bad. Quite the opposite: BTT is excellent as a broad automation tool. This comparison mainly made me realize that some very small trackpad interactions may benefit from a specialized low-level recognizer or at least more visibility into why BTT rejects a gesture.