Thanks,
RackRateRefugee!
I’d treat the result as a risk score rather than a yes/no answer
👆 This is a really good way to frame it.
Agree 100% on risk score vs yes/no. That’s basically how I think about it too — the label is just a shortcut, but the real value should be in the “why”.
Most of the variables you listed are actually what the tool is trying to account for behind the scenes:
- Carry-on vs checked bag → included (big impact on buffers)
- Terminal / same alliance → partially modeled, but still a bit conservative unless confidence is high
- Immigration / security re-clearance → included via connection type / border regime
- Last acceptable arrival time → that’s the “latest viable arrival” output
Where it’s still weak (and you called it out well):
Historical delay patterns → not included yet
Clarity of assumptions → could definitely be surfaced better instead of staying “under the hood”
the “why” behind the score matters more than the label itself
This is also spot on. Right now it leans a bit too much on the label, and I’m working on making the reasoning more obvious (so people can apply their own judgment like you described).
Also fully agree on your DP — once you add bags + immigration, the risk jumps a lot. And the opposite case (carry-on, same terminal, backup option) is a completely different decision.
Glad to hear you’d use it as a sanity check — that’s exactly the goal.