Funding rate as a usable condition

"Moving the judgment into the rules where it can be tested" is the best argument anyone has made for a feature on this board.

Supporting. And it would let us settle whether it's actually predictive instead of arguing about it.

Logged. Being straight about the state of it: funding is display-only right now, it isn't in the condition builder, and I don't have a date. Two things that would help when we scope it — which funding value you'd want (current, predicted/next, or an average over the last N periods), and whether you'd gate on an absolute threshold or a percentile of recent history. Those are different amounts of work and people ask for both.

A few cautions before this turns into a threshold-fitting exercise: 1. Funding is regime-heavy. Tune a cutoff on a long directional stretch and you've mostly fit that stretch. 2. Percentile-of-trailing-window generalizes better than a fixed number, in my experience with similar gates. 3. Whatever you build, log the trades it SKIPS too. A gate that only shows you what it let through can't be evaluated. 4. Expect a small sample. Extreme funding isn't common, so you'll have few blocked entries and wide error bars on the result.

ha, yeah, currently my process is "check funding, feel a chill, take the trade anyway with less size" which is not a process. Re: TraderWe's question, I'd want predicted/next plus a trailing percentile if I had to pick, absolute numbers age badly. And honestly I'd take an ugly version. Doesn't need to be pretty, just needs to be testable so I stop being the discretionary layer.

Genuinely curious — if funding is extreme because everyone's long, does that mean price is more likely to reverse, or just that the trade costs more to hold? I keep seeing both explanations and I can't tell if people mean the same thing. Feels like the answer changes what you'd even build the condition for.