the two clocks in that volume-confirmed momentum entry

Been chewing on the volume-confirmed momentum writeup in the strategy library for a couple of days, mostly the part where it explains why the session-long strength reading and the one-minute buy/sell imbalance aren't redundant. I like the framing a lot - one asks about the day, one asks about right now. That's a clean way to think about it. Where I get stuck is whether they're actually independent enough to be worth ANDing together. My instinct is that they're positively correlated by construction: a session where buyers have been in control all day is a session that generates more of those minute-long buy-heavy windows in the first place. So the joint condition may not be filtering as much as it looks like it is on paper. Or the opposite - it could be filtering far more than intended, and you end up with a handful of signals a week clustered in a couple of trending names, which is a sample size problem more than a strategy. I pulled my own logs on something structurally similar (not the same thresholds) and the number of surviving signals dropped by a lot more than I assumed when I added the recent-window condition on top. Didn't tell me much about edge, just about frequency. So the question: has anyone actually looked at the conditional distribution here? Specifically, given the session-long condition is already true, how often does the minute-window condition add information versus just confirming what you already knew? And if it mostly confirms, is the second filter really buying you selectivity, or just delay?