Divergences are well-known tool for finding trend reversals. Powerful, yet... unreliable. Sometimes they show up, sometimes they don't. Sometimes they really mean trend end, sometimes the trend continues. I've witnessed countless discussions about which indicator to hunt for divergences on. All pointless.
Earlier today I found a very inspirational script by Lonesometheblue, combining multiple divergences together. Actually - summing them up and showing hunted divergences count in a label. Because if a divergence shows up on a single indicator like RSI, it can indeed mean trend reversal, but... yeah, check previous pargraph. But if this divergence shows up on 8 different indicators at the same time, then chances increase, don't they?
I decided to do my own version of Lonesome's indicator, because that one is using Label.New, which doesn't allow analysing entire script, only latest 50 occurences or so. I'm not going to hide it - I copy-pasted a lot from Lonesome's code. And from Everget's built-in divergences script, which you can find in PineScript/New. I didn't even bother to modify variable names or cover tracks in any other way, pretending it is my work from scratch. All with these two gentlemen consent and blessing. I added labels from myself, plus bit of rework and here it is:)
I hope you'll find it useful.
P.S. While analyzing divergences today, I discovered one thing. Hidden divergences are not useless, as many people think. But they should be used at trend dips to join the trend, not at trend ends, as reversals. Experiment yourself.
P.S.2 Still, never use only divergences. They need confirmation, they are not self-fulfilling prophecy.