Catching the final divergence developing?
An example here.
You are looking for a 30-minute fractal print, that comes within 3 pips (normally over) the most recent high / low & is within 8 candles of the recent RSI2 peak’s print setting up a reversal divergence.

Of course, the current candle is not gonna be called a fractal until 2 more candles on the right making lower highs.
Also, you want to have seen 3 progression highs on the upside and 2 on the downside to even start looking for a Final Divergence.
The RSI2 basis has to be halves, not Closes.
…Fractal granted. Final Divergence plotted.

Here’s the code for the final burnout: pedal to the metal RSI (>99 / <1) for 3.5 hours straight (7 candles – after Progerssion Peak 2 / 3 again).
/// final burn out on the upside
if (redcount[i]>=2 && RSI2[i]>99 && RSI2[i+1]>99 && RSI2[i+2]>99 && RSI2[i+3]>99 && RSI2[i+4]>99 && RSI2[i+5]>99 && RSI2[i+6]>99
){
ObjectCreate("120BPMMM"+IntegerToString(i),OBJ_TREND,0, Time[i], High[i], Time[i+6], High[i+6]);
ObjectSet("120BPMMM"+IntegerToString(i),OBJPROP_COLOR,clrBlue);
ObjectSet("120BPMMM"+IntegerToString(i),OBJPROP_RAY_RIGHT,false);
ObjectSet("120BPMMM"+IntegerToString(i),OBJPROP_WIDTH,22);
}

You simply cannot say that you are not learning anything here.