Combining BSGS and Kangaroo Method for ECDLP

Anyone think we can merge bsgs and kangaroo for an ecdlp solver? Like, could it actually work and be faster than just using one of them alone?

It might be doable but it’s gonna take some serious time to figure it out.

So you’re saying it’ll be way slower? That doesn’t sound promising...

I honestly believe it can work this way For example, you set up jumps based on starting points Make a rule for jumps where you use the last 6 digits of X plus 1G After around 5 million iterations, I’m pretty sure all nearby points will converge to one spot... I've tested this and I think I proved it to myself.

Interesting concept, but how would you apply that practically? Like, if we take puzzle 135, you start jumping from that last 6 digits of X plus 1G Then you could dump every millionth point into a babystep file, right? You could end up with like 200 million points and calculate G added from the starting point to that 6th last digit.