Next step, starting tomorrow, is going back to trying to detect altcoin pump'n'dumps. My more recent attempts at traditional TA should be more successful in finding the steep gains involved, and when to sell. I'd be surprised if I can't make 5% per day for starters. Like a bloodhound I'll follow that trail, and in a few weeks I'll know if it's feasible.
The thing that threw me off the last time was that several coins have price rushes which start out very promising, but never amount to anything. I.e. once you get on board, it's already too late, and the price starts falling. Typical pump'n'dump giveaways are minor price increases and an increased interest shown in the volume volatility and surge.
Things that would cause me to fail are:
- unable to analyze and parameterize the volume.
- too few similarities between pump'n'dumps, if you can't pattern-match it in advance in any way, you can't predict it.
- too diverge patters before an average pump'n'dump, which means one'll buy-in too late, which means no gains.
- too much noise. If the signal-to-noise ratio is too high, either within an altcoin or between currencies, it would be very hard to find a good spot for buying.
- too low auto-correlation within and between volume and price in the pump.
Even though I think all will be causing some problems, and the combined hurdle will be quite big, I really think I'll manage.
