[ex_ex_nihilo]

It is true and mostly accurate by some back of the napkin math I've done. What stops people from doing 51% attacks is a few factors:

1) If you can pull off an attack for a few hundred bucks, it's probably on a coin that will only net you a few hundred bucks. Remember, these coins are thinly-traded altcoins, many of which only do a few hundred bucks worth of volume on exchanges in the first place.

2) Once you pull off your attack and people realize what happened, the "free" coins you get back will be worth shit because nobody will want that coin.

It's not as much of a free lunch as it might seem at first blush. I most certainly possess the domain knowledge and skill to pull off a 51% attack, as I have contributed code to Bitcoin and many other cryptocurrencies, written open source mining software, etc. But I would need several million dollars in starting capital to make the kind of money that would make me even consider it. That time would be better spent hacking away on a project with the potential to make me sustainable income, or a client project where I'm paid by the hour.

[slivym]

On point 1 take for example Ubiq, it'll cost you say 10x the 1h rate or $4.7k, and the 24 hour volume is 130k, that seems like a pretty good ratio to me. Or Bytecoin - let's say 100x since it's harder to get the capacity - cost $50k, 24h volume - $18m.

On point 2 I thought the entire point was to double spend - so you have 1 einsteinium or whatever, you put it in the exchange, you use it to buy bitcoin. You re-spend it at another exchange buying bitcoin. You got twice as much bitcoin as you paid for. As long as you're out before it becomes obvious what you've done you make profit.