You still need a significant initial amount, and the right conditions, but you certainly don't need the entire $91 million yourself.
Yes, in the article they put a very simple scenario and a bit of a lazy one, an assumption that everyone will just stick to the trades already set, there will be no one dumping, there will be no one withdrawing their sell order and placing them higher . Also, it is implying that you have to move the price on all the exchanges, not taking into account the herd effect
if a reputable exchange jumps suddenly and arbitrage bots are taken off guard without enough funds.
This was clearly not long ago when simple sudden dumps of 1k range
BTC managed to crush the price
Besides, there is a far easier way and even lazier than this, all exchanges have an estimated trading feature where you simply input how much you're willing to try and they will show you what's the average price you will get and the max price you pay for.
I didn't get it, it says that there was 90 million bought and 67 million sold, lets say 23 million for short and there was 1% increase. First of all we all know it will not be same all the time but lets assume for a moment that its always the same and 23 million difference means 1% increase. That means we do not need 90 million to increase right? I mean there is already some buy and sell going on anyway, we just need whatever the difference is right now and to make it 23 million.
No, he is right, you have to buy 90 million to create that difference at the current sell volume, If nobody would sell, yeah, you would only need 23 million, but that's not the case.
You're assuming your 23mill second buy on top of this will not encounter any resistance.