I have never seen a swing that large but he was obviously renting. I was renting at the same time on the same site and seeing their hash rate drop as I increased mine. Depending which rental site they are using it can swing easy if he is being outbid. Some rental sites you rent/hour and others you pay a certain amount and if you are highest you get it. Others can outbid you.
Oo the economics of it I'm up over 200% on my renting in the last few days (2 btc in and 4.2 btc out) so if they are going after the same thing I am then I can understand. It is definitely luck chasing though. If we hit back to back reds then I would be down. But it does only really take 1 block to almost break even. If we hit 2 or more it is all bonus money. it all depends on how much you pay for the rentals.
AS you can also see both them and myself renting have not actually helped hit a block either. At least not the way that I think of blocks anyways. We did work to solve blocks but didn't specifically find one ourselves.
Fahlcor
Interesting. So, it is basically gambling (unless the rental cost is priced lower than the expected value), correct? If it is priced exactly equal to the expected value than it is more a variance gamble? Right?
So, your expected value of buying 24 hours of 1 PH/s would be: 1 (the amount you purchased) / 254 PH/s (the network size) X 3600 (approx bitcoins per 24 hour period) X 0.98 (bitminter keeps some money) = Expected number of BTC returned.
How does the pricing offered compare to the above?
PS. I understand that hashing (and not solving) helps others solve the block, that was not my implication. Everyone doing work helps whether they are the block solver or not.