Difficulty is part of the calculation of the length of the chain, so the fork would eventually be abandoned.

what the OP is trying to say is that knowing difficulty is changed every 2100 blocks.

he will set up a private network of 2 PC's and at block 2098 he will set both computers time to be year 2100, thus as mining 2 more blocks. in this closed network. then the protocol in his theory will see that the blockchain has not mined blocks for 85 years and then give him super easy difficulty..

now the reality. when he rejoins the network he wil be all alone on his fork and everyone else will mine as usual not realising he exists, his coins from those 2 blocks would show as invalid and unspendable (wont confirm if he tried to send them) and he will get on with his life.

OP you also have to realise that you trying to (in basic principle) solo mine block 2098 and 2099, would be still at an extremely high difficulty before the change, which unless you had a asic farm, would take you not 20 minutes.. but days or weeks if lucky..

so dont waste your time on this. many thousands of people have been there, done that.

so heres the T-shirt



the other theory is to gather soooo much hash power to rival Ghash.oi etc. that if lucky he is the one that mines the critical block 2099 and it shows a timestamp of the year 2100..

goodluck, have you got over $10mill of equipment to try?

I don't pretend to know enough to comment on the specifics of such an idea. I do think as an academic exercise it is good to be aware of this kind of stuff in order to stay ahead of people that would try to game the system. I don't believe most such attacks would succeed in getting a bunch of coins, but they can cause chaos and fear which does affect our wealth. 

That attack won't work because there are validity checks
on the timestamps.

https://en.bitcoin.it/wiki/Block_timestamp

you misunderstood the OP, he is talking about manipulation via difficulty adjustment , but that won't work because the timestamps are validated.

If timestamps were not validated, I think it actually would be a problem because attacker could build a very long chain very quickly and broadcast that whole chain.

I see your point now, sorry... but still I think there could still be problems.
Why wouldn't the network build on top of your long chain since you were the first to broadcast it?
Also, it would be a very chaotic situation if you had miners trying to solve at different
difficulty levels.  A lot of reorgs.  And a distinct possibility someone would come out with
X number of cheap blocks.  Such chaos would be very bad for Bitcoin.

Well, I think people should try to keep thinking of ways to attack Bitcoin.
A new idea can come from any mind.

That said, I agree, the devs are very smart people and have had
a lot of Bitcoin experience.  They are usually several steps ahead
of the crowd.

I didn't do a lot of research, it's just something I came up with.

Can someone expand on how the timestamps are validated?

A timestamp will only be accepted if it is no more then 2 hours greater then the average of the time of all the nodes that you connect to. In other words the most you can manipulate the time is to make it 2 hours in the future. You would not wish to do this for found blocks, as the first block found is generally the one accepted to the blockchain assuming both chains are the same length.