From what I understood, it is the ability of a miner with 51% hashing power of the whole network to be able to control bitcoin transactions.
the thing about 51% attack that i have learned is the fact that 51% is the border line of the limit that you should have to start doing the attack not the amount that you can pull it off with. for that you probably need more than 70% or possibly even more.
They can possibly not confirm a transaction, change a transaction, and other awful effects.
you can not "change" a transaction! that would mean bitcoin is broken
what you do in a 51% attack is that you double spend. and in this process you have to make it worth the effort, otherwise you are wasting a lot of power for nothing!
otherwise no "awful" result would come out of 51% attack if you are just rejecting other blocks and building your own chain!.
This attack almost happened 3 years ago with ghash.io mining pool almost reaching 51% (they were so close, when they reached 50% hashing power).
i believe they were close to 51% for a long time and people were worried and some day they reached 60% (it was also a mining pool after all and people were connecting to them) and nothing bad ever happened!
What if these people colluded to attack bitcoins using the 51% attack? Do you guys think this is a possibility and will those entities that are pushing for segwit2x attack the network?
they already have the hashrate, if they wanted, they could have attacked bitcoin all this time. what would SegWit2x change? if the 80% leaves to mine 2x the hashrate of 1x drops and if they leave 2x to attack 1x the 2x hashrate will also drop.
all these can potentially cause a huge drop in price. i don't think miners would want that. and as i said above 51% attack is hard and expensive already without the price drop!