Reading the paper on so-called 'Selfish Mining' http://fc14.ifca.ai/papers/fc14_submission_82.pdf
On page 12 Figure 2 relative pool revenue equals 1 at 50% of hash power 'selfishly' mining. The authors are saying you only need 50% of hash power to obtain 100% of fee revenue.. Am I missing something or is this just an obvious red flag for the analysis? Referencing their state machine on page 8 Figure 1, every time the private pool publishes their branch there should be a 50% chance of public miners finding the next block. Are they saying there would be a near-perma fork, with the private pool never actually publishing, leaving all revenues in limbo (ie never been resolved to a main branch)?