No it doesn't. You don't understand the whitepaper.
With 33+% of hashing power I can break Bitcoin, this is what the paper says.
https://people.eecs.berkeley.edu/~luca/cs174/byzantine.pdf talks about 33% too. 51% works only in unrealistic assumptions. I can imagine a case where an attacker with 90% of hashing power won't break the system. I hope you got my point, I'm not going to extend it because it's an off-topic.
The 33 - 50% control only gains an unfair amount of rewards. It doesn't convey the power to win every block. The reason that can "break" Bitcoin, is then the other miners have an incentive to join the selfish mining cartel until 50+% control is attained:
This attack can have significant consequences for
Bitcoin: Rational miners will prefer to join the selfish miners, and the
colluding group will increase in size until it becomes a majority. At this
point, the Bitcoin system ceases to be a decentralized currency.
Also with asymmetrical rewards (and all other factors not considered), the 33-50% attacker will over time increase his proportion of the hashrate, because he is more profitable than the other miners.