Fixing the finality_fixedLength sim

[sternhenri]

cc @sa8 Issue: This alternate sim for confirming the fil confirmation time under a selfish mining scenario is quite a bit more optimistic than its counterpart.

Background: The approach here is rather than to try and have the miner run the longest possible attack repeatedly over the sim period, we simply have the miner run an attack for k blocks and see how they fare at the end (ie how often do they come out ahead).

Necessary Fixes:

• As with the early other sim, this is tantamount to only looking at the first attack, and not the best. There may be a fundamental methodology issue here.
• The attacker here is suboptimal: they start their attack immediately rather than at optimal conditions (whenever they win with a headstart, and whenever they win more without).

[sternhenri]

Overtaken by doing in the convergence sim instead!