One reason we may be skeptical of the simplified puppies/kittens example is that I think you are alluding to a simultaneous Nash equilibrium, whereas with Steemit people will vote at different stages, so we have to account for the aspect of taking turns and time element. Steemit is full of 'sub-games' like the one you describe between Puppies and Kittens where previous actions affect immediate and future payoffs.

Also, the Nash Equilibrium can be deduced but people are not always rational decision makers. People tend to have 'bounded rationality' so with the mass of information on Steemit they may not vote optimally.