[Spoke]

I believe the argument they are implying goes like this Molly:

1st. Assume that all MUDs have players that vote the "intended way" by opening a browser, clicking on the vote icon in their sites and checking into TMS.
2nd. Assume that all MUDs have some players that vote "semi-automatically" by some sort of script that takes them directly to the "Do you want to vote for this MUD?" page.
3rd. Assume that most MUDs have players that will use a "completely automatic" script that will open browser-vote-close browser.
4th. Assume that in general (statistically speaking), the proportion of players on each category is similar. It is clear this is maybe the weakest assumption but on average I do think (opinion) that it is valid.
5th. Assume that the majority of the people voting belong to the first two categories.

CAPTCHA would take care of the 3rd category, meaning those voting would be humans, but if we accept the 5th argument that would not reduce the perceived traffic by much. Now the problem is the other two categories, if someone went so far as to set a script that would save them a couple of steps posting a vote (open the browser directly at the "Yes or No" page), it would make sense to assume that a good chunk of these people would be at least inconvenienced by a CAPTCHA and maybe the proportion of inconvenienced people in the 1st group would be somewhat lesser. Now the problem is that any one person that belongs to either of these groups that has been inconvenienced enough to not care about voting means a real traffic hit less. I do not have any way to decide what the proportion is between the first two groups, so nothing further could be said on that direction, but the final argument is that even though you would take care of the meaningless hits you would also dissuade real hits from coming, which at the end, is bad for the site.