I don't see the problem with this, though. If you're willing to give 5 points to a single game as opposed to 1 point to 5 games, that says something about the game getting 5 points.
The problem with that is that there is no common agreement between the voters on what constitutes a 5, a 4, a 3, and so on. Maybe some people only played 4-5 games but liked only two, then gave them 5s. Maybe some other people really liked 15, but couldn't really decide which is better, hence picked their 5 favourites and gave them equal ratings of 1. Etc...
In any case, if you look at the top 50 list with ratings removed, it's a good list, and it's 90% the same as the CRPGAddict's GOTY + highest rated lists, so we haven't discovered anything new, really. Ratings are pointless anyway (no pun intended), especially in a popularity based poll. E.g. how does it help anybody to know that it's game X, Y, Z rated as the top 3 if you don't care about their particular subgenres? By definition, because RPG subgenres are so varied, effectively taking the averages of the favourites of a large group of people doesn't make a lot of sense. Maybe all the results can reveal is that subgenre X is more popular than subgenre Y, but that's it.
On the other hand, if you take the union of the 5 games that people enjoyed most, that gives you a usable selection of good RPGs people actually liked (assuming you have a few old-timers among the voters and not just people who only played games released after 1992 or something). It also bypasses the drowing out of the valuable votes of the relatively few oldtimers problem completely.
Categories like best Wizardry-like, best DM-like, best Gold Box game, etc. would make actually more sense if you're hellbent on ranking things.
EDIT: Why am I going on about this? Well, if some guy wants to play some good oldschool games and starts dutifully from #1 of the popularity based poll, there's a good chance that he will be sorely disappointed. Ratings are quite misleading that way, and the approach of the CRPG book is infinitely better (no ratings, just find something that you find interesting by reading the descriptions). I for one decided to play Planescape Torment a few years ago because the top 70 list made me believe I am going to play the best RPG of all time... Well, I was extremely disappointed, to say the least, had to abort my playthrough after about 30 hours when the realisation hit me that well, this is the game, it's not gonna get any better. I was also led to believe that combat takes a backseat in the game, then I'm confronted with endless trash mobs in the catacombs, etc... Probably I would have enjoyed it a lot more had I not been primed to expect the best game of all time, so might give it a go again later with reset expectations. On the other hand, many of my absolute favourites are around the middle of the top 70 list (or are not even included). So, for me personally, ratings are a non-indicator, they actually just lead me astray, and I'm sure I'm not alone. Same with books, movies, music, and anything really that people like to "rate".