The reason why TMNT (and similar "new" titles) always get "on top", is that early on, nearly the only people who have played them are the people who bought them early. And these people always rate them very highly for a few reasons...

1) If they didn't like the theme/design they probably wouldn't have bought it (except for those few people who buy EVERY new machine).
2) They want to feel that they "made a good choice" and "got their money's worth".
3) The "new" toy is usually the favorite toy, until that newness wears off.

I would think that a machine should not be listed until like 6 months after release. Just limiting it to 100 ratings or something will not do much (as it will still be the first 100 purchasers that are rating it). It really needs to be on location for a while, so people who haven't invested in a purchase also get to vote on it. Of course, the pandemic has made this even harder for now, as location play has dropped off greatly.

I would like to see an alternate Top 10 (or maybe 25) that is computed by everyone submitting their personal favorites list (in order). This would mean that everyone would be voting using the same "scale", and it would be much harder to try to manipulate the results (as the best you could do is rate 1 machine your favorite, and the worst you could do is not list a machine at all). No more "vote a 10 for all the machines I own and vote a 1 for all the others".

The point is the list is not for yourself...it's for everyone else.
And to see how your list compares to everyone else's.

If some game (say TZ) is listed near the top, then it says to people "this is a great game, you should check it out".
And you may think WH20 is the best game ever, but other people list it as 35th.

Obviously, everyone already knows their own list...it's learning everyone else's list that is interesting.

The problem with this is that it would take a long time before you had enough pair comparisons to provide useful results (even if you forced a question at each log in). And by the time you do have useful data, a bunch of new machines will have come out - invalidating the data you finally have.

This is a great method if the data being compared is mostly static (i.e. listing the "best" presidents), but not so good when the choices keep changing (and when people tend to change their minds a lot).

Yes, that's true...but those two cases are the exceptions, and I would argue that this is reasonable.

For example, when I played TWD, I thought it sucked (and I still don't understand what people see in it) and it was likely that this was partially due to early code. But all too often, on location, the early code never gets updated by the operator (unless there is some problem that requires an update)...so that's the code that (non-owners) play and know (and rate by).

The BM66 near me still had no secondary villain modes the last time I played it (not sure what version it was, but was obviously early). And while I still like the game, I'm not going to give it a better rating just because someone on line says "the new code is great".

If you are going to release an unfinished game, you can't complain if people rate the game based on what they actually play, not what might eventually get released.

But that's my point, if you put out a "turd version" of a game, you can't be surprised when it ends up with a bunch of bad votes. Sure, maybe "some time later" the "good" version comes out, but odds are the (non-updated) machine out on location is still a "turd".

The people who buy the game (and update it) may be happy, but the guy who tries it out on location will walk away thinking it's a turd - and he'd be right.

I think MB and MM clearly deserve to be up at the top, and I think EHOH and POTC are over-rated. Just shows you how different people are different.

(and I think WH20 is way under-rated)

They only update the ratings/rankings once a week, not as every rating comes in.