Ranking films — or any other creative work — objectively is notoriously difficult, particularly as you end up comparing chalk and cheese: which is the better film, Dawn of the Dead, Meet Me in St Louis, or Gosford Park? The temptation, whether you go in for academic aspirations or inverted snobbery, is to award ratings based on a sense of 'worthiness', where gore and horror, say, are at one end of the scale, and ambiguous endings and social commentary are at the other... although the direction of the scale in question is likely to vary! Take a look at a few newspaper critics' reviews. High-brow papers — speaking from my personal observations — are apt to award top marks to foreign-language film by a woman director shot under a burqa in Afghanistan, for example, and sneer at brainless summer blockbusters, whereas the mass-market sheets will happily endorse the same blockbusters as the summer's must-see attraction, and poke fun at pretentious art-house filler. (Actually, my own experience is that both kinds of highly-recommended film tend to be equally disappointing, but let's not get into that...)
When it came to allocating ratings on the IMDb for films I'd seen, I naturally enough started off by trying to operate a relative system: "film X is better than film Y, therefore I'll give it one mark more, but film Z is better than both of them, therefore I need to mark it even higher..." It worked — more or less — for a handful of films, when these were all of a similar genre. But it had at least two major drawbacks: it required me to remember (or constantly look up) which other films I'd seen and how, relatively speaking, I'd ranked all of them — plus the fact that every time I changed my mind about the merits of one film, it implied a cascade of demoting to keep all the others in the same position relative to their neighbours. And, as mentioned above, it became virtually unworkable when dealing with films that had nothing in common with each other.
So I very soon stopped trying to rank films by merit, and ended up with a very simple set of conditions based entirely on enjoyment: did the film provoke a level 6 response, a level 9 response, or — heaven forbid — a level 3 response? However great a work of cinema it was supposed to be, if I wouldn't recommend it to a film-loving friend then I wasn't going to award it a mark above 7... although in practice, I confess that I will cut a tedious but famous 'classic' perhaps one mark's degree of slack, bumping it up a level on the grounds that there must be something there that I'm not seeing!
The resulting 'absolute' rather than 'relative' table has served me pretty well, in that if I do want to make comparisons between two films, I tend to find that their positions are accurate. If I marked one film higher than the other, then yes, on balance I did think it a better film — even though I wasn't consciously comparing the two at the time.
In practice I've found that the available spread of marks is a bit mean: I made the arbitrary decision that anything from 1–5 would be in the 'bad' spectrum and from 6–10 in the 'good spectrum', but as a result the vast majority of films fall between 6 and 8, which doesn't really give a lot of leeway for distinction. Some 7s, inevitably, are better than other 7s. Meanwhile, the scale below 5 is so little used that I've never really codified it, and still assign fairly random judgments on so-called 'relative' demerit... This is perhaps unsurprising, as I do try in advance to avoid watching bad films!
One final note: my choice of top ranking does depend on longevity. I'm not going to award a film a 10, no matter how much I enjoyed it, until it has also stood the test of time. No instant classics: it doesn't have to have been famous, but it has to have been known and around, and still liked and played across a considerable number of years.
To be honest, I can't see anyone's caring what I personally think about various old films, especially as my experience quite often seems to conflict with the collective wisdom of the IMDb's users: but to see the above principles in practice, view my IMDb voting history. It may be an eclectic selection from 90 years of cinema (and heavily biased towards the pre-1960 era), but it's honest enough.
A subset of the above are those films that I've written comments on; most of them formal reviews that I've spent a lot of thought on before copying up, some just tossed-off comments written while actually on-line. The difference is fairly obvious, I'm afraid!