For a scoring system I'd prefer a par/bogey/birdie system like in golf. Par (0) means the game is good enough that there's enjoyment to be had there (but the competition probably outperforms it and you'd be better off buying something else if that's available but if it's not you can still buy the game and have fun with it). I believe par would also be the average since I believe that the majority of games is actually good enough to be fun.
However with an uncapped scale I have a feeling reviewers would end up creating score inflation because they rate one game the best ever, then the sequel comes out and is predictably better so they award it a higher score, etc. Personally I think the only scale points worth having are +2..-2 with +1 (if we use positive = better, obviously that's not how golf works but it seems more intuitive as a score) being great (the various must-play games), -1 not fun, +2 so awesome you need profanities to describe it and will probably keep telling people they should go kill themselves for not buying the game and -2 being so broken you're lucky that your computer doesn't explode.
Yes, that ends up with another 5 point scale but I think the hard cutoff between good and bad would completely change the perception, plus making the main scale three point is IMO enough.
What I'm opposed to is buy/rent/ignore, there's still the intermediate step of "wait for a pricedrop" and I tend to place the cutoff point for full price purchases VERY high, moreso if full price means 70€ instead of 50€ (I haven't found a game for the 360 that would warrant a full price purchase for me). The result is that unless the game costs like 10€ to start with a "buy" recommendation usually gets downgraded to "wait" but in some cases an immediate purchase might be warranted, if the scale already topped out by then I won't be able to tell those games from the rest.
As for that PSX article, he may be right or he may be not but since he does not attack the text of the review, just the score his article is worthless and comes off as "my score is better than your score".
EDIT: Also I made no mentions of technical quality on purpose. Technical quality is a means to an end, that is making the technology not interfere with the fun. If the great graphics or whatnot make the game more fun (e.g. a game about exploration probably lives and dies with the interesting presentation of the discoveries) then it affects the total fun and therefore the score, if the game is faulty in a way that makes it frustrating that agian interferes with the fun. If it's neither I don't think it should affect the rating. I'm saying this because I think some reviewers have a tendency to add or subtract points based on the technical merits even when they don't affect much (e.g. giving extra points to MaBoShi for its polish of the extra features or deducting from Earth Defense Force for the obviously low budget design).
