In Europe all TD use AREA to determine seeds in tournament.
Even if not up to date, it's a great tool to evaluete relative players strenght, which is useful if said players comes from all over the world and don't play each other frequently.
For exemple When Alan Smee (an australian) played in Europe it was a breeze to determine that he's a very strong player and deserve one of the top seed (currently he's 40).
This.
And AREA uses only games played at tournaments reported by tournament directors to AREA (i.e. Bruno).
In other words verified games played in tournament settings are used for tournament seeding.
As far as I can tell AREA provides the best ranking list which is available. This does not necessarily mean it's perfect, but it is much better than the one of the ASL Scenario Archive where one or two games seem to let your ranking swing wildly even if you have hundreds logged. Apart from not displaying the correct number of games that you've logged most of the time - probably while it recalculates entries.
By personal experience I can tell, that AREA gives quite a good impression. If another player's AREA-rating is 100+ points above my own, then I have it usually confirmed that his skill-level is better than mine because it can be felt when actually playing him.
You can roughly say that you'll have a crack player with an AREA of 1800+, a really tough one with 1700+, a very seasoned and experienced one with 1600+, and someone who knows his ways quite well with 1500+. If you look at the names next to these ratings, I believe you will find that this assessment is quite sound.
Of course, as always, these are only general guidelines subject to the number of games logged for the player etc. I can't remember at which level a new player enters the AREA ranking either.
As AREA only considers reported tournament games and Bruno's time is limited (i.e. this thread...), the ratings tend to lag a bit behind. For example, if you have a very prolific player that attends relatively few tournaments, then chances could be that his actual skill-level has increased quicker than AREA can "catch up" with it.
Another possible distortion of the world-wide AREA rating could be that the general skill level in some areas of the world might be higher than elsewhere. For example, it is sometimes said that the Scandinavians are pretty good players in general. Since by tendency the attendance of tournaments is highest of the local playerbase, this could mean, that a 1500+ Scandinavian AREA-rating might in reality be a bit better than a 1500+ German AREA-rating. But this is, of course, difficult to figure out in detail.
What remains is that AREA gives a decent idea which type of player you have (crack, extremely good, very good, good, etc.) if the sample is not too small and that the AREA rating is thus useful for seeding.
von Marwitz