I personally find the reviews section for games useless on GOG. The problem is that many of the reviews I see have nothing to do with the GOG game itself. Common review comments:

1) This publisher/game sticks, don't buy it. It's a mod...
2) This is a review of the non-GOG version...
3) Please bring expansion pack XYZ to GOG...

None of these are reviews of the GOG version of the game itself. Why is this even allowed? Downvoting the review doesn't help get rid of it.

I'd personally like to see a restriction placed on the review section. If you don't own the GOG version then it doesn't make sense to be able to review it. It doesn't matter how much you love/hate the original, the Steam version, etc. The GOG version may be different and therefore should be owned before it can be reviewed. This is akin to someone posting reviews of the PlayStation version of a game on an Xbox review site. They are different games. It also would prevent people from going on rants about how the GOG version isn't worth buying for whatever reason. Unless you own the GOG version then should your opinion really matter? I don't believe so. We have general purpose review sites for reviewing games and publishers. When I see a game on GOG I want to know what is good/bad about the GOG version (DRM, crashing, etc).

If restricting access isn't possible then at least post whether the user owns the game and possibly how long they've played. This will help identify people who are reviewing the non-GOG version. The length of play is useful to know whether someone's experiences are reactionary or over the course of the game. I've seen many reviews in other places where people complain about a game being too long/short, hard/easy, etc and they hadn't even played an hour. These types of reviews are less relevant to me.

Don't agree with point 2, but agree with the rest. As for point 2, as long as the reviewer indicates their review is of the non-GOG version, I see no problem with that. After all, the gameplay isn't going to be different in a GOG version (possibly some minor changes, but nothing significant), nor is the story. So a review of those elements is still of value.

Exactly.

Also, I'd add that reviews of the GOG version get out of date, they often discribe problems long since fixed and as result only confuse, even if originally they were very helpful.

We've had this discussion a number of times and I've always thought the best answer was to have two types of reviews;
1) A gameplay review, platform and store independent, that just focuses on whether it's a good game or not

2) A technical review, this would be focused on the specific version; whether there are bugs or technical issues and if it's a good port (where that is applicable) or other info like that. These would really need to be properly dated and maybe have a version number of the game at the time (as well as being tied to accounts of people who owned the game only) so that if the information became obsolete (bugs were fixed) it could be seen easily.

All improvements of the review system face one critical obstacle: there are not enough reviews for it to be worth dev time.

Not enough reviews? I'd like to be able to filter the 20+ pages of reviews you have to navigate through. I'd like to show 1 and 2 star rating to check for colossal game breaking issues that fanboys downvote. I'd like to show low rated reviews to see what fans/haters are hiding. I'd like to be able to filter out foreign language reviews so that they don't get downvoted by retards who only speak English.

Hey! I only speak English, and I'll have you know I have the intelligence of a large plant... :P

What about all those useless "OMG! I used to play this game when i was 5 years old! Haven't played it in 20 years but it was so awesome! OMG! Buy it!!!!!!" that is basically common on every old game, no matter who bad it is because people are driven by blinding nostalgia and "oh i remember some parts!"
I find that way worse than people thinking the review section is a forum where they can ask for DLC because those are usually short and can be ignored, while all those "blinded by nostalgia" reviews probably fooled a lot of people into buying crap.

If you're comfortable with using userscripts, I suggest you install adaliabooks <span class="bold">Adalia Fundamentals</span>. Has a ton of features, and offers some nice options regarding reviews.

Games by # of reviews are distributed as a very steep Pareto. Most games have like a couple reviews, and the few that have 20 pages don't need reviews to sell. I'm all for an overhaul, but I understand why it (probably) isn't going to happen soon except for maybe a language filter.

addendum:
I do think GOG's lack of attention and respect toward conceptually permanent user-generated content / metadata is a major ethical failure, the #1 example being the removal of store pages for removed games.
I don't use Steam, but, from the outside, I like what they've done with the review interface and game guides.

That would realistically only affect problem number 1, though. People with nostalgia would still insta-buy games and leave reviews based on their memories before the GOG versions were even installed. Besides, there are plenty of people out there who play for 30 minutes before leaving a 1-sentence review declaring X game to be the best or worst thing ever made. Filtering out potentially good ones that counter that kind of thing simply because the reviewer owns the game on Steam or something would be counterproductive and arguably make the review system even worse.

Does Galaxy track play time? Even if it does, displaying that would presumably disadvantage anyone playing without it and lead to more client drama. Probably not a good idea to go down that road, either.

Like 227 said, you can't measure how much time someone spent with the GOG version of a game unless they ran it via Galaxy, which many users here don't. Even on Steam the data can be misleading. They might have played the games offline, for example, and in those cases Steam doesn't record the time.