Statistically best games of the decade (2000 to 2009)
Introduction
Metacritic aggregated 50 different publishers' best-of-the-2010s lists and ranked the games that appeared in most of those lists. Unfortunately, they did not do this for the 2000s, so I decided to do it instead. In case you're wondering, you can look at this spreadsheet to see all the publishers and how many points each publisher gave to each game. This is not set in stone, so please definitely tell me if I'm missing a publisher. More data is (usually) better.
Video Game Canon is doing a similar project that involves aggregating lists of the best games of all time. Go give them a look.
(the comic is from here, in case you're wondering)
Finding the publishers
First, I thought publishers in Metacritic's best-of-the-decade lists are more likely to also post a best-of-the-decade list for the 2000s. I visited each of their websites (except the European ones because English is the only European language I can read) and searched for keywords like "games", "decade", and "2000s". Some of them had good search engines that let me specifically search for things posted in late 2009 or early 2010. Some of them had garbage search engines that appeared to give me random results that didn't contain any of the words I searched for, so I also used Google to search each of those websites.
After searching those websites, I searched every website that Metacritic uses to calculate its Metascores.
A bunch of best-of-the-2000s lists are in neither of those categories. I went on Google and searched a lot of different search terms, including but not limited to "games 2000s", "games 2000s decade", "best games 2000s", "video games 2000s", "best games 2000s decade", and "game of the decade 2000".
I ignored lists that ranked a subset of 2000s games, such as best indie games or best role-playing games. I also ignored lists that are copies of other lists, because apparently they exist for some weird reason. I found 42 best-of-the-2000s lists, which is definitely less than what Metacritic found but hopefully still a significant amount of publishers.
Ranking the games
I gave points to each game based on how many times they appeared in best-of-the- 2000s lists and added all the points that each game received. More points are better. Each website can contribute up to 38 points to these rankings, and there are no exceptions to this rule. The next few rules do have exceptions, though.
5 points for placing first
4 points for placing second
3 points for placing third through fifth
2 points for placing sixth through tenth
1 point for placing eleventh through twentieth
0 points for placing worse than twentieth
If multiple games are tied on a list, then they get points based on the last place they tie for. For example, in an unranked list of 10 games, all of them are tied for tenth place, so each gets 2 points.
Guitar Hero and Orange Box
I tried my best to rank the 3 Guitar Hero games, Half-Life 2, and Portal separately. Unfortunately, there were just too many publishers that ranked the series instead of the games inside them, and I did not know what to do with them. They should not get ignored, but I also do not know whether they think Guitar Hero is better than Legends of Rock, or how they would compare to the other games if they had to rank them separately. Ranking the series instead of the games inside them is the best compromise I was able to come up with.
Citizen's News
There are 5 games that each won an award from this site, so each of them got 3 points. LittleBigPlanet and World of Warcraft were the only runner-ups in their categories, which implies that they are definitely second best, unlike the other runner-ups that are tied for second through fourth. LittleBigPlanet and World of Warcraft got 2 points, and the others got 1 point.
Escapist
The unranked top 10 list on this site is really a top 12 list in disguise. This critic put the 2 Halo games in 1 slot and Nintendogs and Brain Age in another slot. I thought this implied they are not fun enough to have their own slot, treated them as tied for ninth through twelfth place, and gave them 1 point each.
Everyview
This site had 6 critics who each posted a top 10 list, so I can give 5 points to each game that is first on one of their lists, 4 points to each game that is second, etc. and then add those scores to determine which games are ranked highest by the average of the critics.
Gamespot
This site posted 2 top 10 lists, 1 ranked and 1 unranked. I gave 5, 4, 3, or 2 points to each game on the ranked list and 2 points to each game on the unranked list and then added the points.
gnd-tech
Each award has one winner and at least one nominee. The winners each got 5 points. The nominees got points based on how many different awards they were nominated for, and how many other nominees were on each of the awards. For example, Crysis was the only nominee for best graphics, which implies that it had the second-best graphics and deserved 4 points. 15 games were nominated for best shooter, which implies that they are all tied for sixteenth place and each deserves 1 point.
IMDb
This site lets users sort lists based on popularity and user rating. I do not know which statistic is a better indicator of how fun each game is, so I looked at both lists, gave each game points for appearing in each list, and added the points to determine which games have high rankings on both lists.
Metacritic
This site ranks almost every game (on every platform) on a scale of 1 to 100. When a game was released on multiple platforms, I counted the score from the platform with the most reviews. When multiple games had the same score, I broke the tie by giving higher ranks to games with more reviews.
SFGate
There are 3 critics who each posted a top 5 list, but they also posted 5 honorable mentions. This lets me treat them as top 10 lists, give 5 points to each game that is first on one of their lists, 4 points to each game that is second, etc. and then add those scores to determine which games are ranked highest by the average of the critics.
USA Today
This site had 21 different critics who each posted an unranked top 5 list. For most of those lists, I was able to simply give 3 points to each game, because they are tied for fifth place. Some critics put 2 games in the same slot, so I thought each game by itself was not fun enough for its own place in the top 5, treated them as tied for fifth and sixth, and gave them 2 points.
Video Game Canon
This site has an Exploring the Best Games From the 2000s page. It appears to be identical to the home page, but games with other release dates are filtered out, and it's outdated. I decided to not use this Exploring the Best Games From the 2000s page and instead get more updated rankings by copying the (more updated) home page and manually filtering out the games with other release dates.
Questions that might or might not be frequently asked
Q: Why don't you include a particular best-of-the-2000s list?
A: I haven't found it. Please send it to me so I can update this list. More data is (usually) better.
Q: Are the publishers credible at all?
A: Metacritic cited Buzzfeed's games of the decade rankings, which effectively set a precedent of not caring about credibility. However, I did make sure each of the publishers at least wrote a few sentences explaining why each game is fun, so I excluded stuff like this.
Q: Did you exclude any games that were listed by the publishers?
A: I excluded ports, remasters, and remakes of games from before 2000 (such as A Link to the Past's GBA port) and games that officially released after 2009 (such as Minecraft).
Q: Metacritic already found a bunch of reviews for 2000s games and aggregated their scores. Why did you waste so much time to aggregate a bunch of other lists when you can just rank games based on Metascores and spend that time finding a real job?
A: The short answer is that the different ranking methods have very different results. If I just used Metacritic, then the top 10 games would be, in descending order, Grand Theft Auto 4, Tony Hawk Pro Skater 2, Super Mario Galaxy, Metroid Prime, Halo: Combat Evolved, Grand Theft Auto 3, Tony Hawk Pro Skater 3, Perfect Dark, NFL 2K1, and Uncharted 2: Among Thieves. Only 3 of these games are also in the top 10 of this list. Evidently, the games that critics enjoy near their release dates are very different from the games that they still remember and like at the end of the decade. You're totally allowed to believe that Metacritic's list is more accurate than this list.
Q: Were the 5, 4, 3, 3, 3, 2, etc. rules based on some other function?
A: They were in fact based on y = floor(-log(2, x / 5)) + 3. It's not the simplest function, but it does its job.
Q: When a game got points from multiple publishers, why did you add the points?
A: I honestly don't know why I added them when I could have done so many other things. I don't even know why people and calculators keep telling me that 1 + 1 = 2. I apologize for choosing an arbitrary option without first considering better options, and for assuming things without proving them.
Q: Why is this list so inaccurate and useless?
A: I didn't floss enough when making this spreadsheet. Remember to floss often or your spreadsheets will turn out bad like mine.
Q: What if I have a question that's not in this section?
A: Send it to me and I might answer it.