Saturday, August 14, 2010
Not enough data points
You know the MMO blogosphere is in a summer slump when the hot topic of the week is why WAR failed. The theories range from "because it was a WoW clone" to "because it just wasn't very good" to "because they failed to meet explicit promises". I find the latter two explanations more believable, but there is a fundamental problem with the discussion: The fate of a single game, or even a few similar games, provides not enough data points to make an accurate prediction of player preferences.
Syncaine thinks that World of Warcraft is a "unique snowflake", a perfect storm of unrepeatable circumstances. But if we believe that, than why should we believe that all other games are not "unique snowflakes" as well? Why would it be a bad idea to design a game based on the success of WoW, but a good idea to design a game based on the success of EVE? I'm pretty certain that if game companies decided to produce a bunch of half-baked EVE-clones, the result would as much a failure as the fate of the half-baked WoW-clones.
I don't believe that you can reduce the success or failure of a game to one single point. Certainly EVE is doing quite well, but is that *because* it has free-for-all PvP, or *in spite of* having free-for-all PvP? According to CCP's official own data, over 80% of the EVE players never ever leave safe empire space, so drawing a straight line from EVE's success to the potential success of future PvP games is somewhat spurious. Neither do I believe that you can take a single feature of WoW, lets say it "being easy" or "soloable" and pin all of the success of WoW on that feature. If anything, the relative failure of "WoW clones" shows that even game designers are unable to find the secret sauce recipe of success of a game and then copy it, even if given millions of dollars to try.
I do not believe that most players play a specific game for one specific feature and reason, but they tend to look at the game as a whole and decide whether they like it or not. They might *buy* a game because of hype, but they certainly aren't going to play it for long because of hype. For example I would say that I don't like first person shooter games, but I still try them from time to time, and occasionally I find one I like and play through, like Call of Duty or Bioshock. Even my MMORPG preferences aren't that easy to categorize: Yes, I like WoW, but I didn't like a lot of other games that were similar to WoW, for example LotRO. And yes, I didn't like hardcore sandbox EVE very much, but I do like A Tale in the Desert, which is more hardcore and more sandbox than EVE, albeit in a different way. And when I quit playing WAR, it wasn't because of PvP, but it turned out that PvP in WAR was still the part I disliked the least. It would be difficult for me to take a list of features from an upcoming MMORPG and accurately predict whether I will like it or not.
If anything, the best predictor for me liking a game is quality of execution. Again, that isn't an absolute factor, A Tale in the Desert is not a very polished game by any means. But when I pre-ordered Final Fantasy XIV without having even played the beta, or when I'm certain that I will buy Star Wars: The Old Republic, it isn't because these games have this or that feature, but because I have a certain trust in the companies making these games that their games will be reasonably polished.
Now I don't want to claim that "I think this way, so everybody else must think the same way", which is a typical fallacy on blogs. But that is exactly why the whole "game X succeeded/failed because of Y" discussion should be taken with a grain of salt. In most cases the reason proposed reveals more about the author of the argument than about the game he is discussing. If anyone really would be able to find a specific reason for the success or failure of existing games, then why are game companies obviously unable to learn from successes and avoid failures? Talk is cheap, developing a MMORPG is expensive, so a game developer should be not only more qualified but also more motivated to find the reasons for success and failure, yet obviously they can't.
Any single MMORPG has such a huge amount of features and characteristics, that it would take statistical data mining of thousands of MMORPGs to measure the popularity of each single feature, if that was even possible. As mushy as it is as advice, the best approach is still just trying to make "a good game", without even trying to identify what worked and didn't work in previous games. MMORPGs are a greater whole, larger than the sum of their features, and the features of a new game fitting well together is a far better approach than putting in a feature just because "this was popular in WoW".