Some playtesting goes a long way to making your product rise a grade or two above how you might design it. For all the clever ideas one might put in, seeing how people react to it can help tune the game to be an enjoyable or exciting experience, or a frustrating evoker of cuss words that never gets recommended to friends. I picked up LOTR:ROTK by EA the other day, it was in the $20 bin. After playing it for a few hours, I have to wonder if they playtested it at all beyond basic function. Perhaps the single most annoying thing is cut scenes that come up mid fight, in some episodes continually, which cause a fight you might otherwise have won to go badly for when it goes back to game, positions have changed & other things you were unable to react to. The second biggest one was camera angles views totally obscured by intervening explosions, smoke etc. During those times you are unable to fight as well. Drawing the camera in close, or being able to move its angle would have helped. After excercising my longshoreman''s dictionary many times, I have to conclude there is no way they observed how player''s reacted to those aspects, if they had, they would have changed them to be less disruptive of what would otherwise be a fairly exciting combat game. Then again, that is probably why I got it in the $20 bin. For the quality of your game, and return of your investment - playtest, observe people''s reactions to situations in your game. If they cuss, scream, punch their monitor, chances are there are changes you need to make unless they are just poor losers. Losing a game to things beyond your control though is a situation most people will avoid the second time around.

Deus Ex 2 would be a recent example of a lot of hard work fucked in the ear by poor quality testing. The mistakes they made with the PC version were so elementary, members of the community would be patching them up and providing workarounds within hours of something being found. Unfortunately, this doesn''t help the majority of people who buy the game without attaching themselves to the online community surrounding it.

I recommend having a brick on hand to throw at publishers who insist you have your game shipping the moment it''s in what they consider to be a ''complete'' state, because it has the potential to completely obliterate years of good work.

While I can think of many games that were shipped broken, the worst, for me, was Ultima 7. There was a "key deletion bug" in the game where if you camped and went to sleep unique keys needed for quests would just disappear from your backpack.

This actually made it impossible to finish the game (I think Ultima 9 shipped like this too, though at least now you have the internet to download a patch from).

It was obvious that no one had ever played the game from start to finish, without using whatever "cheat" codes were built in for testing.

As a professional software developer not in the game industry, I find the lack of proper testing in games abhorrent. I wonder if any game studios follow the "unit test first" methodology? If you''ve never heard of this before, it''s where a developer, after creating a detailed design document, creates a unit test script for what they are working on before ever writing a line of code.

I''ve heard arguments before about how it is impossible to automate testing for games, or how testing is vastly different for games as compared to other commercial software. For the most part I think that''s bunk -- with a proper design you can setup automated testing for many areas of gaming (for example, your pathfinding routines can be used in a special test program where the results can be compared against expected results).

I doubt this will ever change though...I doubt the number of people moving into games development that have prior non-game development experience is large enough to ever influence coding practices.