"The intrinsic reward for knifing dudes is knifing dudes."

I waded into the debate on game achievements with my lecture at the 2010 Game Developers Conference titled Achievements Considered Harmful?, with a strong emphasis on the "?". Since the game industry seems to be careening head first into a future of larding points and medals and cute titles on players for just starting up a video game, I wanted to raise awareness of the large body of research studying the impact on motivation from various types of rewards. Trying to be "fair and balanced", I delved into what the data show and what they don't show.

Sadly, there is a contentious debate amongst psychologists about how rewards affect motivation, and I spend a bunch of time in the talk discussing this debate. Psychology is at the soft end of science, to put it mildly, and so it's easy for people—including academics—to have an agenda or opinion and "interpret" the data in a way that backs up that agenda or opinion. This is human nature, of course, and confirmation bias is everywhere in life, but reading some of the papers reminds me more of a schoolyard yelling contest than of peer reviewed research.

Highlights

To hack my way out of the thicket, I focused on the two results that both sides seem to begrudgingly agree are true.

For interesting tasks,

1. Tangible, expected, contingent rewards reduce free-choice intrinsic motivation, and
2. Verbal, unexpected, informational feedback, increases free-choice and self-reported intrinsic motivation.

I define all these terms in the slides below, and I'll fill out this page more as time goes on.

The first is the scary one, since it seems to have a lot in common with the ways games reward players with achievements and the like. I reiterated many times that there are no studies that I'm aware of on achievements for games, but if they really could be sapping intrinsic motivation to play games—and I can't see how you can argue with the possibility this is true after looking at the mountain of available data—then I think somebody should start funding research into this. I think Microsoft Research would be the perfect people to do this work, since they have both Xbox Live (meaning a great source of data, and a vested interest in figuring out the truth for the longevity of the platform) and a bunch of smart psychologists on staff. Hopefully they'll heed my call.

After talking about the results for "interesting tasks", I ask the following question when I pondered covering the results for "dull tasks":

Why are you making games?

If you’re intentionally making dull games with variable ratio extrinsic motivators to separate people from their money, you have my pity.

If you’re making intrinsically interesting games and want to make them even better, be very careful with extrinsic motivators.

Towards the end of the talk, I outline a potential Nightmare Scenario based on all the implications of the research going the wrong way for games:

1. make an intrinsically interesting game, congratulations!
2. use extrinsic motivators to make your game better
3. destroy intrinsic motivation to play your game
4. metrics fetishism pushes you towards designs where extrinsic motivation works
5. BONUS: women lose even more intrinsic motivation than men do given extrinsic motivation!

Who knows whether things will actually go this way, but it seems clear to me that the potential is there, and so we should look into this more instead of blindly moving forward.

In the talk I also address a bunch of the Common Buts:

• Players like them!
• Our data shows they work!
• We make lots of money!
• Just ignore them if you don't like them!
• They show players different ways of playing!

I go through each of these in turn, trying to address the core of the point.

Finally, I talk a bit about how to Minimize the Damage, if you're forced to have achievements and rewards in your games. As you may know, I'm working on an indie spy game called SpyParty, and since some platform holders currently require you to give away achievements to pass certification, I gathered a list of ways of implementing rewards so they do less harm:

• Don’t make a big fuss about them.
• Use unexpected rewards.
• Use absolute, not relative measures.
• Use endogenous rewards.
• Make them informational, not controlling.

Again, the data shows even following this advice reduces intrinsic motivation, but it's at least something you can do. I talk about this in more detail in the slides below.

Metrics Fetishism

I had a section here about the trend I have started calling metrics fetishism, but it was long enough that I moved it to its own page.

Materials

I usually just dump a ppt and mp3 up here, but I decided I'd try something fancier this time. After searching far and wide for a good tool for putting presentations online, I found MyPlick, which is a silly name, and the site is unquestionably ugly, but the tools for syncing your slides to audio are really great compared to all the others, and their player is simple and efficient and not bogged down with menus and whatnot.

So, here you go:

Welp, the web service I was using for the slide syncing died, so I'll have to redo it on youtube I guess.

If you're on a phone that won't play Flash, or want to listen to my dulcet tones on your Walkman/iPod/whatever-the-kids-are-using-these-days, or you just want to study my poorly constructed PowerPoint deck, here you go:

• The mp3 audio
• The ppt slides

Links and Notes

Here's where I put miscellaneous notes and links, both to articles talking about this lecture, and related stuff:

• Anthony Burch at Destructoid wrote about the lecture, as did Chris Remo at Gamasutra
• Hmm, there are a lot of links, and I haven't had time to go through them all, but you can see them here
• John Teti tried to get (and succeeded in getting) 10,000 gamer points in one day, and lived to tell about it here: Overachiever For A Day
• Soren Johnson wrote up a bunch of notes from the GDC on this and related topics, called Fear and Loathing in Farmville
• Daniel Carvalho wrote What Have We Achieved
• Alex Meer at Rock, Paper, Shotgun on TF2 Achievement Servers, Pyro Maniacs: Achieving Nothing in TF2
• There are more, but I haven't had time to go through them yet...
• As I say in the talk, Alfie Kohn's book Punished By Rewards, The Trouble with Gold Stars, Incentive Plans, A's, Praise, and Other Bribes was the introduction to this topic for me. I read the book because I have a child—Kohn is a respected name in education circles—and most of his books are about education oriented topics. I think it's an important book, and you should read it, but it comes across as pretty biased, and like most of these popular psychology books, it could be ⅕ the length. In a similar vein, more recently but with less detail, Daniel Pink wrote Drive: The Surprising Truth About What Motivates Us. It covers some of the same research as Kohn, with a few more recent studies, but then also spends a lot of time on a kind of self-help motivation take on the results, which might annoy you if you're looking for rigor. Pink also gave a talk at TED about the book. After reading some of the back and forth between the anti-rewards people and the behaviorists I was introduced to the term tendentious, which definitely applies here.
• As an example of the softness of all this research, you can start with the Wikipedia entry on Pizza Hut's Book-It! program. Book-It! is often mocked for rewarding kids with pizzas for reading books. As Kohn joked, we should really reward kids with books for eating pizzas, since we want them to eat less pizza and read more books. The Wikipedia entry cites a paper, Effects of extrinsic reinforcement for reading during childhood on reported reading habits of college students, Flora & Flora that says there was no impact on reading. However, looking closer, the Flora paper heavily quotes the Cameron meta-analysis, which has been discredited by Deci, which was questioned by Cameron, which was responded to by Deci, and on and on. As I discuss in the slides, it's really hard to know who or what to believe. This is why I use the two conclusions above, because they are the only points on which Cameron and Deci seem to agree.