Gamification, random reinforcement, and “the metagame”

The following is in response to an article titled The Current, and Unfortunate, State of Gamification. It was written by my work acquaintance Dakota, who is a relatively new colleague of mine (though I don’t work directly with him). We have interesting conversations about board games, and he was the one who linked me to David Sirlin, (discussed in my last post). If you don’t go read his article, you’ll need to at least know that gamification (in this context) refers to enhancing websites with “game-like features”. I think Dakota’s point is generally that sites aren’t being creative enough in how they add gamification, and generally just add badges and leaderboards, but he touches on a lot of interesting topics while making it.

I’ve been thinking a lot about this article since I read it. Only a few days ago, so I’m late to the conversation, but this blog is stale, so I don’t feel bad posting on an older article.

Anyway, one point I’d make is that, as David [one of the previous commenters] touches on above, the definition of “game” is quite loose. Just as peek-a-boo is a game, full of delight for the extremely young mind (and a game that arguably evolved into the “hidden object” genre that some adults play), so is “collecting” a game that toddlers and adults play. Lots of people refer to collecting badges and getting highscores as the “metagame”. For many, (myself included), it’s not that I find a new game less interesting (far from it) it’s just that a new game is one whose mechanics I might end up taking or leaving, while the metagame is one I already know and love. The metagame is a game that we play in (almost) every game, and (increasingly) in every aspect of our lives! To bring back your ketchup metaphor, [Gamification was compared to food condiments, Dakota argued that its not required for a gourmet meal.] some people just LIKE ketchup, and they put it on everything, breakfast eggs included.

The heart of gamification for me is competition. Messages like “Your friend XXX posted XXX minutes ago” can be a reminder that they are playing against you. To me, even the metagame isn’t nearly as interesting if you can’t compare yourself to your friends and (at the very least) hordes of strangers. (Obviously, this is what leaderboards are about.) The harder it is to see your competition (who else has the badge that I just got?), the quicker you loose the hard-core metagamer.

I do take umbrage with comparing gamification to the Hawthorne Effect. The Hawthorne Effect is increased productivity due to external observation (and I’m no psychology expert, so I have no opinion about whether its effects are debatable). It’s possible that gamification causes people to “feel observed”, but I would argue that gamification is much more like a simple feedback loop, the psychological effects of which are not debated. (The concept was popularly lauded by this wired article in one of the last few issues.) If the Hawthorn Effect is external observation, feedback loops are internal observation (or self observation). One racing game example would be seeing “ghost” images of your previous laps so you can compete against yourself. I think this translates into knowing that you’ll get a badge after one more check-in, or more instantaneously, seeing the counter of how many characters you have left for your status messages on twitter.

This segues nicely into a rant about Random Reinforcement… (I thought I could just link to this, but amazingly, I haven’t immortalized this yet on the interwebs). Essentially, animal trainers know (and research probably shows, although I honestly haven’t looked too much into it) that using positive or negative reinforcement is not nearly as effective as randomly applied positive reinforcement. (Looks like Wikipedia calls this concept “variable rate” (VR) reinforcement. If you follow that link, the graph on the right sums it up pretty nicely.) The rational that I remember is that the animal (or person) is encouraged by the reinforcement they do get, but because they don’t get it every time the effect is that they try harder, thinking the fault somehow lies with their behavior. (The “rant” that I have been known to make is about how this is why people remain steadfast in relationships with partners who treat them badly, but I’ll save that one for later.) Anyway, I followed Kathy Sierra’s horse training link, and I’ll admit that it definitely made me want to read Daniel H. Pink’s Drive. (I’ve added it to my amazon wishlist.) I can very easily imagine that “autonomy, mastery, and purpose” are HUGE motivators, but I don’t see them as in opposition to the concept of gamification at all. In fact, I think there are direct mappings from leaderboards to mastery, and from achievements to purpose.

But let me step back for a minute. It should be obvious by now that I disagree with at least part of the original Hide & Seek article. Specifically the premise that “neither points nor badges in any way constitute a game” seems pretty wrong to me.

I knew it sounded familiar, and after reading through the comments, I realized that I posted a pretty similar sentiment in the comments over there! Here it is for posterity:

Definitely agree with the your semantic argument here. “Pointification” is a better word than “Gamifcation”, and they are definitely different things, but I do not agree with the premise that points != a game.

Remember the meta-game! What about someone who buys and plays through an XBox live game they don’t even want to play, just to get those achievement points? Aren’t they playing a game? I think they are.

What Margaret (at Hide & Seek) is saying seems to be that we should remove all metagame elements from the definition gamification, and just talk about how you can modify the “core mechanics” of your website. (No more ketchup, only entree.) And maybe there is something there in the user experience realm that can approach gameplay. The feeling I get when browsing kickstarter, for example, looking for that next project to back… it’s akin to playing a game, maybe. But I’d argue that a site like Facebook is almost no longer applicable at that point, since so much of the facebook experience is the metagame. Maybe all of social media is the metagame. I know it feels like a game to a lot of people. We just need to remember that, like investment banking, it’s a game with real consequence.

2 Responses to “Gamification, random reinforcement, and “the metagame””

  1. Kathy Sierra Says:

    So much to think about with this post… thank-you for that.
    I’ll start with this: with animals and clicker-training (purely positive reinforcement using a variable ratio schedule), one reason people often don’t find the problem I did is because the animals are playing their own kind of meta-game: the Get The Human To Do X game.

    The joy for most dogs and most calm horses becomes far less about the reward itself and much more about creating behavior in the humans (“what can I do to get them to press that clicker button?” “what is the least I can do?” “What new thing can I do that will produce the *jackpot* (the extra-large reward given only for very special efforts)?”

    With the horse in my video, two things were happening: I was using clicker-training (I.e. Gamification) to reinforce what *should* have been an intrinsically motivating behavior (as opposed to, say, a tedious task with no potential for intrinsically-rewarding feelings associated), AND this horse was not interested in the meta-game. He didn’t find joy in figuring out what I wanted him to do. So he quickly became less and less motivated (self-determination theory explains this quite well) with ultimately dangerous consequences.

    Today, this horse is unrecognizable to those who knew him during that time. He is engaged and energized and spirited… all which happened when I stopped “rewarding” him for being engaged, energized, and spirited and used those principles of autonomy, mastery, and purpose. It has now been just over a year since I saw the “error of my gamification/Skinner box ways”, and the changes have been lasting and richer as time passes. He now is even more proud and willing than when I made that video.

    I do still use extrinsic rewards for the things he would not otherwise EVER want to do, and those pose no problems. All upside, for him since there is no danger he will lose motivation. But I am extremely cautious using it for things which could/should be motivating.

    I fear, deeply, for the ways in which extrinsic rewards are being applied to education in a well-meaning but ultimately dangerous attempt to “increase motivation and engagement for learning.” If it merely had a short-term boost and then eventually it stopped working and things returned to their pre-rewarded state, I’d have no cause for alarm. But it has the potential (demonstrated over decades in more than 100 different studies) to leave learners LESS motivated than they were before extrinsic reward structures were put in place. Just a few of the specific examples:

    *kids given ribbons for drawings began to draw less than those not given rewards
    * writers rewarded for poems began to write less sophisticated poems
    * monkeys that enjoyed solving wooden puzzles began solving fewer puzzles (and making more errors) when a treat reward was given for solved puzzles

    And on and on.

    These results seem counter-intuitive, so it has been tough for people to recognize the harsh and lasting implication of applying extrinsic rewards *where there is or could be intrinsically rewarding activities*. Further, the fact that so many people and animals DO discover and enjoy a meta-game around extrinsic reward structures makes the problem even harder to spot: “but look how engaged and motivated they are?!” Yet if we look more deeply, we find the meta-game they are enjoying is producing engagement around the reward structure and not the thing-we-wanted-them-to-engaged-with. In many cases, this distinction simply does not matter. They are active and appear motivated, so who cares what’s driving it?”. But the difference can be staggering, because you end up with phoned-in behaviors for the thing you care about, masked by excitement around the meta-game.

    Marketers in it for the short term, especially consultants, will get a lovely spike in engagement and sometimes a long-term meta-game spike. This makes everyone happy, sort of. But it produces no lasting effect of producing deep engagement, and in some cases can drive people to now care LESS than before about the actual thing you wished they cared about.

    You can see it in the way marketers will often point to Nike+ as an example of success, forgetting the fact that people actively WANT to exercise more, and that for most people running is itself *not* an intrinsically rewarding activity. In other words, it feels good to HAVE run, but most never get the endorphin rush that makes it feel good in real time. So extrinsic motivation is pretty much all upside where exercise is related.

    Oops, sorry for the novel-lenth comment…

  2. Martin Says:

    Kathy, thanks for your comment, lots to digest here! (No apologies necessary.)

    As a father of a 1 & 1/2 year old, rewards and motivation are very relevant right now, so I appreciate your insights. (There will be no gold stickers in my household! …at least, not for tasks I want her to like. They might be helpful for getting her to do chores though.)

    It strikes me that a book my wife and I are reading is probably relevant to this discussion, Nurture Shock: New Thinking About Children. It outlines a bunch of ways in which conventional thinking about child development and rearing have been discovered to be… if not wrong, then at least flawed in some way. I went and looked at the index, half expecting there to be a chapter that deals with this stuff. One of the later chapter titles suggest at least a tangential relationship, but I haven’t read that one yet. Anyway, it’s been well worth a read so far, and frequently comes up in conversation.