How is an old-school D&D wizard like two marshmallows? How is an old-school D&D fighter like one? And what can four year olds teach us about delayed gratification, reasonable expectations, and game design? Read on, dear reader, and we shall see together that timescale and trust are two of the most powerful factors in tabletop RPG gameplay.

Waiting For Superman
There is a promise in any game with advancement:

Someday, you will be great.

Regardless of the starting point, be it rat-catcher or superhero or demigod, there’s always greater power just around the corner, just out of reach. It’s a carrot of RPG gameplay: Keep playing, the game whispers to you. Keep going. Look at how awesome your future will be.

That hope should be a powerful motivator, a nice carrot at the end of a long stick. That’s the logic that went into early designs of things like the D&D wizard, after all: Suck now, rock later. Put up with d4 HP and one magic spell that might suck, and soon you’ll be granting wishes and traveling the cosmos. All you have to do…is wait.

Sounds easy enough in theory. But in practice, our patience is perhaps not what it might be. There’s a famous psychological experiment done at Stanford University in the early ‘70’s about delayed gratification that might help show us how shaky using it in game design might be. I give you, the Marshmallow Test:


The test is fun enough by itself (who doesn’t love tormenting toddlers?), but even more interesting for its follow-up years later, where there is evidence that the patience in some toddlers is correlated with other positive indicators in grown people, such as higher SAT scores. Turns out, that being able to deny yourself pleasure to reap greater rewards later is a pretty useful human skill in this society (this guy calls it “self-discipline”).

That self-discipline is a skill that RPGs can help players to practice and learn. Today you suffer. Tomorrow you dominate. Have confidence, have faith, believe in the future, and chase it. That's the essential promise of the old Wizard class: practice self-discipline, and be rewarded later. Fighters were for people without that self-discipline: you can be good now, but you'll lag behind later.

And's a pretty exceptional D&D player who thinks that we should go back to 2 hp PC's who eventually become demigods, or who thinks that linear-warrior/quadratic-wizard is a positive thing. If the Marshmallow Test shows us that practicing self-discipline is constructive, even positive, why does the game yank us back from that in practice?

To answer that, it’s worth looking at how this concept of "rewarding self-discipline in PC classes" works in our actual games. It's true that to a certain degree, any gained level works this way. Even if a Wizard has more to gain from a level than a Fighter, the Fighter still gains things -- every level is a carrot held out there for you to pursue. Everyone gets more marshmallows, the people who opt to wait the longest just get the MOST.

Or is that really true? While D&D still maintains an advancement of levels, the idea that you actually gain improvement over levels has been weakening, mathematically. In 4e, Level 30 looks a lot like Level 3, just with bigger numbers. The odds remain roughly even at all points. Call it the “flat curve” effect: rather than have a lousy chance at level 1 and a great chance at capstone level, we have an even chance all the way through. Take a look at an average first level 1e thief’s chance of hiding in shadows, and take a look at the average first level 4e thief’s chance of doing the same, and you’ll see that the 4e thief is clearly better, as a matter of a percentage chance. Compare them at 16th level, and now the 1e thief is hiding much more often than the 4e thief.

This has been heralded as a positive thing, in general. It certainly sounds nice on paper, right? A 15% chance to hide in shadows is absurdly low, but a 95% chance to hide in shadows is absurdly high, so to fix both problems, you flatten the curve, and make it closer to a 50% chance at all levels. Now you've got 1st level rogues who don't have useless abilities and 15th-level rogues who don't have overpowering abilities.

Only, if the idea of D&D teaching self-discipline is accurate, that’s a disaster! D&D has gone from teaching you to endure (and in enduring grow strong) and has gone toward giving you instant gratification. Everyone gets all marshmallows all the time! The levels become a treadmill, measuring time spent playing, not offering you power! Having a +5 vs. a DC 15 at level 1 and then a +50 vs. DC 60 at level 25 becomes fairly meaningless. You may as well take out the math and just flip a coin!

Well, maybe. If we stop judging kids these days with their rock and roll music and their dragonboobs and whatever and start looking at causes and effects…we might see a different story.

Endless Summers Don’t Last Forever, Kid
Say, you run your game once a week. Maybe you meet on Saturday afternoons for about 4 hours (you play D&D, order pizza, hang out a bit). You meet pretty regularly, but holidays and stuff get in the way, maybe you meet about three times a month.

Say, your group levels up “when you feel like it.” When you feel like it happens to be about once a month, when the players get restless and you start getting bored of the same abilities over and over again.

At that rate, it’ll take you one year and eight months to see 20 levels of the game That’s 240 hours, or about enough time to see all the extended cuts of all the Lord of the Rings movies twenty-six times. That’s assuming you don’t lose a month in December, and nobody goes on a two-week vacation, and nobody gets board of their dwarf fighter and wants to start over at level 1 again, and that you don’t spend a few extra months at level 7 because the DM wants to hold off on giving levels until a plot point is resolved.

In other words, it takes a lot of time to get to those high level marshmallows. When you start off your character at level 1, all the promise of that high-level goodness might not get delivered on. People move. Get busy. Have kids. Discover alcohol or romance. Life events happen. Games change. It requires a fairly awesome alignment of the stars to have a single game going for 1 night, let alone on a regular basis for a year and a half. Those promises of power get broken. You never do get to see your awesome high level abilities, because your DM got a new job and passed off responsibilities to someone else who started a brand new game in Eberron where everyone starts again at level 2. Le sigh.

This has a bearing on the marshmallow test. When promises get broken, the rationale for eating your marshmallow early changes:

[ame=]The Marshmallow Study Revisited - YouTube[/ame]

It turns out, those kids who eat the first mashmallow aren’t necessarily impulsive or short-sighted or greedy. It's not that they lack self-discipline. Rather, they lack trust. Eating the marshmallow is a rational act from their view: the future is uncertain, adults say things they don't mean, and there is certainly this thing here in front of me now that I want.

In game design, we can see the same reaction. The promise of high-level powers gets broken regularly. Very few campaigns actually see even 10 levels. Campaigns end unexpectedly, and the longer a campaign goes on, the more likely SOMETHING is to get in the way. So, rationally, the players want it delivered on sooner, and better. They’ll take a 50% chance now and forever in exchange for a 15% chance now and a hypothetical 95% chance later, because chances are good they’ll never even see later. That 50% chance will last for a few months, and then they'll start over again. Better to have that now, than to trust that lives won't change for the next 18 months.

So you can see that far from catering to laziness and instant gratification, changing the game so that early-level characters can actually do stuff is more about the broken promise of higher levels. People rarely see the upper echelons of those levels, and only infrequently inhale the rarified air of high-level play (at least, not without starting there).

You can see this at work in the design of the game. High level play has been, throughout D&D, unbalanced or at best exceptionally unwieldy. Thanks to the theory of rewarding self-discipline, spellcasters begin to dominate. Even if that doesn't happen, options and abilities explode and cascade out of control. High-level play simply doesn’t get as much play time as levels 1-5 or 1-10 do. Campaigns don’t often hold together for longer than that, even if the group does. Rather than 1 year, 8 months, you get more like 3 months. 6 months. 9 months. Certainly not enough time to hit 4e’s 30 levels in anything other than a frickin’ rocket ship.

Which means all those promises – those paragon paths, those prestige classes, those epic destinies, those high-level spells – are broken. You don’t actually get to play with them. If you played a wizard, you sucked for 3 months, and then we started over again because we swapped DMs. Whoops.

In fact, this is one of the phenomena at work in the ages-old “wizards vs. fighters” debate. One side might claim wizards are overpowered because just LOOK at what they can do with their high-level magic! The other side might claim they’ve never seen this overpowered wizard in all their years. Neither side realizes that a high-level wizard itself is a rare beast at any actual table, one just argues from the RAW (which is rarely used) and the other argues from experience (which is hardly universal).

In fact, in talking about experience, it's useful to point out that not all tables fall apart so fast. Some last for long periods of time, and play about in the high levels on occasion, a natural outgrowth from their regular meetings. Most groups who meet that regularly often start over again at level 1 on a fairly regular basis anyway, but not all. I think it's fairly safe to say that those groups who manage to regularly hit high levels are quite the exception to the rule, but it does happen, and there, the promise of high levels is kept, and self-discipline is rewarded. Those games are at one pretty dramatic end of the bell curve, though.

Challenge: Accepted!
So we’ve seen that a game with delayed rewards can have a rewarding psychological experience. It can teach you to make due with inadequacies, trust the future, and reach for the stars. But we’ve also seen that this is only true if the game actually delivers on these promises, rather than delaying them indefinitely. If the game never delivers on those promises, people will grab what they can when they can.

For your own games at home, you can take a few details into account to use this idea to make your games better, instead of letting it hurt your games.

First, you should figure out what the minimum time you need to play your game is. You should have a comfortable, natural end-point in mind, before you even start playing. Ideally, you'd also have a good idea about how long it would take your group to get there. How long does combat take with your groups? How much do they actually get done in a night of gameplay? Can you reduce your sprawling quest web down to a single adventure? A single night? A single encounter? Do it. The shorter and sweeter the experience, the more likely your players are to get what you promise them. It’s always easier to do a sequel to a successful adventure than it is to get people to sign up for an entire 20-some-level Adventure Path. Rather than write Nabakov, do an episode a week, one week at a time.

Second, you should keep in mind how much you want the characters to advance while they’re playing. Do you want them to experience all 30 levels of your 4e grandeur? Is that realistic in the time allotted? Could they maybe do it over one level? Or even less? Maybe they should only expect magic items at some point, and not any actual XP, since it’ll be real quick. Share that information with your players. Let them know how big they can dream. That way, they can make educated choices about what they want to experience. If you're only playing for 5 levels, they won't bother picking out prestige classes for their characters -- they'll never see it.

With those two things in mind, you have the ability to promise them marshmallows you can actually deliver – to actually give them the experience you claimed this game would have, and to avoid failing to deliver on your promises. Then, you can make some use of a limited kind of delayed gratification. The players can actually advance in a small stretch. They can fantasize about being able to use abilities they actually get to use. This can encourage self-discipline, if only in small doses. But if a toddler can wait 10 minutes for two marshmallows, maybe you can wait 3 months for Fireball.