[In this in-depth design article, game designer and balancer (Street Fighter II HD Remix) David Sirlin explains his methodology in designing intriguing achievement-based 'metagame' CCG Kongai for Flash game site Kongregate.]

While Jim Greer was the technical director at the casual game site Pogo.com, one of the things he oversaw was the Pogo badges system. This allows players to earn icons in various games on that site and show off their achievements to other players. Microsoft used this same idea with their aptly named Xbox Achievements. Now Jim is stealing his own idea back for his new free-to-play casual Flash game site Kongregate.com.

This time around, there's an interesting twist. When you complete the challenges for various casual games on Kongregate.com, you don't just win an icon; you win a card that's part of a metagame tying the whole site together. Jim asked me to design that game -- Kongai, which is currently in a closed beta and will debut soon. I told him there are many pitfalls in this idea and I could think of at least one hundred ways to do it wrong. Jim asked how I'd do it right.

The first thing that came to mind was avoiding the style of game where we have artificial rarity for a few very powerful cards. There are going to be some players who get caught up in the fun of collecting cards, and there will be others who actually want to play this metagame.

For that second group, I want them to have relatively easy access to all the cards. This doesn't conflict with Kongregate's business plan because the challenges aren't meant to be incredibly hard -- they're meant to be interesting enough to bring the average player back to the site.

We could make some cards extremely hard to get, but only if they have no gameplay differences from the easier-to-get version. For example, a very difficult challenge might get you a different border on the card, or different art, or a different icon for the edition of the card.

Along these same lines, I wanted all the cards to be approximately the same power level. I'm aware of Mark Rosewater's stance over at Magic: The Gathering that there should be a lot of bad cards on purpose to give players the fun of not choosing them.

Mark is brilliant, and I love his work, but on this point I disagree. I like the Guild Wars philosophy -- as you gain more cards (or abilities in that game), you are gaining the ability to create a wider and wider variety of decks, but not more and more powerful decks.

Another thing I wanted to avoid was a game that required a lot of cards to play. So-called constructed decks in Magic: The Gathering have 60 cards, but winning 60 challenges on Kongregate.com just so you can try a new deck would be way too hardcore. Even if we gave you 60 to start, winning 60 more to make a totally different deck is way too many. I wanted a game that could be played with relatively few cards.

With these ideas in mind -- no too hard to get cards, no intentionally bad cards, and small deck size -- I needed to actually create a game. I had several candidates, not to mention three other card games I was already working on for my own amusement, but one idea rose to the top: Pokemon Netbattle.

Years ago on my own site, Sirlin.net, I created a thread asking people to name a game that satisfied my long list of requirements for a good competitive game. I could not really think of any game that met them all, so I asked my readers. An unusually high number suggested Pokemon Netbattle, and it took them a while to get through to me that they didn't mean the Pokemon trading card game, which is a totally different game.

What they meant is the turn-based battle system that's inside all the Pokemon role-playing games on GBA and Nintendo DS. The fans of the game extracted the combat portion only, including all equations and stats, and created a PC online version that removes all of the RPG elements.

Even though I do not have much experience playing this game, when Jim asked me for a metagame for Kongregate.com, I remembered all the good properties of Pokemon Netbattle, and it seemed like a good fit. It has good strategy, requires only one action from the player per turn (simple), and has small decks of 6 cards.

Here's a quick note on "good" versus "new." If I copied the Pokemon game exactly, it would not be new. I personally don't care at all about new, I only care about good. Besides, this game would be new to the vast majority of our audience at Kongregate because most people are not familiar with Pokemon Netbattle.

That said, I decided to make lots of changes to how the game works, but they were all in the interest of creating a better game that's easier to learn. None of the changes were made for the sake of being new.

I'll tell you how the game actually works, but first I'll list the major areas I changed from the Pokemon game:

Most of you probably have no idea how these games work in the first place, so first, here's a quick explanation of the Pokemon game. In that game, each player has a deck of six characters. Each turn, each player makes only one decision, but it's done in a double-blind simultaneous fashion. Your only choices are a) do one of your character's four possible attacks or b) switch to one of your remaining characters.

So you might decide to do your attack A and your opponent might do his attack C. Then these choices are revealed simultaneously. The faster attack hits first, then the slower attack hits second. If one player chose to switch out, then his incoming Pokemon will get hit by the enemy's attack. When you lose all your characters, you lose the game.

These mechanics are very simple, but learning the game is actually not simple, so there was a lot I thought I could improve on. To understand this, let's look at how a Pokemon player should go about deciding what to do during his turn. Why choose Attack A over B? Why switch characters instead of attacking?

A lot of the strategy of the game comes from a concept called resistance. There are 17 attack types in Pokemon (such as fire, grass, psychic, dragon, etc.). There is a 17 x 17 chart telling you how good each attack type is against each other attack type. In some cases, you'll do double damage, others you'll do normal damage, others you'll do half damage, and in others, no damage at all.

If you understand and internalize this chart, you can use attacks that are very effective against your enemy's resistance, and force him to attack with attacks that are weak or have no effect because of your resistances.

On top of all that, each move is classified as either a normal or special attack. Your character might have good resistance to normal attacks but is weak to special attacks, in addition to the layer of resistances provided by the 17 x 17 chart. When both players understand this, it leads to interesting mind games. If your current active character has a very favorable matchup against my currently active character, we both know I'd like to switch out. But will I?

Because of the metagame (players know which Pokemon are popular and tend to be in decks), you can probably guess which exact Pokemon I'd like to switch to in order to counter yours. If you're really clever, you'll do an attack that's strong against the Pokemon I might switch in (that you haven't even seen yet!) instead of doing the obvious move of a strong attack against my current active character.

The problem is that none of this is interesting at all until both players have internalized an unreasonable amount of data. Ironically, little kids have a better chance at this, because they have been exposed to so much Pokemon media (games, tv show, movies, card games) that they already instinctively know whether grass-type beats bug-type or not (nope) and whether dark-type beats ghost-type or not (yep). We need a simpler system that doesn't require a 17x17 chart.

I decided to go with only three attack types: physical, light magic, and dark magic. Each character has three resistance numbers, one for each attack type. I also got rid of all complicated math that goes on behind the scenes.

If you attack with for 15 physical damage and your opponent has 4 physical resistance, then you will do 15 - 4 = 11 damage. Very straightforward. Note that it doesn't matter how much or little resistance the enemy has to light or dark magic when you do a physical attack.

This makes the game much easier to understand, but it removes too much strategy. There isn't enough "play" in dancing between three attack types as there is in dancing between 17. I needed to create more nuances. The first was the concept of multiple-hitting attacks. The rule is that resistances are subtracted from *each hit* of a multi-hitting attack.

For example, if an attack does 10x4 damage (that's 10 damage four times in succession), then a resistance of 3 would make it do (10-3)x4 = 28 damage. But if that same attack had been a single hit for 40, then a resistance of 3 would only take it down to 37 damage. So all things being equal, you'd rather do a single hit for 40 than a multiple hit that adds up to 40, because the single hit is less susceptible to the enemy's resistances.

Pokemon Netbattle

There are a couple other factors to consider though. Each character has a health meter and an energy meter. The energy meter is similar to a Rogue's energy meter in World of Warcraft. It holds 100 points of energy, starts full, and refills quickly (20 points per turn). But moves cost energy to perform, so now you have to consider not just how much damage a move does, but also how much energy it costs.

As part of the basic game design, it costs less energy to do a multi-hitting move for X damage than it does to do a single hit move for X damage. So the best possible case for you is if you fight an enemy with, say, no resistance at all to your multi-hitting move. You will then get to do all damage from every hit, and you didn't even have to pay the higher energy cost of a single hit.

If the enemy *does* have resistance to whatever type your multi-hit move is though, it will probably be very ineffective. You'd be better off paying a bit more for a single hit move that isn't effected much by resistance.

A second thing to consider is that any bonuses you have (such as +1 damage) apply to each hit of your moves. So a multi-hit move can be powered up much more than a single-hit move. The point is, this system creates several nuances, but all of them are governed by straight arithmetic and no elaborate chart is needed.

Character Switching

One of the major changes I made was the character switching mechanic. As always, when you switch characters, you gave up your chance to attack that turn. But Kongregate's game needed more ways for you to maneuver around attacks. Without the vast design space of the 17 x 17 chart, you needed some extra ways to avoid stuff when you know it's coming.

This is why in Kongregate's game, switching characters lets you completely avoid all damage from your enemy's attack. If you know they will attack, you can make them waste the energy they paid to attack, and make them deal zero damage. Of course, they'll need a counter to this if they know you will switch, which is why I added the new mechanic called intercept.

Intercept does nothing at all if the enemy attacks--you just get hit. But if the enemy switches characters, your intercept will prevent the switch and deal 35 damage, a huge amount. That means your opponent skipped his attack (because he chose to switch characters instead), he doesn't get to switch, and he takes a huge amount of damage.

This intentionally creates a game of paper, rock, scissors with highly weighted outcomes. If you have an opponent down to very little life, everyone knows he wants to switch out (he'll heal one hit point per turn while switched out, by the way). Or, if your opponent's character has little or no energy left to pay for moves, everyone knows he wants to switch out. So the "textbook" thing to do is to intercept him in this case.

This creates a good mind game where you have to read how crazy your opponent is. Is he crazy enough to actually attack when his character has two hit points left? Is he crazy enough to attack two turns in a row? Three turns in a row?!

So far, we have energy meter management, we have paper/rock/scissors system of attack/intercept/switch, and we have single/multi-hit attacks and three types of resistances. This almost gives the players enough wiggle room to use good strategy, but I wanted players to have one more tricky way to influence the fight: attack ranges.

Each turn, the fight will take place at either close range or far range. Each attack in the game is designated as either a close range attack, a far range attack, or both (can be done at either range). Some characters must be far to be most effective, others must be close to be most effective, and others are able to fight at both ranges.

This mechanic lets you try to change the range in order to get an advantage, but it's intentionally expensive to change ranges: it costs 50 energy points (half your energy meter) to attempt to change it.

This brings the total number of decisions per turn from 1 to 2. Now, you must first decide whether you want to try to get close (50 energy), try to get far (50 energy) or just go with the flow (0 energy). If you decide to go with the flow (which you usually will because spending 50 energy is a lot), then you're allowing the enemy to pick the range for the turn. If one player chooses close and the other chooses far, then the choices cancel each other and the range is set to whatever it was last turn.

The double-blind nature of the choice can make it a hard decision sometimes. Imagine that you are playing a character who is great at close range, but poor at far range. The range is currently close (yay!) but now you must choose which range you want for this turn.

You'd like to pass (go with the flow), so you'll get to keep your 50 energy and fight at your optimal range. But your opponent might move to far and then you'll be very unhappy. To guard against this, you decide to choose close range (50 energy) even though you're already at close range. This guarantees you'll fight at close range, because if the enemy chooses far, that will just cancel out your choice and the range will remain the same. So you choose close (50 energy).

Remember this choice is double-blind, so after you committed your choice, it's revealed that your enemy chose to pass (0 energy). You psyched yourself out into spending 50 energy for nothing. The fight would have been at close range even if you passed.

One good thing about the range mechanic is that it's visual. Your characters on-screen are either standing far apart or close together, and it's obvious which range you're at. It's also a lot easier to deal with three resistances in your head than it is to deal with 17.

It's easier to deal with simple arithmetic that you can easily compute yourself before you attack, than relying on a hidden algorithm to determine stats and damage. And finally, it's easier to actually participate in the paper, rock, scissors part of the game with attack/switch/dodge than it is to participate in the 17 x 17 version of the paper, rock, scissors in Pokemon.

And yet for all this simplification, you still have a lot of opportunity to be smart and sneaky. I've only told you the basic skeleton of the game, but there's also a lot of twists and turns added because every character has his or her own special ability (that automatically takes place--you don't have to click anything to make it happen).

Also, every attack has a chance at producing an extra effect of some sort. And finally, as in Pokemon, you can equip one item card to each character which gives him even one more automatic ability. None of these require any extra clicks from you, but they create more opportunity for strategy.

Kongregate.com's Kongai

With attack/switch/intercept, resistances, changing ranges, and automatic special abilities, you can really size-up what kind of person you think your opponents is and start to outplay them by reading what you think they will do. There's enough going on that players tend to develop patterns you can use against them. And most important of all, it's relatively easy for players to go from beginners with no clue about anything to intermediates who grasp enough of the game to develop decision tendencies.