a series of interesting choices thoughts on game design from paul sottosanti


the shields system of bit pilot

The goal of Bit Pilot is to stay alive as long as you can. Well, scratch that. The goal is to listen to as much of the music of Sabrepulse (an incredible chiptune artist) as possible, and the way you do that is by living as long as you can. You control a little ship in a huge asteroid field, with your only input being surprisingly tight touch-based movement controls. There are two powerups that occasionally float onscreen; a score powerup that grants 1000 points, and a shield powerup that grants 250 points and adds a hexagonal shield surrounding the outside of your ship. Each shield will block one asteroid that otherwise might kill you, but there's a catch: each one also makes your ship significantly larger. With no shields, you can nimbly slip through the tiniest cracks, but you're also instantly dead to any laser or asteroid that makes contact. As with all games of this ilk, death equals starting over from the beginning.

Recently I tweeted: "Bit Pilot's shields system is one of the best and most elegant examples of dynamic difficulty I've seen. Simple but brilliant."

Greg Marques requested that I go into more detail, and seeing as it was just recently his birthday, who am I to turn him down. Here's why I made that claim:

It keeps the game interesting for players of all skill levels. The basic goal of any dynamic difficulty system is to provide a suitable challenge without forcing players to self evaluate their skill level. (Since players tend to be fairly bad at the self evaluation, especially without enough context, and also are often incentivized to choose something far too easy in order to achieve unlocks or higher scores.) The system works admirably in this respect, as a strong player can build up a huge stockpile of shields, making it difficult to weave in between even the smallest asteroids.

The difficulty increase is visual and directly player driven. A common problem with dynamic difficulty systems is that they're hard for players to comprehend. Once a player knows that something is going on behind the scenes but doesn't understand it, it's easy to feel cheated and blame the dynamic difficulty system for providing artificial challenge. In this case, though, it's easily understood because it's carried out in a visual way directly at the point of the player's greatest focus, the ship.

It works towards solving a common problem with the genre where the early game is boring for strong players. I have a hard time playing any of these "survive as long as you can" (Canabalt, RunRunDash!, Monster Dash) games for long because the game experience gets worse the more skilled the player becomes. Not only does each individual game start to take longer, but there's also a steadily growing period of boredom at the beginning as you wait for the difficulty to ramp up. While the shields system doesn't completely solve this problem, it does mitigate it greatly, as you can now spend the early game trying to build up a reserve of shields to get you through the harder later stages, and manuevering with those shields is tricky.

It turns getting hit from a negative experience into almost a positive one. This is a subtle effect that I think is actually quite important. Typically in these sorts of games, getting hit (or falling off a building, or whatever) is a strictly negative experience. There's nothing wrong with that, and in fact the whole game is built around avoiding it. In Bit Pilot, though, the fact that your ship suddenly becomes smaller changes the feeling completely without doing something drastic like removing challenge or consequence. So many times I've gotten hit and found myself thinking, "Sweet, I'm smaller now!". Most of these games have an arc that just slowly builds linearly and then eventually plummets off a cliff when you make a mistake. Bit Pilot's is more like a traditional roller coaster, with a steep initial climb and then numerous peaks and valleys.

There's no question whether or not the shields are worth it. Another common problem with dynamic difficulty systems is that they will drastically warp player behavior. Oblivion is a famous example, with players intentionally staying low level while training critical skills and trivializing much of the gameplay. However, in Bit Pilot the shield powerups are obviously worth picking up, since an extra life is always worth having no matter how difficult the gameplay becomes (assuming it drops back down to the old difficulty when you lose the life, which in this case it does). I haven't found the system changing my behavior in negative ways at all.

So there you have it. An incredibly simple design choice (each shield makes you bigger) with far reaching positive effects for the game. If you get a chance, try out the game and let me know if you agree or disagree.


the seven pillars of drafting-based games

Some time ago I played Fairy Tale for the first time. It's a game designed by a former professional Magic player named Satoshi Nakamura, and it's based around drafting, which is a format that has fascinated Magic players since the game's creation. Drafting is the perfect mix of luck and skill, with infinite replayability because every draft presents wildly different situations and challenges. In fact, there are sites online where you can draft Magic over and over against AI opponents that learn over time from the picks made by humans. The only problem with these sites is that, as fun as drafting is, it loses something when there's no validation at the end of the draft as to how successful you were.

Fairy Tale takes a stab at providing a more casual drafting experience with some quick resolution at the end of each round and a simple victory point system for determining a winner. It's a fun game, but after playing it a couple more times a month ago I felt like drafting games deserved more exploration. The problem was, I didn't have a great sense of how to make drafting compelling as the main feature of a game. In both Magic and fantasy football, it's used simply to distribute resources, with other mechanics carrying the bulk of the weight. Like Fairy Tale, I wanted the bulk of the strategy and decision-making to happen during the draft, with the resolution phase just providing a quick injection of validation to the process.

I had an idea for a simple game almost immediately, so I created a prototype and played a few games. Cards were split into three types and assigned a base value, a bonus value and a conditional. You first drafted three packs worth of cards, then shuffled up your deck and played a modified version of War where you have three piles in front of you. (I wanted to try the extreme of not having any decisions in the resolution phase at all.) Based on the three cards you had visible, the conditionals of your cards might or might not trigger, which would determine if you got the bonus points. The highest total in each round would win a point card with a random value. One of the three types focused on low base values and high bonus values, another focused on the opposite, and a third was average in both.

Right around that same time I started hearing about a game called 7 Wonders that claimed to let seven people play a game in half an hour. Turns out 7 Wonders is a wonderfully crafted drafting game that advanced my understanding of this budding genre by leaps and bounds. Having played both it and my prototype around five times each now, here are my thoughts on the seven most important aspects to keep in mind when designing a game based around drafting:

1. Card types that do wildly different things so that the draft choices are not just straight comparisons.

My prototype: All cards had the same point numbers on them. This was simple but it made almost every decision into just a straight expected value calculation. The EV of each card was the base value plus the bonus value times the chance of succeeding on the conditional. The types could change that a little bit if you were setting up for a certain strategy, but not nearly enough. This was a disaster.

7 Wonders: There are many different card types (resources, science, military, civilian, commercial, guilds) and almost as many ways to score victory points. While this does add complexity, it's key to making the draft choices interesting. For example, think about a fantasy football draft where you're only drafting quarterbacks. It breaks down into a straight EV calculation to try and decide which QB will give you the highest number of average points each week. When you add in all the other positions, though, you have to think about what positions are being drafted aggressively, which ones you can leave for later, which positions generally score more points, and which ones have large point differentials between the strong players and the weaker ones.

2. Rare cards that can excite drafters and give them a direction.

My prototype: In the first version I tried balancing almost all of the cards against each other, but players had no obvious direction and weren't excited about any of the cards that they got. Perhaps unsurprisingly, the fact that Magic has cards that are somewhat obviously better than others and also rare cards that are labeled as such are both key to making draft more dynamic and interesting. There's no better feeling than opening that "bomb rare" out of a Magic pack. That said, it is a little more dangerous to have overly strong cards in a drafting game where the card will automatically come up. There is a lot of variance in what cards appear in a game of Magic, and also a decent number of cards that can answer the bombs. Still, in the second version of my prototype I added a second deck of powerful and narrow rare cards and created each pack with five commons and two rares. This was a huge improvement.

7 Wonders: This is one area where I feel 7 Wonders falls a little short, although the double resource cards fill this role to an extent, and the starting resource and wonder requirements help as well.

3. Not having to memorize or continually look at a bunch of hidden cards.

My prototype: It was easy for me to fall into the trap of having a hidden pile of cards that you're creating as you draft the cards, since that's how Magic works. Worse, a lot of the cards you could potentially draft depended heavily on you knowing the contents of that pile to properly evalutate them. So players had to keep picking up the pile and sifting through it to remember what they had drafted.

7 Wonders: Uses an obvious but elegant solution of just playing the card immediately and resolving the effects. One of the great things about this is that you get immediate payoff from each card you draft rather than having to wait until the end of the pack or the end of the draft. You do lose some strategy in trying to figure out what your opponents are drafting, but that's a pretty subtle skill that isn't necessary for a compelling experience.

4. Not having to draft and/or play every card.

My prototype: The final card of each pack is discarded. I also tried having players draft all of the cards and then letting them cut some from their deck, but that was too much work for not enough payoff, especially since it just came down to EV calculations again.

7 Wonders: Turns out that the designer of 7 Wonders came to the same conclusion, as the final card of each pack is also discarded. On top of that each drafted card can be sold for coins or placed facedown as a section of the wonder itself, which allows savvy players to "hate draft" (taking a card that would be good for a neighbor) and also mitigates the bad feelings of getting a pack that doesn't have anything that they really want.

5. Cards that have different amounts of value to different people.

My prototype: I was actually reasonably successful here with the conditions. Someone who was trying to draft an "A deck" would value a C type card wildly different than someone else. There was some amount of skill in trying to find the underdrafted archetype.

7 Wonders: There are tech paths that are represented through certain buildings in the game allowing you to build other buildings for free. They aren't necessary to succeed (you can always pay for the building the hard way) but they provide a clear reason why one player really wants a certain card whereas another might not. Additionally, science gives victory points in a way where the value of the science buildings fluctuates wildly from one person to the next. And finally, military varies from being incredible (when you're slightly behind your neighbors) to useless (when you're either far ahead or far behind both neighbors).

6. The opportunity to change your strategy based on what other people are doing.

My prototype: Sadly, not much of this exists. At some point I added conditionals that depend on the active cards of your neighbors, which helps, but it's hard to tell what they're doing anyway. It's all far too subtle.

7 Wonders: There is a mechanic where players can buy resources from their neighbors, so right off the bat you're intrigued by what sorts of resources your neighbors have chosen to develop. Then there are markets, which allow you to buy more cheaply in either or both directions. And of course, science rewards you heavily for jumping in when it's underdrafted, and military rewards you for staying slightly ahead of your neighbors.

7. Players should start the game from different positions.

My prototype: This one didn't make it into my prototype design, but I think it's a strong tool for accomplishing multiple other goals as well as adding replayability.

7 Wonders: Each player is attempting to build a different wonder, which determines both a starting resource and also a set of costs and rewards for building stages of the wonder. As mentioned above, this helps give a starting direction and also helps with players having different value for cards. Completing your wonder would have been a more flavorful victory condition, but relying on victory points instead does allow the game to keep wonder building as an optional task, which means that wonder balance doesn't have to be perfect. If you get one you don't like you can just focus on other goals for that game.

So there you have it. If you can't tell, I found my prototype to be fairly terrible (although illuminating) and 7 Wonders to be relatively exquisite. I highly recommend it. One of my only complaints about the game is that you can't "table" cards in larger games. (Meaning to see a card early in a pack and then still have a chance of getting it later when the pack returns to you.) However, it's not clear if the benefits of tabling outweigh the disadvantage of players having to deal with much larger pack sizes. Magic gets away with fifteen card packs because of rarity and colors; rarity shifts the focus heavily to the rare when you first open a pack, and colors narrow the relevant cards pretty quickly once the draft gets going. 7 Wonders has neither, so a smaller pack size makes sense.

As always, I've rambled on a bit much at this point, so let me know in the comments if you disagree with any of these "pillars" or if you have your own ideas about the genre.


diamonds in the rough

Welcome! Eventually this will be a repository for my thoughts on game design, but I wanted to start things off by talking about the work of others. Here's a list of ten games that you might not have heard of that I think are more than worth the time and/or negligible amount of money it takes to check them out:

Pax Britannica and The Night Balloonists

These two are only great if you have some friends in the room to play them with, but they are indeed great. Originally created for the Gamma IV competition, both are extremely clever takes on working within a significant constraint; in this case, that the player must control the game with only a single button.

Pax Britannica is an RTS (yeah, you read that right) that allows four players to battle it out using giant spaceships that are capable of producing other, smaller, ships. The player holds down the button, causing a ring to slowly fill, and then lets go when the filled area is in the quadrant matching what they want to produce. The fourth quadrant improves the player's economy, giving the rush/tech/econ relationship that RTSes are famous for. I wish that they had found a way to give you at least a modicum of control over who your forces attack, but what is there is more than enough for interesting gameplay.

The Night Balloonists is a deliciously atmospheric competitive game where four players are trying to collect the most pollen over a predetermined amount of time. The interaction comes in the form of ink that can be collected from a sea at the bottom, and can be dropped on opponents who are beneath. Dropping ink causes a balloon to rise rapidly, and being hit by ink causes your balloon to lose half of its pollen. Movement occurs based on fixed wind currents that carry the balloons around the screen, but skilled players can use the upward boost provided by ink dropping to maneuver fluidly around the screen. There's a surprising amount of drama and strategy given the simplicity of the one button control scheme.

These games are understandably bite-sized, but it's an inspiring bite that will stay with you for awhile. An amuse-bouche of gaming, if you will. (You can tell I've been watching too much Top Chef.)


I can't say enough good things about this game. Created by Derek Yu (of TIGSource and Aquaria fame), Spelunky takes the best parts of a roguelike and combines them with the best parts of a platformer, while eliminating the more tedious aspects of both. Essentially, it's a game where the player explores a randomly generated cave (hence the name) filled with traps and monsters, but in real time, with a very simple inventory system. It borrows the concept of perma-death from roguelikes, so the danger in the game is very real, which makes progressing that much sweeter. Unlike most roguelikes, the character doesn't level up, but that just removes the need for grinding and leaves that much more room for player skill to matter. Interesting enemies, clever traps, and the realtime nature ensure that every death is entertaining, which gives the game plenty of replayability as the player learns how to deal with the variety of challenges spawned by the randomly generated levels.

Desktop Dungeons

I'm a big fan of roguelikes in theory, but I don't have the time nor the inclination to spend 50 or 100 hours learning a new one, or even relearning an old one like Nethack. Desktop Dungeons takes the experience of a roguelike and distills it down into a tight ten minute experience, complete with plenty of different races, classes, items, enemies, and challenges that keep you coming back for more. The game also ramps up in difficulty and complexity over time, unlocking new monsters and items each time you win with new race/class combinations, and the experience culminates in special challenge dungeons that change the rules in interesting ways. Finally, the game has tileset support, with the official tileset at this point coming from the talented hands of Spelunky's Derek Yu. This one is highly recommended if you've ever been at all curious about the appeal of the roguelike genre.

Small Worlds

A study in the joys of exploration and scale as well as the potential beauty in simple pixel art. Your character, which consists of just two red blocks and a pink block as the head, moves through worlds created by those same blocks. As you move, the surrounding darkness peels back and the camera zooms out so that the entirety of the world you're exploring can always be seen on the screen. There is no combat, achievements, or even direct feedback really, but the game doesn't need it; quite simply, it stands on its own.

Dungeons of Fayte

I can't say that this game is as polished as the others on the list, but it was created as part of the Assemblee Competition over at TIGSource, where a bunch of people created art/sound/music assets during the first phase, and then people took those assets and assembled games out of them in the second phase. This is another one that is best enjoyed with four players, and it essentially harkens back to old school tile based RPGs, with dungeon crawling, training, shopping, and frequent little dilemmas where one player or the entire group is faced with a choice that has an immediate impact. The replayability isn't high, as the game suffers once the choices start to repeat, but it's a fun diversion that is worth at least a couple playthroughs.

Miner Dig Deep

This was a pretty early discovery on XBox Live Indie Games (XBLIG). It's a charming little game about digging deep into the earth in an effort to uncover precious minerals. There is a shop where the player can purchase an upgraded pickaxe, increased bag space, a brighter lantern, and all sorts of other useful items like ropes and elevators. There are no monsters, but it is possible to die from falling or from cave-ins caused by digging under unstable rocks. At its heart, the game is a big efficiency optimization, but it's a surprisingly fun one. As you play you start to learn ways to build your tunnels in ways that minimize both danger and travel time. The deeper you go, the more valuable the minerals become, but you'll need better tools to mine them, which cost more money. A tried and true gameplay loop that functions perfectly here.

Ancient Trader

Recently I returned to XBLIG for the first time in months, in search of something new, and found this little gem. It has the feel of a board game, with gorgeous hand drawn art and simple mechanics. The goal is to collect a set of ancient artifacts that unveil the location of a final sea monster that you must defeat to win. To become powerful enough to win that battle, you take advantage of arbitrage opportunities at various ports and then upgrade your ship with the profits. The game has that unmistakable polish that only comes when someone really loved the core idea, had a clear vision, and worked their proverbial ass off to bring it to life.

Monaco (unreleased)

I've been following Monaco since the initial announcement, and was thrilled to see it win the grand prize at the IGF this year. I'm not sure I can explain exactly what gripped me about this game early on - part of it is that I'm always on the lookout for good co-op games, and I can't wait to pull off some epic heists with my friends. The other half is that I'm a big fan of both the visual style and the fast paced nature of the game. Seeing Andy Schatz's humility and wonderment at winning the award on the big stage only cemented my love for this upcoming title.

Spy Party (unreleased)

An asymmetrical game where one player is the spy, attempting to complete missions while blending in at a party, and the other player is the sniper, watching and waiting for the spy to reveal himself so that he can kill him with a single bullet. I've had the pleasure of playtesting Spy Party quite a bit at this point, and it's safe to say there's nearly limitless potential here. It feels like it's exploring areas of game design that are almost untouched. Can't wait to see where this one ends up!