clock menu more-arrow no yes mobile

Filed under:

Fixing chess: One game developer's redesign of a 1,500-year-old game

The weekend's chess matches with my father ended as they always do, but with a twist.

I lost, mostly, but it wasn't to a checkmate. Instead it happened four times when he managed to get both of his warrior kings across the board's midline. That last midline crossing happened despite my defensive line of elephants, wild horse and pawns.

This is Chess 2, a chess variant created by MIT grad and famed game designer David Sirlin and given computer game life by developer Zac Burns, to correct what Sirlin views as major design flaws in the original board game.

There are thousands, maybe tens of thousands, of chess variants like Chess 2 in the world. Creating new rules for one of history's oldest games is a popular exploration for the cerebral sport.

Sirlin's own take on chess was inspired by a particularly lopsided match of tournament Street Fighter.

In the match, a player managed to dominate his opponent by pushing their character into a position that trapped him.


"I remember thinking that it was somewhat ‘chess-like' in that it was kind of a memorized pattern he was doing and it was sort of ‘turn-based' if you use your imagination," Sirlin said. "What I mean is that he intentionally played to create situations where he could react as much as possible. He wanted that match to be like ‘Honda does his move. I see it. Now I do my move.' He didn't want there to be any guessing."

In another match, Sirlin noticed a player take the opposite approach, creating a situation that forced his opponent to continue guessing what was going to happen next.

"I really liked that the game of Street Fighter allowed for the player to push the game toward being ‘more chess-like'(zero guessing) or more ‘yomi-like' (all guessing)," Sirlin said.

Yomi is a fighting game term that means reading your opponent to try and figure out what they will do next, a sort of educated prediction. It's a concept that Sirlin is so intrigued with that he created a card game around the premise.

"One day I was thinking how even though I like chess and yomi, it's too bad each of them is on such an extreme end of the spectrum," he said. "Chess does not have any double-blind simultaneous decisions; you can't use prediction really. Even if you know what the opponent will do, that actually doesn't help in a game with perfect information.

"On the other hand, the Yomi card game is so much about prediction that it's hard to do the equivalent of [that first match]. I thought about games in the middle of chess and Yomi, then I thought about injecting some prediction into chess. As soon as I thought about that, I realized that doing so would reduce memorization, which is a quite an issue in chess. Then I thought while I'm at it, of course I should address draws, and given my background and interests, of course it should be asymmetric. It sort of fell into place then."

Chess 2 was designed around what Sirlin saw as four major issues in the traditional game: too many ties, too much reliance on memorization, dull symmetrical gameplay and a lack of "yomi."

Chess 2 introduces three major changes to the traditional game. The most obvious change is the addition of five new "armies," or traditional pieces with new moves. These new moves include things like a queen that can only move as a rook or a knight, or an army that features no queen and two kings, one of which has a "whirlwind attack," or a rook that can teleport to any open space on the board, but cannot capture or be captured.

"The many different dynamics that come from 21 different matchups give you so much more to sink your teeth into than one," he said. "Each one offers a different play style, so different players will find their own personalities or skills matching up with one army more than others. You get specialists of certain armies. You only have to learn one army to play the game, but you get to take part in the rich and varied environment that six armies creates."

The game also adds a new way to win. Players can still beat an opponent by capturing their king, but this almost becomes secondary to the new method, which requires a player to move their king (or kings) across the board's midline.

The Armies

chess 2

Chess 2 is played with one of six armies. Here's a quick breakdown on how they differ from traditional chess.

  • Classic: Regular chess and the only army that can castle.
  • Nemesis: Queen can only capture or be captured by the king. Pawns can move one space in any direction toward a king.
  • Empowered: When a knight, bishop or rook is next to another such piece, they gain that piece's movement powers in addition to its own. Queen moves as a king.
  • Reaper: Queen can teleport anywhere to capture, except the backrow or king. Rooks can teleport, but can't capture or be captured.
  • Two Kings: Two kings, no queen. Second king has an optional extra move at the end of the turn. Both kings can destroy all pieces around it in a whirlwind attack.
  • Animals: Knights can capture own pieces. Bishop can only move two squares but jumps back to original square in an attack. Rook moves three squares and captures all pieces in the way. Queen moves as a rook or as a knight.

Sirlin says that in modern chess, grandmasters are drawing up to 62 percent of the time in matches.

"If a new competitive game came out today that took like an hour to play and had a 60 percent draw rate amongst experts, it would be unthinkable," he said. "This is a major problem.

"In chess, getting ahead means the opponent is losing pieces. This snowballs. Playing out some long sequence of moves that eventually guarantee your loss would be boring, so that's why in Chess 1 you can still play for a draw to keep it interesting. In Chess 2, most of those sequences are eliminated anyway by the 'midline invasion' rule."

Finally, Chess 2 introduces yomi to the game, requiring players to try and read their opponents and predict their move. This is done through something called a duel. When a player captures a piece, the defender has the right to call for a duel. Each player places zero, one or two stones in their hands. The player with the most stones wins — though tricking a player in a bluff by using none, can beat an opponent with one or two stones in their hands. If the defender wins then both pieces are lost, not just the capture piece.

"Some people (including myself when I was younger) say that games without any hidden information or randomness have more skill," Sirlin said. "The problem is that perfect information turn-based games are inevitably going to be more and more and more about memorization. That type of game is dangerously solvable, and that's what chess is inching toward right now.

"Chess 2's dueling system adds hidden information to break this up. This makes it less solvable, it allows you to have advantage for predicting what the opponent will do, and it raises the skill ceiling of the game quite a bit. At first I know it can seem jarring to be bidding stones over pieces dueling, but the skills it tests are more natural and integral to chess than you might realize."

As a lifetime player of chess, though not a very good one, I found that Sirlin was right. Chess 2 feels more like chess sometimes than the original. That is, the spirit of chess, of trying to outthink and outmaneuver your opponent, surfaces above the edge given to players who have memorized vast libraries of opening and closing moves.

Instead of trying to recall the best opening with two knights, or how to push a player into a closing game, I found myself thinking on my feet and being genuinely surprised with the outcomes.

While it was Sirlin who penned the rules for Chess 2, it was Burns who breathed video game life into it, bringing it first to the Android-powered Ouya console and now to computers.

Burns said he came across Sirlin's concept for Chess 2 while doing some design research and reached out to the designer to see if he would let him turn it into a video game. Within a month of getting the approval, he had a rough version of the game up and running.

To create Chess 2, Burns had to come up with new designs for all of the pieces and implement the rule set and a way to duel easily in the game. He even reinvented the way chess Elo rankings work. But the biggest challenge has been trying to create a single-player version.

"People forget that all of those chess AI were developed over 20 years by some of the best minds in chess," he said. "I had to make mine from scratch."

chess 2

And for it to be competitive, Burns would have to create 36 different AI, one for each possible matchup. Instead he created a light trainer, a place where players can get used to the new rules but not expect to be too heavily challenged.

While he hopes to continue to build out the game's brain, Chess 2's ranking system is a slick rework of the original game's Elo rating, built mostly around the way those numbers are surfaced to players.

In chess, players start with a 1,200 Elo ranking, which then goes up or down based on how they do in matches. In Chess 2 players start with a one and that number grows depending on how you do.

While it sounds vastly different, Burns said it's basically a surface change. The traditional Elo is still being used in the background, so what's really being displayed is the change in your ability. Over time, the more a person plays Chess 2, the closer they get to finally showing what is essentially a traditional Elo ranking.

"Eventually," Burns said, "the score converges with an Elo."

He implemented the change because he thinks starting with a 1,200 and often going down in your first games is a terrible user experience.

chess 2

Taken together, Chess 2's rework of the classic seems to move the game forward, "solving" a lot of issues for both newcomers and grandmasters, but not everyone sees it that way.

Tony Rich is the executive director of the Chess Club and Scholastic Center of Saint Louis, which has hosted the U.S. Chess Championships for the past six years. At the end of the month, the world's top players are converging on the center to face off in what Rich says will be the strongest tournament in the history of chess.

For him, Chess 2 is just one of the many chess variants that have been around almost as long as the original game.

"There are literally thousands of chess variants out there, but that's pretty much all they'll ever be — variants of chess," he said. "The classic game of chess, as it exists, has been around for 1,500-plus years for a reason: It is already one of the most perfect forms of you versus me."

As for the idea that Chess 2 solves some of the problems of the original, Rich believes those "problems" are in fact a reflection of the game's perfection.

"We should be careful not to confuse ‘memorization' with ‘technique,' which is the real reason why chess games end in draws," he said. "It's important to remember that there is no element of chance in chess, and there never has been for 1,500-plus years. It is completely a game of skill — or lack thereof. The game of chess as it exists is just a continuous series of puzzles — solvable puzzles and it would be literally impossible to ‘memorize' these infinite puzzles, all of these tricks chess has to offer. But having ‘technique' an understanding, a feel of the current position and the ability to execute toward the next puzzle is exactly what separates one chess player from the next. On the top levels of chess, we very often watch two perfect techniques clash and these games almost naturally end in draws."

What issues some see in chess, Rich says, are more spectator issues than player issues.

"There is nothing necessarily wrong with chess, it's just that everybody wants to see a decisive result," he said. "Nobody likes a draw! But draws are a simple fact of the game itself. Some tournament organizers have introduced some tweaks to the system to encourage more fighting chess. Soccer-style scoring, for instance, awards three points for a win and one for a draw versus the traditional 2/1 ratio, which gives players added incentive to go for a win."

The idea that any variant, including Chess 2, could ever hope to topple chess is improbable, he added.

"The biggest obstacle will always be trying to teach 600 million people worldwide a new set of rules," he said. "These days, the most popular variants have been skipping the rule changes and just playing the same game of chess, only faster — with Rapid and Blitz Championships featuring quicker time controls."

Both Burns and Sirlin know that Chess 2 is unlikely to become the new chess, but Burns thinks this particular variant has an edge over others.

"More than 10,000 variants of chess exist," Burns said. "It's very easy to come up with one."

"But I don't think a lot of those other variants are done in a very thoughtful way. Sirlin really approached this in a different way by thinking about how we can amplify position play, how we can amplifying finding inbalances. How can we diminish things we don't like, like memorization and opening book theory. I wouldn't have dedicated the last three years of my life to someone else's game design if I didn't think it was interesting. I was genuinely inspired by Chess 2."

Good Game is an internationally syndicated weekly news and opinion column about the big stories of the week in the gaming industry and its bigger impact on things to come. Brian Crecente is a founding News Editor of Polygon.

Sign up for the newsletter Sign up for Patch Notes

A weekly roundup of the best things from Polygon