Breaking the Tie

A few friends on Twitter recently began a short debate regarding tying in a game. More specifically, what should the designer do (if anything) to reduce the chances of a tie. Furthermore, they pondered whether ties are bad in the first place, and what a reasonable tie breaker is.

I was prompted to throw my thoughts into this rather gentle ring, so I am doing so via the blog.

This is a bit of a difficult argument to enter as I don't have a great deal of experience with tie breakers. I rarely encounter them in the games I play and the games I've designed haven't needed to rely on them too heavily. I've heard of player's using the tie breaker for Farmageddon, but I've never personally used it. In various stages I think we've had ties in Empire, but they too are somewhat rare.

I have used a tie breaker a few times in Alien Frontiers, a game that scores very tightly. In these instances, the player who usually lost the tie felt a bit cheated. The tie breaker, if I recall correctly, is having the most Alien Tech cards. Catching up here while also landing colonies and stalling an opponent isn't really possible.

I'll argue, therefore, that games are better when they do not result in a tie. A clear, decisive win by a single player or entity is the best outcome. How does one go about designing a game that tends to not result in a tie?

There are a few possibilities. Most obviously is to create a game that features elimination. One cannot tie if he no longer exists! Another is to create a finish line or objective, i.e. the first player to accomplish a goal wins.

Another possibility is to fill your game with multiple ways in which to earn points of varying quantities. For example, every round every player can earn points based on a certain metric. The point values assigned can differ based on superior performance.

You can also create multiple paths go gain points. For example, in Empire, players earn a small amount of points for controlling territories throughout the game. At the end of the game, they earn a more significant number of points for winning battles. The most points can be earned by earning the Strategic Victories, which are earned for going deep on specific aspects of the strategy. Typically, players respond and react to one another and in a four player game, you see 1-2 players focus on battles, a player focus on territory, and a fourth player throwing a wrench in things.

One option is to limit the possibility for winning players. In general people don't care for a runaway leader, especially when it's painstakingly obvious. But, if you make the distribution of points such that it is very unlikely that more than one player can win, that helps. Just be sure to make this information hidden so everyone doesn't check out.

As a final suggestion, you can create varied opportunities for players that are outside their control. In Farmageddon, the crop cards you draw are random. There are 4 basic crop types and 10 unique FrankenCrops. The first choice you must make every turn is what to plant. Players can't control their point opportunities, but they can choose which opportunity to pursue based on its likelihood of success. Because these point values vary significantly (4/7/10/15) it leads to a very tight, but rarely tied experience.

But, and I think this isn't terribly negotiable, unless it is impossible to tie your game, you need to create a tie breaker. Creating a tie breaker can be quite simple or complex depending on the number of "knobs" your game has to twist. If your game is incredibly streamlined and simple, it may be difficult to craft a compelling tie breaker mechanic.

I think it is of the utmost importance, both for simplicity and to reduce player frustration, to make the tie breaker both obvious and deeply integrated into the core experience. When the tie breaker must be referenced, none of the tied player should say "what?" when it's revealed. Nobody should feel frustrated or cheated for not memorizing the rule book and knowing the tie breaker is a path that isn't well-tread or obvious.

In fact, the tie breaker shouldn't be something anyone has to think about. It should be as pre-baked as possible, like a delicious cupcake.

In Farmageddon, the tie goes to the player who has harvested the most crops. You must harvest crops to win in the first place, so if you've harvested more, that seems like a good way to denote the victor. In Empire, the first tie goes to the player with the most Units on the board. This is something players want to do regardless and fictionally, the player with the biggest army present should get the nod. The second tie goes to the player who has won the most battles. Again, fictionally, the player who has the best army should be favored.

A deeply integrated, intuitive, obvious tie breaker is essential. Don't get clever in this instance. If people spend an hour on your game, don't belittle their time, or leave a sour final impression, with a clunky and dissatisfying tie breaker. Be sure that the winner still feels like he or she earned it.

What are some of your favorite tie breakers? What are some of your least favorites? Which games tend to result in ties? And why?


Some excellent discussion points, folks, and thanks for getting this written, Grant.

I view a "tie" as a "win-win" victory condition. Which is to say, not a victory.

A win-win negotiation allows two or more parties to depart with a mutually agreed upon resolution that they both benefit from. But both parties also have to give up something as a result, which means a loss.

There's the rub of it. A "loss".

A tie does not proclaim a winner and does not give a satisfactory end condition if we view it from a game's definition of the final goal. The assumption here is that every game designed always intends to have one player or team succeed at a higher level than the others. I doubt you will find any game that says in the rules that the "goal" is to do well enough to be of equal measure and standing when compared to opponents.

I continue to believe that a tie is not necessary and is altogether a weak attempt to placate or otherwise avoid the issue of upsetting participants by supplanting the goal of the game, which is to succeed above all others.

The spirit of the game remains untarnished regardless if a tie conditions exists or not. We play games to have fun, and if you aren't having fun playing the game, you should play something else.

As a society and as individuals, we should always attempt to succeed and become more or do more than what is "average". By attempting to do so, it is perfectly OK and acceptable to lose. Through defeat, we have an opportunity to learn and grow, both cognitively and emotionally. Designers are already attempting to challenge the players and entertain them. Don't cheat the players out of an opportunity to learn something along the way.

And if you can't stand to lose, it's not the game that has an issue, it's you.

"It's important that you score the most points. But you'll also REALLY want to be watching this other element that is NOT points. Because if you've got the same amount of points, you'll also need this other thing.

So, y'know, do that.

And if you have the same amount of those, then here's three other things."

Yeah, I've got a REAL issue with tiebreakers and I have no problems admitting it.

If the central premise of the game is a contest (get the most/least), ties are more than OK with me. You are delivering the right level of reward -- you didn't win outright, but you certainly didn't lose. You were among the winners. Good on you. If it's an issue, go into "overtime" like in Nefarious.

If the central premise is a race (first one to X), ties don't happen as often. You just need to ensure that there's fairness in the race setup and execution: youngest player goes first, balancing mechanisms to prevent outright runaway leaders, and so on.

The tie problem isn't the game itself. The problem is the culture gaming has become. People, especially in America, play to win. And when they don't "WIN win", it feels like failure. Accepting ties as "not loss, but not win" is a first step.

Interestingly enough, and I should have used it in the article, one of my favorite games, 1812, has a tie. And it has happened a few times (I'd say more common than not). It makes sense both mechanically and historically and I've never felt saddened by such an outcome for that game.

Honestly having no tiebreakers may be the best approach. Fewer rules, fewer things to learn. Try to make it such that ties are unlikely, but if they do occur? So be it.

Why did you go 5 deep? Is that legitimately necessary or did you just think of 5 tie breakers? My gut reaction is that it may be a bit much!

So I'm clear, I'm not criticizing, I want to understand your logic.

So, I asked for this article because I am fairly opinionated about it. When I designed Scallywags, I put in a great deal of effort to reduce the number of ties possible... the game changed a bit in the publishers hands, and now, ties are not uncommon. It bugs me in that specific instance because I was so conscious of it in my initial design.

I also feel that a game should present enough opportunity for one player to be dominant. Whether that dominance is by a single point or fifty is irrelevant... only that one person should have the opportunity, through good play, to make a slightly better use of their resources than the other players. If one player cannot, within the limits and time-frame of the game, legitimately beat the other players, it may be time to examine the game: is there too much parity in decisions, is there too much randomness, are the point spreads too even.... that sort of stuff. If ties still exist, there needs to be a tie breaker. Adding in more randomness or weird scoring mechanisms to try and avoid this situation is a disservice to the better players of the game.

Yes! Just as I was trying to say on Twitter. A tie-breaker should be relevant to the game. It should be a fair judge of who, among those that performed the same on the primary criteria, played the better game. If you can't find such a measure, leaving the game at a tie is fine.

That assumes, however, that ties don't happen too often. As a player, I don't mind tying for the win if we played equally well. Some games have strong catch-up mechanisms, though, and that could make a tie very frustrating.

Reading this article, I can't help but think of Uwe Rosenberg, who has no problem with ties, but hates tiebreakers and refuses to put them in his game.

I recently struggled with a good tiebreaker for Sword Merchants. There are a couple of directions players can go, and I didn't want the tiebreaker to favor one over the other. Right now I'm testing a tiebreaker that should be balanced between the two, but it took a few weeks to settle on my current course.

It fascinates me how players evaluate their performance in a game. In the worst case, you only care whether you won or lost and nothing in between matters. In the best case, you rank yourself among the other players. I've never seen a published game where the scores of your competition are irrelevant and all that matters is your own performance on an absolute scale.

I've tried making a game like that, but players just aren't interested in their objective performance—all that matters is how they did relative to everyone else. I'm sure this is the same psychology at play when you try to pass someone driving slower than you on the highway and they speed up so you can't. I just wish I understand the root of it better.

I'm a big believer in tie-breakers. I can't abide a tie. When two players have the same score, there need to be other ways to assess which played the better game. I have literally five layers of tie-breakers in my current work-in-progress (East India Company) - not just for determining the winner, but even for determining which ship gets raided by pirates when they strike.

Unfortunately, that means I'm guilty of violating one of your rules:
'Make the tie breaker both obvious and deeply integrated into the core experience. When the tie breaker must be referenced, none of the tied player should say “what?” when it’s revealed.'
Nobody can memorize five layers of tie-breaker criteria.

The few times I've had to reference the tie-breaker during playtest, the players didn't know what to expect before I read it off, but when it was determined, I don't think anybody felt cheated, as though the tie-breaker was too arbitrary. Rather, people seemed to agree that the tie-breaker criteria were consistent with the theme and feel of the game. So I still feel okay with the way at is at this point.