Thursday, October 11, 2007

Tie Breakers

Last week, when playing Thurn & Taxis, we momentarily thought we had a tie. (Momentarily, I say, because I added up my 21 points of chips and got 19, but that's neither here nor there.) This inevitably led us back to the rulebook for the perennial question, "What breaks ties?"

In Thurn & Taxis the first answer was, "the player who earned the 'game end' bonus tile'", which makes a lot of sense, because that's a definitive goal that players should usually be going for. However, the second tie-breaker, didn't make sense, because it was, "if [the person with the tile] was not among those tied, the player closest clockwise from this player who was tied with the most is the winner!"

So Thurn & Taxis, to offer a reminder, works like this: when a player goes out, play continues until all players have had an equal number of turns, and thus ends to the right of the start player. This means that unless the last player is the one who went out, the winner is a player who was advantaged because he had more of an opportunity to react to the game ending, which seemed to me to be the opposite of what the tie-breaker should have been. I suggested that going counter-clockwise from the ending player would have worked better, because that would have been a player more likely to be disadvantaged, which led me to a general pondering about how tie breakers should be written.

A Philosophy of Tie Breaking

So what makes a good tie-breaker? I have three criteria: it should be obvious, fair, and ideally unique.

Having a tie-breaker that is obvious is the most important criteria. Inevitably, if a part of the rules doesn't get explained when you're learning a new game, it's how ties are broken. So, you want a tie-breaker that feels obvious: in other words, even if you don't know what the tie-breaker is, when you find it out you want to be able to say, "That makes sense", because the opposite case, where you suddenly find after the fact that you should have been hoarding sheep (or whatever) for the tie-breaker can put a damper on a game.

Almost as importantly, a tie-breaker should be fair. My complaint about the secondary Thurn & Taxis tie-breaker is that it didn't seem fair to me. It would have seemed fair if it in some way either rewarded a player who was truly disadvantaged or else rewarded a player who had extra resources (which especially in a resource-to-victory-point game engine are essentially fractional victory points).

Finally, if possible a tie-breaker should be unique, which is to say something that can't result in yet another tie. Having the end-game marker in Thurn & Taxis is a pretty good example of this sort of thing, because it will usually be held by one of the winners; the designer just didn't think beyond that for the rare cases in which it turns out to be held by a loser.

Looking at Some Examples

So how do different games deal with tie-breakers? I've decided to offer up a few examples, each of which I've looked at by my criteria.

Primordial Soup; Torres
: First, the Holy Grail of tie-breakers: games where you can't tie. This is a pretty rare game design element, but usually, I think, a good one. Torres and Primordial Soup are both good examples, because they're games where you literally can't have the same score as another player: instead, you skip over them.

This technique is often put to good use in any sort of game where you have some sort of absolute positional difference, offset in subsystems where ties are relevant. For example in Entdecker you place figures on jungle paths, and if there's a tie, the person who placed first wins; conversely in Patrician you place floors in towers, and if there's a tie, the person who placed last wins.

The Settlers of Catan: Settlers is a game which allows no ties, because you win by having the right number of points on your turn. This really shows the difference between games which go a set length of time, and thus allow ties, and games which just go until someone wins, and thus usually don't.

Ingenious; Tigris & Euphrates: These two Reiner Knizia games offer the next best thing to no ties: a tie-breaker that is so entirely obvious (and fair) that trying to reach it is just a standard part of your gameplay. In each game, you win based upon your worst score in multiple colors, and in case of tie you drop down to your second worst or third or fourth. Thus the entire schoring mechanism is an organic whole.

Havoc: The Hundred Years War: In this Poker-like game, the person who has won the more battles (hands) is the winner, and if there's still a tie, it goes to order of placement in the final battle, making it fair, relatively obvious, and with the second tie-breaker unique.

Ticket to Ride: This is a pretty standard game with good, but not great tie-breaker. The person with the most completed destination tickets wins ties. That strikes me as fair, but it's neither obvious or unique. I'd guessed that the tie-breaker would be the person who has the longest-route bonus, since that's usually unique, but I'm not unhappy with the actual rule.

Carcassonne; Caylus: These games have my least favorite tie-breaker. Either the game explicitly says there is no tie-breaker, or else just doesn't mention one. Besides being anticlimatic, it feels lazy on the part of the designer. I think some game designers feel like they can get away with it because you earn enough points that a tie is pretty unlikely ... but they will come up sometimes. For Carcassonne a potential tie-breaker is immediately obvious: a count of unused meeples. For Caylus a good tie-breaker is a bit more difficult because unused resources have already been valued with points. I'd be tempted to offer a tie-breaker based on total contributions to the castle, with earliest contribution being an additional tie-breaker, since building the castle is the theoretical purpose of the game. Alhambra and Coloretto were another few games that I found that had no tie breakers.

In looking through games, they generally did better than I expected ... other than those which didn't include a tie-breaker at all.


Smatt said...

I posted on BGG this past week about the Thurn & Taxis tiebreaker. We had had a game where the scores were 22-31-31. The 22 went out first. Then my friend won with the 'first to the left of the person who went out' rule. I was especially bummed for two reasons:

1) This was a game about establishing postal routes, and I had used 19 of my 20 houses; the other 31 player had only used 18 of 20.

2) The 22 player had started the game, so we all had equal turns.

Tiebreakers are rare, but when they happen, players really do want them to be intuitive and sensible. With the above, I felt gypped.

Anonymous said...

As always, nice analysis, Shannon!

I really don't mind a game that ends with players "rejoicing in a shared victory" as Jay Tummelson puts it. Ties aren't the end of the world. During single elimination torunaments or if there is an indvisible prize for first, they can cause problems, but otherwise ties are great! (In the conditions it matters, there should obviously be a tiebreaker specified by the organizer of the event.)
However, if there are tiebreakers, I fully agree that obvious and fair are good characteristics, but as long as there is an eventual unnique one, I like having layers upon layers. It's more fun to be able to say, "I lost (or won) on the third tiebreaker!" than to say, "I had more money, so I won."
Both of these are obviously personal tastes, but I'd like to hear reactions to them.

Shannon Appelcline said...

I find uniqueness important for two reasons.

First, that you have some ultimately unique tiebreaker, so that after the nth tiebreaker you're not forced to declare an anticlimatic tie.

Second, to keep the tiebreakers within the realms of obvious.

Conversely, I think I'd be OK to have some large number of tie breakers if it's entirely obvious what each of them is. Ingenious and T&E are fine examples of this.

Unknown said...

Thanks for this interesting read.
As a player, I am not a big fan of tie-breakers, especially in games where lots of points can be scored - finishing second on 100 points when someone has 101 is bad enough, but scoring 101 and losing on a tie-breaker feels even more frustrating.
Regarding Caylus, as a lazy designer, I thought that ties would be rare enough so that the problem would not arise very often. Also, which satisfying unique tie-breaker could have been used? Castle contributions would make sense and be nice thematically, but ties are likely to happen there as well (even if you decide to check each castle section individually) - and I prefer no tie-breaker at all than a whole arsenal of tie-breakers which prove inefficient in the end. The order of contributions could be an option, but besides forcing the players to remember it (not too much of an issue), it would give an advantage to some strategies (with early castle-building) compared to some others, for no very valid reasons (additionally, it wouldn't work for the very rare cases where several players tie with no builds at all in the castle - unlikely enough, I agree).
As for Caylus Magna Carta, which inherited the absence of tie-breaker from its big brother, you can rightfully argue that ties are many more common there, since the overall scores are lower. Castle building could have been an option here as well - with the same faults, maybe even stronger, than in Caylus (going by the order of contributions would be a huge advantage given to the first player, if stone is available among the neutral buildings). That's why the publisher and I decided to keep the absence of tie-breakers - even though some players like them.

MrHen. said...

I will chime along with those who said that some games without tie-breakers is fine. In games where there is a particular objective, a tie-breaker makes sense. In a game where the goal is to accumulate points, I think ties are perfectly acceptable.

Some games really have no good method for breaking ties. How would you break a tie in Blokus?

Shannon Appelcline said...


I can see your argument based on high-point games that came really close. Still, I personally prefer to see who was that fraction ahead, even if it's really lost in the "noise" of the score.


It's quite possible that some games just don't allow tie-breakers.

For Blokus, I'd at least consider a tie-breaker based on number of pieces left (vs. the score, which is number of squares). I'd suspect that a player who has less pieces left played more skillfully, though I'd need a more experienced player to verify that.

Anonymous said...

No good method for breaking ties in Blokus?!

Playing earlier in the turn order is a pretty strong advantage--there's more space available on each of your turns.

Obviously, last player in the turn order should win ties.