The Good Score

My development focus the last few weeks has been examining and minutely tweaking the score tuning for Five Ravens. When designing, you should always make changes with purpose, and my purpose here is threefold:

  • Reduce the number of ties. I had a week where we were tying 1 in 3 games, which tells me something is wrong. 
  • Improve the game's scaling between player counts. I was previously altering the rules for changes between 2, 3, and 4 player games. I removed those exceptions, and took the simpler, but more nuanced approach of altering the distribution and number of cards in the Relic deck. 
  • Introduce a broader range of final scores. When games are consistently coming down to 1 point, it feels a bit tight, and it feels a bit cheap. You don't want pepole to write-off a game because "eh you just won because of that one card." Well, sure, that one card and the other four I took! But, perception is always key to a game's success, and you must take isuses of perception seriously.

I really enjoy making light, tight card games like Five Ravens. It's the result of years of similar work, and as that is the case, I realize that I've internalized many of the lessons in regards to score tuning. Much of it is muscle memory now from years of doing this. But, it occurred to me that there may be value in sharing these lessons with others, so a blog post formed. Hopefully this case study and these lessons help you in your own projects!

If scores are too close, widen the range. In Five Ravens, players are competing every round to claim one of the Relic cards in play, with one being shown per player. Everyone will get one! This means a player's score is primarily driven by the five cards they claim, with a few other exceptions that are not as relevant to this. Previously, the dominant scoring cards in the deck had a range of 2-4 points, with a heavy focus on the 2 and 3s. Firstly, I added a second 4. This encouraged players to use their high value cards more, but also did more to break up the average of 2.5 point block. Secondly, I removed a 3 and added a 1. This further broke up that block, but also put another card that isn't that good, again, encouraging competition and tension for the more valuable cards. Finally, I converted one 4 to a 5. This meant that in a round in which one player gets a 5 and another gets a 1 really amped up the risks! This increases the average value, allowing for more score discrepancy, increases the value of using more potent cards, and does so in a way without addin gmore rules.

If someone thinks they can't win, they'll stop trying. One of the worst sins a game can commit is giving you a really good argument for checking out. If you're playing a two hour game, and in minute 30 you know you can't win? It's just less fun. At least, for some people. The highly successful designers at Splotter Spiel roll their eyes at this. One of their publicly stated goals is that a player should be able to lose a game on turn 1. I love that they have that premise! It's not one I share. You can also see the recent Shut Up & Sit Down review of Tigris and Euphrates from Reiner Knizia. Their praise for the game is glowing, but they dedicate some time to providing one caveat. The game is hard, and strategic, and you may get completely and royally hosed.

I'll clearly state one of my personal design principles: players should feel invested until the end, assuming more good play than errors. If someone is making horrible choices, repeatedly, I do not think the game should help them. To me, this means opportunities to excel and come back. In Five Ravens, the game has five rounds. If your opponent grabs the 5, then the 4, and you took a -1 and a know you're in trouble. This is why I tried to alter the distribution through scaling. The 5 point card only exists in 4 player games. The Illustrious Gems, which rapidly increase in points based on how many you claim, are limited based on player count. I do, however, leave most of the negative cards in every game. Why? The game has a negative feedback loop, similar to Dominion's. If I spend a high value card to get a really valuable Relic, I lose that card from my deck. This means I'll have to use craft, guile, or luck to make up for its absence. This also means if I spend my 8 to gain a 3 point Relic, and the next round I cap out at a 4, I need to mitigate those negative points. 

Another way I handle this overall issue is by obfuscating information. In Farmageddon, crops you harvest are placed face down. Yes, you can memorize them, but eventually it'll get muddied and you'll only have a vague idea of how well somebody is performing. In Five Ravens, the cards enter your deck, and only a few cards require I reveal which cards I claim to opponents. Therefore, again, it's largely secret. 

A final method to use, which is particularly useful in race condition games (i.e. the first to earn X points wins), is to allow for wildly disparate scoring opportunities. In Blood Rage, the scoring escalates as the engine grows and new cards come into play. In Hocus, the scoring is entirely player driven, leading some players to come roaring from behind with a big 20 point pot. In Modern Art, the top 3 cards are typically worth far more in subsequent rounds due to the historical value add. A game more recently exposed to me, Star Trek: Ascendancy, effectively forces players to keep each other in check. The problem is, the Federation and Klingons and Romulans are resilient, but also finite in their reach. The game creates a sense of whack-a-mole, so while you can counter the Federation now, you'll need to stop the Klingons next turn.  This makes it so players may have an off, or bad turn, but then next turn they're left alone to progress again.

Try to use such tools, if you share this goal, to keep players invested in the entire experience!

Note that which people seek. The quality of your tuning choices will be visibly reflected in what players desire. While you may understand the mathematical or probabilistic value of a scoring option, its value or potency may be too opaque to your players. If nobody chases it, it's not good enough. It doesn't matter what the math says!

Conversely, if too many people desire a thing, and put too much weight against not getting a thing, you have a imbalance perception. It is going to drive you batty as a designer if you hear repeatedly "Oh, well whomever scores card X wins. Game over man." Calm down, Private First Class Hudson (RIP Bill Paxton). 

Sid Meier has a popular approach to this, which is applicable to both issues. "Double it, or cut it in half." While this brute force hammer approach may not sustain through to the final product, it'll help you zero in on an issue. If you find nobody will go after a thing when it's 3 points? Make it 5 points. If you find a reward is far too easy at 8 points? Make it 4, or make it harder to get the 8 (though then people may see the dependency chain and back off). 

When creating your rewards in a game, it's good to begin with a set of goals. What do you want people competing over? Why? What do you want people to be scared of? What do you want to excite people? Use this to guide your tuning as you begin to observe player behavior. Having goals early prevents chasing rabbit holes, or muddying your vision as an assortment of data comes through.

Avoid the tie. I wrote a post about tiebreakers over five years ago called Breaking the Tie. It's pretty short, and I still think pretty relevant. You'll be amused by references to "Empire," which is one of the early names for the game that ultimately released as Cry Havoc. I worked on that thing for forever. I think ties are ultimately dissatisfying. I think tiebreakers feel like a cheap solution, and should be tightly baked into the overall structure so that when you reveal the tiebreaker, players go "oh yeah that makes sense." In Five Ravens, the tiebreaker is granted to the player whose captured pool is the lowest. Abstracted, this means: the player who spent the least but gained the most doing so.

There are a variety of ways to discourage ties through tuning. One is by expanding the range, shown above. Another is by increasing the distance between tiers of numbers, which is similar to broadening the range, but it makes it harder to arrive at the same number. In Farmageddon, for example, the crops that comprise 2/3 of the Crop deck have a range of 4/7/10/15. A third method is to fundamentally introduce variance in how players arrive at their final scores. In Five Ravens, if your 9 is captured, you lose 1 point. If your 4 is captured, you gain 1 point per Refined card. If you have specific cards refined, they give you a big jump, and only one person can do this. These unique patterns lead to more unique scores. 

A final method is to introduce completely alternate paths of scoring. This is difficult in a narrow game such as Five Ravens, where you're trying to score a single card each round. But, in broader euros, there are typically different paths to earn points. Concordia has approximately seven different scoring conditions, which force different decisions and player behavior in game. Caverna rewards you for livestock, mines, and a wide variety of options based on the buildings you purchase. Cry Havoc bases it primarily on Crystals, but also conquering territory, eliminating the enemy, and other conditions based on the asymmetric races. If you create multiple paths, it's less likely your gamers arrive at the same destination.

Think about it fundamentally as such: how can you, the designer, increase the probability that players arrive at unique scores?

Keep the math simple. As a child of both the video game and board game worlds, I'm often asked how video games are different. Beyond the obvious notions of graphics and technology, a key one to me is calculation. Video games calculate yoru complete armor set's score, which allows Blizzard and other developers (I'd argue often excessively) to bake in dozens of numbers to the equation. Meanwhile, board games like Mice and Mystics often have a simple defense stat. I have 4 defense, I roll 4 dice. 

In a board game, you must keep your math simple. It needs to be no more complex than napkin math, somebody can quickly do in their head. If Five Ravens let you score 500 points, players would spend half the game doing math in their head and in the end they wouldn't quite realize why they were a.) exhausted and b.) didn't have fun. 

Keep your math simple. Figure out methods of reducing repeated calculation, such as once you score a card, you increment a tracker, or gain tokens, and then put the card away. Figure out ways to eliminate division. Rules that say "reduce by half, rounding down" means players have to go 7, 3.5, 3, which is 3 calculations. That adds up, especially when done repeatedly. Instead of including the rounded down clause, perhaps tune the base value to even numbers? Or, use board information and UX to simply communicate the numbers. If you arrive at the sums by halving a number and rounding down, just print 7/3/1 on the board for them to reference. 

One of the best lessons for me on this topic in recent memory is the point scoring in solstice. Previously players raced to 25 points as cards were worth 1-5 points. But, that range was too great, the number involved too much constant counting, and players struggled much more to evaluate a 5 versus a 3 versus a 2. I took a step back and recognized that the game was about small, medium, and big scores. I went to a 1/2/3 model. This let me shrink the final scoring number, reduce the amount of tiny math, and made it very intuitive to evaluate a 1 versus a 2 versus a 3. People can easily grasp "small, medium, and large" versus "tiny, a little bigger, decent, quite good, really great."

Keep your math simple.

In conclusion. I hope this was valuable, or at least leads to you scratching your head, challenging my thesis, or using it to redirect your thinking. If you have any thoughts, counters, or additions, note them below in the comments.