June 19, 2012

The Choices We Make


The title of this blog post is “The Choices We Make.” This statement is intended to explore two intertwined manifestations of the statement. First are the choices game designer’s make in how they sculpt a game to meet (and hopefully exceed) their intended goals for their design. Second are the choices players make during the game and how those choices are shaped by the game’s design.

More specifically, I want to talk about different scoring mechanisms/systems employed in games and how the quest for the right scoring mechanism in my own game, Hegemonic (not again!), has slowly progressed. The impetus for writing this was sparked by a reading of this article on the blog “Illuminating Games.” I will cover some of the same territory, but also go into more detail.

Finding Mr. Right

One of the biggest struggles I’ve had in designing Hegemonic has been developing the “right” scoring system for the game. Now, I’m sure all kinds of designers, great and small, have said something to the following effect: “The scoring system IS your game.” Yeah, I know. Thanks. Well, the funny/eerie thing is this; the scoring system in Hegemonic was never at the front of my initial design sketches, for better or worse.

As a result, the scoring system has gone through dozens (literally) of variations over the game’s design and development cycles. It’s the one element that has gone through the most radical changes from version to version. This isn’t to say the tested (and discarded) ideas didn’t work. Not that at all, most of them worked quite well. Rather it is about finding the “right” scoring system that best supports the goals of the game. In the case of Hegemonic, I wanted the scoring system to support the following goals:

#1) Encourage a high level of direct spatial interaction on the board
#2) Create tough decisions + reward strategic play
#3) Maintain tension and create a compelling “arc”

For competitive games with a strong spatial element and low-randomness environments, the list of above goals might be pretty universal. I would venture to guess that most designers want to create interesting games with player interaction, strategic opportunities, tension, and an enjoyable arc! For reference, this series of articles posted on the game’s journal does a wonderful job expanding on these topics. Regardless of the intended goals, I feel any game’s scoring system is central to achieving its goals, as it drives the play and the total experience.

Scoring Systems

The scoring systems themselves are quite diverse across the plethora of games that exist. I will attempt to break down the countless forms and variations of scoring systems into three tangible dimensions that we can talk about: (1) Scoring Type; (2) Scoring Frequency; and (3) Scoring Information.

Scoring Type – The “What”

The first division in scoring systems (and the highest level mechanical division in my mind) is whether scores are in fact numerical “scores” and/or whether victory is based on meeting a fixed/discrete win condition. The scoring type defines “what” the game is all about, as it drives what players are fighting or competing over during the course of the game.
In general, modern boardgames appear to be favoring actual numerical “scores” that players compare at the end of the game to see who won (i.e. VP’s). CARCASSONNE is a clear example, as are any game with victory points, score tracks, tally charts, or any sort of continuous value range of possible scores. Even an abstract like GO is also going to going to fit here, as the game is scored in the end and points are compared. In most cases, game’s with numerical scoring rely on a different mechanism to determine when the game is over. In Carcassonne, the game ends when the draw stack of tiles is empty. Go ends when both players pass. In these examples, the scoring is decoupled from the end-game trigger.

Games with fixed/discrete win conditions are usually focused around accomplishing a specific “objective”. These type of games appear to be less common today (totally unsubstantiated observation) but there are plenty of examples of traditional and abstract games that fit within this division (and of course modern games too). CHESS is a typical example, where the winner is the player to capture the opposing player’s king (by any means necessary). Games of attrition (CHECKERS) where the winner is the last player standing generally fit within this bucket. Other games include a fixed board state that is a win condition, such as alignment games like CHINESE CHECKERS or TWIXT that give victory to a player able to achieve a specific formation with their pieces. In general, these “Discrete” win condition games are linked with the end-game trigger. When a player captures an opponent’s King, they immediately win and the game is over.

Of course, you can have a hybrid approach where the gameplay hinges around numerical scores but the win condition (and game trigger) is fixed at a certain “threshold” level. SETTLERS OF CATAN is such an example, where players accumulate VP’s over the course of the game, but the winner is the first player to achieve a fixed victory point objective (10 points). EUCHRE, ROOK, CRIBBAGE, and plenty of other games use such a “threshold” system to determine the winner, where the first player (or partnership) able to reach a pre-set score limit is the winner. DRAKON’s winning condition is both numerical and discrete. You have to get 5 coins AND get out of the dungeon first, and good players plan their escape route as carefully as they plan getting the last coin or two.

Another hybrid approach is where there is a numerical score comparison, but the gameplay hinges around competition for a “majority” claim over a discrete number of scoring entities or categories. As claims can shift, the game end trigger typically is typically decoupled from the scoring system, and relies on a different mechanism to trigger the end game. For example, in the DECKTET game MAGNATE, the winner is the player able to have a higher card total in a majority of the 5 districts when the draw pile runs out. In Knizia’s SAMURAI, there are a finite number of three different types of scoring tokens that players fight over. When all the tokens are claimed (or all of one of the three
types is claimed) the game is over and players compare who had the most majorities across the token types. In TALUVA, players win by building all of 2 of 3 of their building types or having the most of 1 specific type built if the game’s stack of tiles runs out.

And yet another approach occurs in ANTIKE, which “hybridizes” the above two approaches. In Antike, players earn personality cards (essentially VP tokens) for doing certain things in the game (i.e. building X number of cities, conquering a temple, being the first to research a technology, etc.). There are a finite number of these VP cards players fight over (so it’s like Samurai in that regard), but there is also a threshold number of cards a player needs to win (determined by the number of players in the game) and trigger the game-end.

So we then have 5 approaches along this dimension:
- Scored
- Objective
- Threshold
- Majority
- Hybridized

Scoring Frequency – The “How”

The next biggest division (in my mind) in scoring systems is whether scores are accumulated “incrementally” over the course of the game versus based on a final board state only. While I’m presenting this as the second dimension, for reasons I will get into below, this dimension has the biggest impact on the game’s dynamics, pacing, and tensions as it relates to how scores are accumulated.
CRIBBAGE and QWIRKLE are examples where scores are entirely accumulated incrementally over the course of the game. QWIRKLE doesn’t use a score track, but players jot down their scores after each play and keep a running tally. The key in incrementally scored games is that points are generally accumulated immediately as an action is resolved. I.E., you place a tile and collect points.

A variation on incremental scoring is “Periodic” scoring, where scores are accumulated or tallied at a particular moment in the game, and often involves all players in the scoring process. GLEN MORE is an example, where a “scoring round” occurs as each of the three stacks of tiles are exhausted.

Compare the above systems against GO, for example, where scores are only ever tallied once both players pass in succession (and hence end the game). The scoring in GO is principally based on the territory you surround/control at the end of the game. Of course, in either of these cases, one can look at the “board state” or the tally sheets and see who is in the lead “at this moment,” but it isn’t strictly speaking a predictor of who is going to win.
There is a mixed approach too, which many euro players are likely familiar with. STONE AGE combines incremental scoring around the score track from buildings, but at the end of the game has “final” scoring that also adds the VP’s gained from all the civilization cards players have acquired during the game. RACE FOR THE GALAXY is similar in that your consume actions typically generate VP’s (and you take the VP tokens right away) but you also score the VP’s from your tableau at the end of the game. CARCASSONNE (I’m most familiar with Hunters + Gatherers) falls into the same boat, as you earn VP’s from forest + river segments during the game, but then get bonus points for your hunting groups and total connected rivers.

Hence, there are 4 variations on this dimension:
- Incremental
- Periodic
- End Game
- Mixed

Scoring Information: The “Who”

The last category, but no less important, is whether the scoring information is hidden or open. Hidden scoring, not surprisingly, comes in a few flavors as well. 
Most obvious are games like SMALL WORLD where players accumulate coins at the end of each player’s turn based on their control of the board. These coins are kept hidden so no one is supposed to know who is in the lead. A pitfall of this approach (in my mind) is that it is forcing hidden information which in practice is quite easy to track (awesome memory skills and/or actually writing down how many coins players take). So in concept the idea is to prevent players from knowing everyone’s relative position/score and presumably cut down on AP or leader-bashing, but how this mechanic is implemented doesn’t always work in practice. DOMINION is another example, as players buy and add the VP cards to their decks. It’s hard to remember what every player has taken, but in theory it can be tracked.

Another flavor of hidden scoring pertains to games in which each player may have a different/unique victory conditions and the condition itself is hidden and unknown. ILLUMINATI DELUXE is a great example of this in action, where if you play “Illuminati Groups Down” (aka hidden) no one knows which main group you are playing (although this can be deduced over time). Each group has its own victory condition based on acquiring certain types of cards or cards with certain properties. It is a bit of a guessing game to see what cards a player is going for and trying to deduce what group they are and whether they are close to victory. I’ll call this category “Masked” scoring.
A third flavor is what I’m going to call “Obscured Scoring.” In this approach, the scores are technically right there in the open for everyone to see (i.e. no hidden information), but the complexity of the scoring is such that it is quite difficult to keep tabs on the leader. For example, in a big game of 7 WONDERS, you can theoretically pause the game and look around the table to see who has how many points in what categories and make a determination of who is in the lead at that moment - whether you can do anything about it is a different matter! In this sense, the scoring is obscured because it is quite difficult to tell relative position. 
AGRICOLA is another example of obscured scoring. There are many scoring categories in the game and it takes a lot of experience with the game to be able to glance at another player’s farm and determine what their relative position is. Similarly, one might argue that many abstracts are fairly obscured. It takes a lot of skill in CHESS to be able to look at the board state and make a determination of who is ahead (or in the better position) in a close match. At the most obscured level, the question of who is ahead isn’t even relevant because it can’t be measured (and perhaps CHESS does fall into this category?)
Open scoring means that the relative positions of players are readily known/identifiable to all players. In CRIBBAGE you move your peg around the track, and I know and you know who is in the lead at that moment. I typically play 2-player SAMURAI with open scoring (since hiding it just adds time spent counting pieces), in which case we both know where we stand on capturing the three token types. There seems to be a movement behind open scoring in TIGRIS & EUPHRATES (perhaps prompted by the iOS version of the game?).

The distinction between truly open scoring versus some level of obscured scoring, is than in open scoring the scores are “broadcast with clarity” and it doesn’t take any extra legwork to sort out the scores. RACE FOR THE GALAXY is obscured scoring because while players have a visible pile of VP tokens, to get the exact score of a player you’d need to go around the table and add up everyone’s card VP’s, which isn’t immediately apparent. Likewise, many abstract games, particularly those that hinge on discrete (non-numerical) scoring systems, can be very opaque in gauging relative position until players are much more experienced with the game.

So, the lineup for this dimension is as follows:
- Hidden
- Masked
- Obscured
- Open

The Goal of it All

I’m going to re-iterate my intended goals for HEGEMONIC (which again may apply to lots of other games too):

#1) Encourage a high level of direct spatial interaction on the board
#2) Create tough decisions + reward strategic play
#3) Maintain tension and create a compelling “arc”

The biggest challenge I’ve faced as a game designer is balancing the achievement of these goals. While they certainly go hand in hand with one another, the specific scoring mechanisms that are employed can favor certain goals in lieu of others, and coming to a satisfactory mixture in the scoring mechanisms can be a real challenge!

#1) Encourage a high level of direct spatial interaction on the board

The Scoring Type (Dimension #1) relates to the “what” of the game. It’s “what” players are competing for or over, irrespective of how or when the scores are accumulated. And the “what” of the scoring system has a strong impact on the kinds of player interactions that are created in a game.

To focus a bit on Hegemonic, the above goal was critical to deciding what scores points in the game, as it drives the entire incentive structure for player choices. Many empire builders (Eclipse + Antike come to mind) have a number of different “things you do” in the game that earn you points. In Antike, you get points for your number of cities, for your number of temples, for your number of sea’s, etc… While that can be compelling, I wanted Hegemonic to do something different. Specifically, I wanted to points and scoring in Hegemonic to be focused around one principal scoring element that all decisions feed into, rather than categorizing decisions and actions into discrete scoring silos.

Furthermore, many empire building games follow a pattern where your empire expands out linearly into adjacent territory until you run into your neighbor. Then there is some fighting and the game is over. For Hegemonic, the goal was to create a space empire game that was very fluid and dynamic, with no discrete territories and where player’s empires were highly intertwined in different ways. While I’m terrible at it, Go was one of the inspirations. The idea that the whole board is open to you and you can theoretically play a piece in any open location was a compelling concept. Coupled with a galactic scale taking place over 1000’s of years, Hegemonic thematically started to gel. In space, you can go anywhere, the question is where “should” you go.

With the desire to create high levels of spatial interaction, scoring focused principally on territory control was the clear (obvious?) choice. I had explored a number of other scoring methods (one based on just your constructed assets, one based on discrete categories, etc.) but in the end they all fell flat because they didn’t incentivize interaction well enough. Without incentivizing interaction, defensive (turtling) strategies become more effective and pushed the game towards being an optimization game light on conflict. This wasn’t the experience I was going for at all.

So in Hegemonic, your scores are based on the relative control (power) you have in each of the many regions of the galaxy. There are a finite number of build locations within each region as the board evolves, so players engage in conflicts to maintain majority positions in whatever regions they are positioned to fight over. Most importantly, the incentive structure encourages widespread expansion and getting a foothold (preferably a leading position) in every region. It can be potentially a waste of an action + resources to attack someone in a region you already have a majority control of. But then again, this needs to be considered in light of players’ relative score positions. If someone is nipping at your heals for points, it might be worth it to try and oust them from a region if it widens the gap between your scores, irrespective of the absolute score you end up with.

Wrapped into the Scoring Type is a consideration of the end-game trigger. I explored a number of options for ending the game in the early design stages, but ultimately settled on a “decoupled” approach, where the end of the game was triggered by having a sector tile placed everywhere in the galaxy. Like TIGRIS + EUPHRATES, players progress through the game earning points, and must keep an eye out for when the end game trigger hits. And to some degree players can hasten the end of the game by taking extra explore actions. Often, it can be worth it to try and push a game to its conclusion if you are able to get a sizeable lead on your opponents.

#2) Create tough decisions + reward strategic play

Let’s consider Scoring Frequency (Dimension #2) in light of this goal. In an incrementally scored game, score gains are generally visible to everyone at the moment points are accrued (otherwise cheating can be an issue). More importantly, incrementally scored games often create a tough choice between “a few points now” or “more points later,” and the strategic + tactical depth of the game often hinges around this agonizing choice.

Many VP-engine building games (Race for the Galaxy, etc.) create a tough choice between “when” you switch from engine building to VP-churning. In QWIRKLE, there is an ever present risk-reward tension between using tiles to get a few points now and potentially giving your opponent a great follow up move, versus holding onto them to score big later when you can combine your tiles into a higher scoring combination; at the risk of not getting the tile you need on a subsequent turn. GLEN MORE is a good example where most of the play focuses around how to best capitalize on “this turn.” Yet you need to be conscious of your tile selection + placements so that you don’t sacrifice long term opportunities for future scoring by doing a huge move now when a series of smaller moves over the next few turns would net you more points.

Compare these incremental scoring cases against CHESS or GO, where keen players routinely make short-term sacrifices of their assets (stones, pawns, etc.) in order to set themselves up for a longer strategic move (the same is true for many wargames and miniature games). By avoiding incremental scoring, I often feel these games open up more emergent possibilities for surprising/daring moves because you aren’t held to the need to maintain a stream of point accumulation or maintain a specific engine to accumulate future points. Essentially, you are playing for the “endgame,” with all of your moves being an ends to a mean rather than the means themselves.

Another way of saying this, is that final scoring approaches can allow a lagging player to make an amazing and/or unforeseen move and take the lead in the final moments. Dropping the proverbial “bomb” on the game state. The extent to which a player builds towards this move strategically over the course of the game or whether salvation is handed to them by the forces of chaos is an important issue, ultimately playing into the broader mechanics at work in the game and the amount of chaos going on. Tigris is a superb example of a game where a single move can radically shake up the board state, yet the game's balance is such that other players can still follow up with an equally brilliant move.

The downside of final scoring methods, particularly in multiplayer direct conflict games (like Hegemonic), is that kingmaker concerns can be quite high; even more so if the ability to gauge relative player position/score is also high. If it is easy to spot the leader, it can be relatively easy for all other players to gang up on them in the final moments of the game. Illuminati (though I love the game) is a culprit here if played with groups “face-up,” as you know what each player’s goal is and it can be pretty easy to stop them from winning if other players gang up. Often the player in second place ends up being the winner. Hence the so-called Munckin Effect is born. Of course, the counterargument is that this is the nature of high conflict free-for-all style games and one should know this going into it!

So the real dilemma is how to reward sustained smart play during the game in light of tough decisions, while still allowing for bold moves and longer term strategic planning. This may be the reason why we see so many games with a mixed scoring approach, combining incremental scoring with final bonus points at the end. These mixed approaches start to create compelling tradeoff decisions and push players in one strategic direction or another. In Stone Age, do you go building heavy and score points on the track during the game or try to acquire lots of cards for a big score at the end of the game. In Race for the Galaxy, there is a tension between cranking out VP’s from produce/consume versus the opportunity to score a lot of points from synergies in your tableau.

For a long while I tested various final game scoring approaches in Hegemonic. The runner-up rewarded points to players at the end of the game based on multiplying two scores together. One score was points earned for control of the galaxy, boards the second your total power production. This was an interesting mechanism that created balance points between your level of expansion vs. raw power production. However, this approach was highly susceptible to bash the leader and kingmaker effects, as only the final board state mattered. You could “play well” all day and then get ganged up on the end and have it all tumble down. All the “patches” I explored to circumvent this problem added more noise to the game experience and bloated the rules. Additionally, the difficulty in calculating relative scores + position under this system led to a very opaque experience and one filled with a lot of mental arithmetic, both things I wanted to avoid.

So I combined the best of both worlds in my mind. I took the scoring mechanism based on board control (incentivizing interaction) with end-of-turn scoring for all players. Each turn in Hegemonic consists of three rounds of action selection + resolution. By placing the scoring at the end of all three of these rounds (periodic scoring), it creates a mid-term strategic objective to work towards. Compare this against something like CARCASSONNE, where you score on your turn after closing a feature. It is much more tactical because you can make a move and immediately score points from it. In Hegemonic, you make a move and have to wait 1-3 action phases to see if your choice pans out in light of what other players are doing. It injects a strong timing and sequencing element in to the game.

This also takes the inter-player dynamics resulting from turn order and turns the kingmaker effect on its head. Since player order can be largely dictated by either action selection or the arbiter (or both), the final round of action selection often sees players jockeying for the “last word” on a turn, such as making a final attack to grab a higher claim in a region. But again it creates an tough choice on another level too; do you take a later action to get “the last word” or do you act earlier, build up a stronger position, and then try and whether the storm? As these choices are faced each turn (typically 9-10 turns in a game), the periodicity of the scoring rewards consistent smart play over the course of the game, rather than piling the tension all on the end.

#3) Maintain tension + create a compelling “arc”

Scoring frequency has a strong bearing on a game’s pacing, arc, and tension. The strategic weight of an objective scoring game is all oriented around a singular end-point, rather than dispersed as a series of incremental shorter-term objectives. Under an incremental system, frequency of scoring rounds is critical and can promote (or not) cyclic mid-term objectives as players plan out the next few turns in preparation for scoring (as we discussed above).

As mentioned, many games feature specific “scoring rounds” that occur at defined intervals during the game. What’s the purpose of these? Compared to an “end state” scored game, these mid-game scoring rounds help establish shorter-term goals for players to work towards and possibly enforce greater interaction between players. GLEN MORE again comes to mind, as the two mid-game score rounds award VP’s based on player’s relative positions in three different scoring categories (whiskey barrels, chieftans, and special locations). Without these mid-game score rounds, the game could feel somewhat aimless as you move through the mid-game stages. Compared to a fully incremental scoring game, the lag between score rounds allows for a slightly longer strategic outlook as you plan towards the eventual scoring round.

Complimenting this rhythm, is how the Scoring Type contributes to the “arc” of game and it’s narrative evolution. Part of what allows an arc to emerge in a game is how the game end is triggered and how compressed or spread out the game’s arc is. TALUVA, which uses a majority scoring approach (build all of 2 or 3 of your base types) has a distinct but very compressed arc driven by the relative ratios of the three building types (2 towers, 3 temples, and 20 huts). The early game tends to be fairly open and opportunistic as players establish multiple settlements to support multiple towers and/or temples. The midgame generally finds players formulating a specific eng-game strategy, i.e., “I’m going to try for my 2nd tower and all my huts.” The end-game usually sees each player within a turn of winning, and player moves are driven to block their opponent while working towards an inexorable checkmate for themselves.

To take another example, CYCLADES has a very discrete win condition (build 2 metropolis’s) and I would argue a very compressed “arc.” It feels like you have entered the late game within a few turns of the game even starting, driven partially because you can go from zero metropolises to a win in a single turn if you plan carefully (and are a little lucky). ANTIKE is an example of a game with strong arc by use of a combined scoring mechanism. Early in the game, it is quite easy to get a scoring card, but as the board fills up, competition over a dwindling number of cards increases (pushing into the middle game). The final game is generally when a player is within a card or two of winning, and players begin actively working to inhibit others from getting a card and winning, even in some cases at the risk of losing their own ability to win on their next turn.

The scoring information dimension, in particular the opacity of the scoring, can also have a huge impact on the tension and arc of the game. In the case of “Hidden” information like that in SMALL WORLD (a multiplayer conflict game), I’d contend that it does the game a disservice to its strategic possibilities and potential tension. Knowing clearly who the leader is in the game gives players a clearer purpose and set of tradeoffs in the game. Do you attack the leader while trying to gain some coins in the process to narrow the gap or do you just maximize your own possible points for the round? That fundamental question is much harder to answer if you don’t know who the leader is.

Even more so, tension in games so often derives from understanding your position and making the best move to advance your position, in the face of some future uncertainty. Normally in SMALL WORLD, the uncertainty is focused around what your actual position is (Who’s ahead of me? Am I in the lead? Am I way behind?), which is backwards to me. As a result, I often feel the ending of a game of Small World is anti-climactic because who the winner is feels too arbitrary. I’m not seeing openly how close the players are tracking each other in the scoring over the course of the game and miss that entire tension caused by relative position.

The purpose of this “falsely” hidden information is no doubt to make the game feel lighter and less competitive. Likewise, hidden tokens in multiplayer SAMURAI makes it harder to tell who the leader is, and by default pushes players to just optimize their own turns. But that assumes people aren’t playing competitively and paying attention. If you are playing hard to win and can keep track of who has what tokens, it’s a huge advantage in a supposedly casual game to do so. Why not just keep the scoring open and let the game be what it wants to be?

In terms of Hegemonic we have already established that the scoring system is based on controlling regions and those points are accumulated at the end of each turn. You can probably guess where I’m going, but I felt strongly that the scoring needed to be open and visible to support tension and a good arc. The score track allows players to immediately see relative position, and derive a strategic direction from that knowledge. The territory based scoring forces players to expand and creates no shortage of opportunities for conflict.

So the tensions in Hegemonic are created both in the lead up to the scoring at the end of a round (are my choices going to pan out based on what other players do?) and as a function of the varying rate at which players are accumulating points. Fast expanders might surge ahead in points early in the game as they grab more territory. But empires will be more spread out and fragile. A player that builds up concentrations of power earlier gets fewer points early on, but can make a strong and defensible push later in the game.

Conclusions + Disclaimers

I think I bit off more than I can chew in this post! I’ve been pecking away at it for a while now and just needed to get it off my shoulders. So I apologize for the somewhat rambling and aimless nature of it.

That said, I hope that I’ve shed some light on different scoring approaches that designers can employ, and how those choices manifest in the game’s dynamics and the kind of decisions and tensions facing players.

It has been a frustrating thrill to consider this milieu of scoring options in light of my own design efforts. Being attentive and aware of how your scoring system impacts player choice, dynamics, and the game's progression is key to make a game that is exciting to play, and more importantly, exciting to play again and again!


No comments:

Post a Comment