#drafto

I'm still working on Drafto.

Playtesting over the last couple of weeks has honed the game into something everyone seems eager to keep playing.  I think the rules are more or less locked down, now - it's just the individual objectives that require tweaking. Once this is done, I'm going to write everything up and release it under a Creative Commons non-commercial license, allowing people to freely reproduce and modify the game (as long as you credit me, and don't sell it).

It's good.  I look forward to sharing it with you.

Way back in the...two weeks ago...I discussed a set of pillars proposed by Paul Sottosanti that he saw as important mechanics drafting games should possess.  Some of these I agreed with, and others I felt were more optional.

Here is my counter-set, which for lack of a better descriptor I'm going to refer to as the core of drafting games - central mechanics which I feel are important enough that all good drafting games should possess them.

Some cards should be clearly more valuable than others. This is to do with the size of the decision set. If all cards are of roughly equal value, then the player has a number of choices to consider equal to the size of the pack - which is problematic unless packs are very small. Conversely, if cards have clearly different valuations (up to a point), the player can ignore all of the weaker cards present and have a much smaller set of "true" decisions to choose from. This is what occurs in Magic. Players don't start with any external direction, so the first pick of the first pack is ostensibly a choice from 14 unique cards. However, each pack is stratified into 10 common cards, 3 uncommon cards, and 1 rare card - the higher rarity usually correlates with card power, and this allows an inexperienced drafter to ignore most of the 10 commons and face a much easier choice of between 4-6 cards.  (Veteran drafters ignore rarity altogether and will be able to perform a similar pruning operation on the decision set based entirely on learnt card evaluation / experience.)  As it's important for players to understand and enjoy your game before they become proficient in it, I feel this overt strength signalling is critical. That, or tiny packs :)

Cards should have different value to different players - and each player should be able to approximate the valuation functions of his or her opponents. This was on the original list of pillars (albeit in a simpler form), and it's a property I strongly agree with - in my opinion, this is the whole point of drafting games.  If every player values all cards the same at any given point in the game, then you lose all the subtlety and nuance of the genre which you would otherwise get for free. My addition is to specify that each player should have the capability to be at least partially aware of the valuation functions of the other players (by "valuation function", I mean the criteria that player is using to judge how valuable a particular card is to them if they were to pick it).  This will usually be a consequence of the weaker condition, as cards valuable to one player will be missing from packs they pass, allowing that player's valuation function to be inferred by their absence (this is known as cutting a particular subset of cards).  However, I felt it was worth stating explicitly, as other mechanisms could be introduced to strengthen this property. For example, consider the strategy known as wheeling or tabling (as in "to wheel" or "to table" something). Imagine if, given a pack, you're presented with a choice between two cards that you'd ideally like to draft both of. If all players value all cards equally, then whichever one you don't take, the next player will - so you take the one with the highest immediate value (which isn't a real decision at all, because it was strictly better than the alternative).  If all players don't value all cards equally, then the correct choice suddenly becomes less obvious - whichever card you don't take won't automatically be taken by the next player (or indeed any of the other players), making the decision process involve some non-obvious function of how valuable the guaranteed card would be to you now and how likely it is to make it all the way around the table so you can pick it in the future (based on your beliefs about the valuation functions of the other players). Another strategy this enables is hate-drafting, which is when a player takes a card that is not particularly valuable for them, but is something that they believe is of significant value to an opponent. As the players are in direct competition, this can often (but not always) be a correct action - the accuracy of the player's beliefs about the valuation functions of their opponents is a key component of this, which makes deducing these a skill intensive part of the emergent gameplay.

Some cards should have unknown value. Let's be honest - expecting a new player to immediately grasp the intricacies of forming beliefs about other players' valuation functions is a bit of a tall order.  While these are cool parts of drafting that reward invested players, you still need enough of a game to be present for players who are unaware of this, else they won't stick around long enough to find out. For these players, drafting becomes more of a puzzle than a game, as their opponents are irrelevant. To introduce meaningful decisions into this puzzle, the player must be presented with several attractive choices at each step, and it must not be obvious which (if any) is the best of these choices. To use Magic as an example (again) - assume a player is completely accurate at assessing how powerful a card is, apart from its cost.  The cost of a card affects how easy it is to play with other cards that have been drafted - for example, if you've drafted mostly green and white cards, picking a card with a heavy black requirement is unlikely to be a good move, even if the card itself is very potent.  The kicker is that this doesn't just depend on the cards you've already drafted, it also depends on the cards you have yet to draft.  The true value of the card won't be known until after you've finished drafting! In a more general sense, if you know that when you pick a card it's worth a fixed amount, the game devolves into calculating this fixed amount for all of the cards in the current pack, and picking a maximum.  If the value of a card is (partially) unknown, then the decision incorporates all the other information you've been exposed to so far in the game - i.e. "will I get enough other cards to make this card work based on what's already been passed to me / what I've passed"? This doesn't need to apply to all cards in the game - this mechanic adds complexity, so if you want a more accessible game it's worth using this sparingly - but it is a good way of adding clearly signalled interesting decisions to a game.

  I think that's it.  Everything else I could think of seemed to be overly specific, optional or adjustable (e.g. visible objectives, starting asymmetry, amount of complexity).

It's worth noting that Drafto v1.0 satisfied all of these (but was still terrible), as does Drafto v2.0 (but is much better).  So too do Magic and 7 Wonders, though in these cases I don't think they do a particularly good job of clearly signalling card value (point 1) without a reasonable amount of prior knowledge about the resulting gameplay - this doesn't mean I think they're bad games, it's just that the drafting portions of them perhaps don't work as well as they could do because they have this other, more complex game attached.

It's worth pointing out now that I consider this first prototype a failure.

I mean, that's to be expected: it was a first attempt at a problem that hadn't been solved to my satisfaction in any released product that I'd come across, which implied that it was hard, and here's me attempting this as a hobby in the downtime from my actual job.  But I still think it's worth saying - I frequently get pissed off at science for encouraging only positive results, when the negative ones matter just as much, even if only to stop people making the same mistakes again.  The good news is that I used this as a stepping stone to something better - but that'll have to wait, because I want to tell this story chronologically.

It's probably best if I just describe the prototype now, then discuss my choices behind it - so without further ado:

-----------------------------------------------------------

  Draft Deck - 125 cards

This was equally split between five colours (red, green, blue, yellow, purple), each of which was split into five guilds (soldiers, workers, scholars, leaders, artists), each of which had five cards of different values (1, 2, 3, 4, 5).

For example, there were five green soldiers (1, 2, 3, 4, 5), but only one red scholar of value 4 - making each card unique.

Objective Deck - 30 cards

This was a separate deck consisting of threshold conditions and point rewards.  Examples:

Rational - Collect 10 blue cards - 20 pts

Defensive - Collect 15 soldiers - 30 pts

Peerless - Collect exactly one scholar - 20 pts

Setup Phase

Each player was allocated three objectives at random from the objective deck (or chose them via a pre-draft - but for initial testing allocation was random). These were private information. During the scoring phase, a player was allowed to score any/all of these as long as they'd achieved the threshold condition.

Draft Phase

Each player drafted 30 cards from the draft deck - over three packs of 10 cards, passing left/right/left.

Scoring Phase

Each player revealed and scored all achieved objectives, plus the total value of the cards they'd drafted (making a 4 twice as good as a 2, and a 5 five times as good as a 1. MATHS).

-----------------------------------------------------------

I chose the deck size / draft structure due to personal preferences on the time of a session (maximum 15 min draft, 2 min scoring).  I chose the concepts of colour and guild because I wanted overlapping axes for players to collect, and I could make these very visual in my physical prototype by varying ink colour and symbol.  I chose the division into fives because four seemed too few and six seemed too many.

The draft deck was essentially the common "filler" which existed only to support the objective deck's "rares".  I intended the bulk of the decisions to come from trading off the raw value of each card versus the potential gain it provided to achieving a player's objectives (through the less tangible colour/guild).

As two of my three reference points (7 Wonders, Sottosanti War, not Magic) suggested an initial external asymmetry, I decided to try allocating the objectives at game start, rather than introducing them as part of the draft (as Magic does). This gave players an initial direction to aim in, with the objective difficulty being such that achieving them wasn't guaranteed unless they were being ignored by the other players in the draft - the intention being to identify which objectives were "open" and maximising those and raw value in general.

Comparing this against the seven "pillars" mentioned in my previous post yielded:

Wildly different card types - ? (objectives/draft decks - is this enough?)

Rare cards - YES (objectives provide direction and different experiences)

No Memorisation - YES (three objectives is enough to hold in your mind at one time, and the draft deck is simple enough that you don't have too many moving parts)

Not drafting every card - IRRELEVANT (all cards are positive, apart from for some weird objectives, but we did experiment with variants where the last few cards were discarded when we had fewer than four players, just to increase the number of cards seen)

Card have contextual value - YES (objectives give cards differing value to different players, and with expected value that's hard to calculate on the fly)

Reacting to opponents - ? (hopefully colours/guilds being cut could be identified, and objectives abandoned/adopted)

Initial asymmetry - YES (objectives)

  -----------------------------------------------------------

This seemed to be in a state worth playtesting, so I cajoled three helpful work friends into helping me out over a lunch break or two.

Things immediately started to go wrong:

Players weren't happy that they didn't have any choice over the objective allocation.  As objectives were worth varying numbers of points based on "difficulty", this initial asymmetry meant that some players were disadvantaged right from the start, and had no chance of winning overall.  While this could have been fixed by some very careful balancing, this felt like it would have taken a lot of work.  Pre-drafting the objectives helped a little, but not enough.

I'd rather idiotically included objectives such as "collect 10 of any one colour" or "collect 10 of any one guild".  I wanted these to encourage dynamic strategies - adapting to what was being passed to you - but instead they just stacked with other objectives (e.g. "collect 10 red") to compound your bonus.

As I wanted the objectives to be fairly easy to keep track of, there wasn't much design space available to make them unique.  This meant that guild and colour objectives felt very similar, and by extension so did guilds and colours.

Because the objectives were hidden, and cards were distributed along two similar axes with often high rewards for taking an off-objective pick (4s and 5s), it wasn't at all clear what the other players were doing, and therefore it was extremely difficult to adapt to it.

Because all the choices in the draft were between very similar cards, the game was simple to pick up - but it wasn't as fun or as deep as I'd hoped.

We tried fixing these on the fly, by (e.g.) only allowing one objective to be scored (and making them harder), or tweaking objective thresholds/rewards, but nothing really felt right.  We briefly discussed increasing the asymmetry between colour and guild by making guild objectives more interesting and granting a static +2 bonus to each player's most drafted colour, but we never got around to trying it out.

So, I wanted to design a drafting game.  I was aiming for approximately 30 cards to be drafted over about 5-10 minutes, with a very quick (1 minute) scoring phase at the end.  I wanted it to be easy to pick up / elegant, but with sufficient strategic depth and nuance, as these are the kinds of games I like best - I was counting on the randomness and hidden information inherent in the genre to obscure the complexity.

As I mentioned in my last post, Paul Sottosanti came up with a list of seven pillars of drafting game design, which seemed like a good starting point for my own attempt.  These were:

Card types that do wildly different things so that the draft choices are not just straight comparisons. I don't quite agree with this exact statement, but I agree that cards need different contextual value (so that they do "wildly different things" for different players at different times).

Rare cards that can excite drafters and give them a direction. The fun of drafting comes from identifying the correct strategy from increasingly visible information.  "Rare" (read: powerful, restrictive) cards do this, but this could also be achieved through multiple hidden objectives that aren't all satisfiable (7 Wonders essentially does this).

Not having to memorize or continually look at a bunch of hidden cards. I don't think this is actually necessary, as Magic drafting is very successful without constant visibility (drafted cards are placed in a facedown pile, and are only able to be reviewed between packs). It does lower the complexity of the game, though - but this can be achieved through other means, making the memorisation aspect a feature, not a bug.

Not having to draft and/or play every card. This depends on the scoring mechanism used.  You certainly don't want a system where drafting cards can be detrimental, but as I wanted my scoring mechanism to be rapid I didn't want players to have to solve a subset selection problem at game end (as happens in Magic).  If drafting cards can be detrimental, then throwing away the dregs of each pack seems sensible.

Cards that have different amounts of value to different people. This is the crux of drafting games - without this every player is taking the numerically best card in each pack, and you cease to have a game at all.  No arguments here.

The opportunity to change your strategy based on what other people are doing. I don't think this is essential for an enjoyable experience, but it's certainly a very good thing to have.  It turns what would otherwise be a competition with n participants into an actual game.

Players should start the game from different positions. Asymmetry/variance is important, else all players are doing the exact same thing in every game, which will quickly become boring. However, as drafting games are generally going to involve random sets of cards during the drafting phase, you should get this for free.  Of course, starting the game with different predefined objectives gives you asymmetry and a goal (see point 2) all at the same time, so it's worth considering.

So.  It seemed apparent to me that while these points were good things to bear in mind when designing a drafting game, they were far from essential - apparently I'm not one to learn from the mistakes of others, especially when they're clearly annotated (thanks Paul!).

I'll write my own list of lessons learned in a future post - it's worth mentioning here that my design is not finished, so any "lessons learned" will likely be in-progress assessments and should be taken with a pinch of salt (of course, this depends on when I write it...).

So, content.

I've played the card game Magic: The Gathering for a scarily large number of years.  As there's a nonzero chance that you haven't heard of it (o mysterious wanderer of the internet), it's a magical duelling game where you have a deck of spells and you use them to brutally murder your opponent, in what's been described as a cross between poker and chess.  It's quite well-known, exceptionally well-designed (nowadays), and I'm sure I'll discuss it more in the future.

Anyway, Magic cards come in booster packs of 14 randomly assorted cards. While modes of play exist where you construct a deck out of a subset of your entire collection, there are various "limited" formats where each player constructs a deck from just a small number of packs. The most popular way of doing this is through drafting - 6-8 players sit in a circle, open a pack each, pick a card, then pass their pack to the left (repeating until this pack is empty). This process is repeated with a new pack passing to the right, then again passing to the left once more. This gives each player 42 cards from which to extract a ~23 card subset (+ ~17 freely available basic land cards = a deck).  There are other ways of drafting including methods with fewer players and a different number of packs, but this is the default - it provides a good mix of hidden information, variance, and availability of strategies.  Drafting is interesting because each card has variable worth dependent on the other cards in the same deck (e.g. it's easier to cast spells of the same colour rather than spells of all five colours), which means that each player develops a different preference ordering on all cards available that dynamically adjusts as the draft progresses.

Drafting is a meta-game - the winner is still decided by playing good ol' fashioned Magic against the other participants, but the pre-games step of slowly collecting a deck is crucial to this "scoring phase", and requires a great deal of skill to navigate successfully. Players have to pay attention to what types of card they're being passed / are passing so that they can avoid cutting each other off by being in the same strategy, in addition to maximising the chances of them having a cohesive deck. A lot of Magic players (myself included) actually prefer drafting to playing.

This is the motivation behind "Drafto".

I wanted to create a game that appeals to players for the same reason drafting Magic does, but without the additional constraints that come with drafting Magic (having to construct a deck, having to play several hours of games to "score" the draft, having to assemble 8 players, having to buy 24 total packs). Don't get me wrong, playing Magic is still incredibly fun and rewarding, it's just that sometimes you don't want to go through the additional rigmarole that comes with drafting - say you've only got three players who have 15 minutes to spare, for example.

Other board games use drafting as a part of gameplay (the most popular of which is probably 7 Wonders), but I've yet to discover one where drafting is the whole game.  I did manage to find this illuminating blog post by former WotC designer Paul Sottosanti, which discussed his experiences in trying to design a "War" type drafting game - I'll be referring back to this when discussing the initial design for Drafto in the next post, as this seems like a good place to stop for now.