Accepting incremental progress

A couple of tests of the ideas described in the last couple of posts show some promise, and after a long period of dramatic changes after every test, the core structure is stable enough to leave it intact and go into tweaking mode for a while.  (This is especially good news for me, since the last 10 iterations or so have each required pretty major component overhauls.  It would be nice to have a change that is mostly just rules instead of components!)

Not all of the ideas panned out.  The idea about only being allowed to collect cards from certain categories, reflecting your character's strengths and limitations, hasn't yet made it to playtesting.  I decided to try the less restrictive approach first and see if it worked well enough, and it did.  But it nevertheless works to think of your hand limit as 'iconic characteristics and items that are associated with your character', which can include not merely equipment but also quotes and character traits and such.

Unfortunately, the temple really did not work, in the sense that there just wasn't much dice-rolling.  This was as I had feared -- if the number of dice you roll is based on your holdings relative to the other players, then not many dice are rolled when we're all clustered in our holdings, and that was what consistently happened. 

The idea holds some promise, but it needs a tweak so that more dice are rolled.  The one I'm most interested in to have a marker (maybe representing the enemy) that starts at 1 and moves up by one each temple card (of which there are 5).  Each turn, you see how many symbols you possess that match the current temple card's peril, and place your marker at that position on the track.  Then the number of dice you roll is either the difference between your marker and that of the player with the most symbols, OR the difference between your marker and the enemy's, whichever is greater.  This adds an arc to the temple, since the enemy gets stronger and so later cards in the temple are strictly more important in the temple from a preparation standpoint.  Except!  While everyone else is viewing and preparing for those, you can view and prepare for the earlier cards, forcing the other players to take more rolls on those early cards, which gives them more noise, which has a cumulative effect.  Unless they do the same thing, in which case no one takes much damage early on but no one is very well prepared later on.  And so on.  So this arc might add some interesting decisions, perhaps.

Although shared encounters outside the temple were supposed to reduce length and tedium, they don't succeed with respect to the former.  The game still takes about 2 hours to play; ironically, a live 3p session took longer than my solo 3p session!  Latency isn't a big issue, and no one felt bored, but it still just seems like an awfully long game for what it is.  Player decisions whether to join an encounter or not, and how much to commit, and then whether to stay in or get out, and then which rewards to claim, all just take time.

I'm considering whether we might at least make all decisions simultaneous outside the temple.  Basically, choose simultaneously whether you're in for the encounter and what if any you'll contribute to it, and then choose simultaneously after each roll who is in or out.  This last is a bit closer to Diamant than I had wanted to come, but the decisions aren't informed by the same considerations as those in Diamant so maybe it's ok.  The investment decision, at least, seems ok to convert to a simultaneous decision.  My original thought in making it serial was simply that, since players will have the opportunity to back out and claim rewards in turn order, it's better for players to also invest in order so earlier players have to sweeten the pot a bit if they want later players to also join in and contribute.  We could have a compromise of this, perhaps -- the first player must be in, so perhaps he must reveal his level of commitment to the encounter, and then other players decide simultaneously based on that what they're going to do.  Not perfect but maybe quicker.  Another benefit of simultaneous decisions is simply that it's more congruent with the temple phase, in which players simultaneously decide whether to invest any cubes in the current temple card.  So if "simultaneously allocate cubes" happens in both phases, maybe that also helps make the game easier to learn. 

Needless to say this is all quite different from the v7 rules, and I'm not sure what to make of that.  The encounters do seem more immersive and suspenseful than the v7 rules, but they've required a lot of changes to the other aspects of the game to make them work in a reasonable time frame, because they eat up a lot of time.  The simplicity of the theme cards is back, though.  The one thing I guess I miss most was the ease of setup -- shuffle some piles, grab a card from each, boom, go.  But I think for features like this that we've lost, there's a lot more direct player interaction, and I think that's a net gain.  Now, as for whether it works equally well across all player counts...

Return of the son of the temple

A strange inversion has happened.  Several posts back, the temple system entailed each player rolling dice against each temple card until they passed a card-specified threshold, and the idea was to do this in as few rolls as possible so as to reap the benefits of position; thus, it felt like there was an element of jockeying for place.  (This led, conceptually, to the 'running time' track, which players unanimously have liked).  The map system at that time still involved individual isolated challenges.


Now, the map phase is trending towards group encounters but the temple system is back to feeling somewhat solitaire-ish:  you roll dice as before, but now dice are bad and the number you roll determines how much harm you suffer.  So you want to roll less to accumulate less harm in the aggregate, but our rolls have little impact on each other.


While I think there's room in the world for a game that shifts gears, I think that to go from a tense interplayer experience to a solitaire feel is a shift in the wrong direction.  It feels a bit like a skiing race -- everyone runs the course and we see who did the best.  There's no jostling in skiing.  What we want is more of a roller derby:  there's a goal, but there's also jostling as we all try to achieve that goal.

This might be achieved simply by allocating the dice in a different way.  Instead of "roll four, less one for each symbol you hold that matches the current temple card", it could be "player with the most symbols rolls nothing and each other player rolls dice based on how many fewer they have than the player with the most". 

There are two possible issues.  The first is that this makes the temple difficulty entirely relative.  If everyone is poorly prepared for the temple, it's functionally no different than if everyone is extremely well prepared for the temple, because you're only rolling dice relative to what other players have.  There's no sense in which the enemy or the temple itself are going to grind you into dust if you're all unprepared.  This needs thought; it may be a problem.

The second is that even though the rolls are relative, we're still not exactly jockeying or jostling -- we're just seeing who is best prepared and letting things play out accordingly.  This one, at least, I think I see how to address.  I may have mentioned previously that at each temple card, you get to place a cube on an equipment card of your choice to double its effect.  This rewards knowledge, since you presumably will double a card that matches the temple card's peril if you know what it's going to be.


But!  Add to this a twist that you have (say) four cubes for this purpose, but every time you use one it is discarded.  And, any cubes still in your possession count as green cubes in the hubris challenge, which are what you need to acquire the white dice that you'll roll to purge hubris.

This may add some jostling in the temple, because even though you may know what peril a temple card contains, and even though you can see what peril symbols your opponents possess, you don't know whether THEY know that the next card is "fear", and you don't know whether they're willing to spend a cube to double an equipment card.

Moreover there are some equipment cards that are 'parasitic' of other players' peril symbols, and others that can cause whammy effects on other players on certain temple cards.  So these combine d with the cube allocating may give enough interactivity that the vibe of the external phase is preserved.

If I have knowledge of the next temple card and have the symbol for it, why not just pay several cubes on the right symbol to really hit the other players?  I think the rules should allow this but also impose the restriction that the most anyone can be forced to roll is the number of temple dice, which is probably four or five.  So there's not much upside to dropping several cubes on a single card unless you REALLY want to win that particular card.

The next question becomes what the temple dice do to you when you have to roll them.  I still like the idea that you either generate noise or fall into a trap, but that equipment cards help with the latter.  Not sure what noise does, but maybe it's something simple like, the player with the highest number of symbols places a marker on a track -- that is the 'tolerable noise level' for the current card.  Anyone with more noise than this after rolling takes damage, and the player with the most noise (regardless of the threshold) takes damage as well.  Noise is cumulative but maybe you can also spend a cube to reduce your noise (as opposed to using it to double equipment cards).  This is still a bit fiddly but I like the theming of temple dice representing you speeding through the temple in a risky way, which might result in you tripping on a trip wire or alerting the Nazis to your presence.  In other words, there are two kinds of dangers in the temple and they function differently and the dice results allow you to be at risk for both of them, but in an asymmetric way.  But maybe making them more symmetric is worth it if it makes the game easier to learn/play.

Simpler cards and complex characters

We had a three-player playtest of the full game with fully-simultaneous shared turns in the map phase.  There were some fun filigrees.  Each encounter requires two points of 'investment' for each white die the group will roll.  Players contribute in order and then, when a player cashes out of the encounter, they take their investment out as well.  So there were times when a player would decline to join in or would pull out at an inopportune time so as to leave the other players high and dry.  I think this is important and offers something different from Diamant-esque doublethink.

I don't think the game was much shorter.  To be sure, it was a learning game, but we did try to play quickly and still it took maybe 90 minutes of playing time.  There was generally a preference for cards over information as rewards; the game still seems to reward having lots of cards as a way to navigate the temple vis a vis having lots of info and just the right cards.  It may be that too few cards are available per encounter.  Currently it's just two and they're so precious that they're scooped up quickly.  Perhaps with more cards available, it's possible for all of the encounter participants to get one and so it's more about who chose the right cards. 

The encounter has two parts, the 'investment' part in which everyone decides to be in or out, and the resolution part, where we roll dice.  The latter requires a few too many rolls but that can be sped up by tweaking the dice.  But the investment is also slower than it should be, and if I had to guess, I think it's that there's a bit too much that each player can do and so reminding yourself of all of your options is still eating up too much game time.  I think the cards can be simplified further such that each card does exactly one thing, ideally a thing that can be represented by an icon.  If we look back at version 7, it's incredibly streamlined; all of the equipment cards have one challenge category, and all of the theme cards have 1-3 icons for their clue categories.  The amount of information conveyed by v7 of the game was really quite minimal.  I think we probably need to get back to this.

On the other hand, a suggestion was made to diversify what some of the cards do, in particular the starting cards you're given.  If we're all Indy and we all have a Whip, Pistol, Satchel, and Fedora, then we're all the same; some asymmetry might be more interesting.  That might not be so hard to do.  Perhaps one character has a pistol (good at fight encounters) whereas another has something that makes them good in circle cities and another still good in the Middle East.  These might all benefit an encounter in Cairo (say) but won't all be as effective in different places.


One consequence of simplifying to one-effect-per-card is that you presumably need more cards in the game to get the same net number of effects, and this could mean that everyone will acquire many cards over the game.  A glut of cards, though, adds to the decision space. One simple counter is to impose a hand limit, which will force you to keep the cards you especially plan to use, and in particular to transition from cards that are map-useful to cards that are temple-useful.  (I guess each card could have a map phase effect and a temple phase effect, but again, that's getting away from the simplicity that I think the game probably really needs.  One effect per card!!!!)

But, this led me to ask, what does the hand limit represent?  In an adventure game, the obvious thing is your 'carrying capacity'.  But this is a game about an adventure movie, and is there a more cinematic option?  Perhaps the hand limit could represent the attributes that the audience will associate with the character -- a catch phrase, an accessory, an article of clothing, a weapon.  Trying to have too many things would lead to a confusingly articulated character, so the cap says how many things the audience can be expected to keep straight. 

This could be implemented in a way that allows for a secondary effect.  Specifically, the cards could be categorized in a way that's distinct from the effect they create.  Perhaps this is color-coded.  And you could have a player mat that has seven (say) colored boxes, indicating what categories of cards you can carry, but each player's mat could be different based on their archetypical character.  So for example "very strong person" may have three of their seven boxes be for 'equipment' cards (brown, let's say), whereas "slippery eel character" may have only a single box for equipment but two for 'alliances' (pink, let's say).  'Very strong person' can choose what equipment he's holding and 'slippery eel person' can choose which entities he'll ally with, as represented by the cards they each draft, but the competition for cards is informed in some ways by these considerations. 

I'm a bit worried about simplifying one thing only to add additional complexity somewhere else.  But, I'm intrigued by this idea as a way to shape player strategies while still allowing for flexibility.  I'm not sure I've seen this implemented previously, although I suppose that it must have been used somewhere.

Immediacy is important

We had a chance to playtest a few turns of the simultaneous system described in the previous posts.  In some ways it works -- it does reduce the total number of turns overall, it does provide some brinksmanship, it does speed up the resolution.

But there was one important thing that it lost.  When you're facing an encounter in the previous version you roll your white and red dice at the same time -- so you get some good results, but also the bad results at the same time.  Thus, with each roll, the enemy is getting closer and closer and the tension mounts.

In this version, we each take turns rolling, then the enemy rolls, and based on that brief description you would think that the tension is similarly pretty palpable.  But what really happens is, for each roll, you  collect your dice, then roll, then increment the track, then decide if you're going to get out, then (if you do) collect your rewards.  If you take a reward it may be upwards of a minute before the next person rolls while you do that.  Then the process repeats for each player and THEN, at last, the enemy rolls and maybe moves.  And then the process repeats.  The encounter just takes long enough that the sense of suspense is hard to sustain.

We had a very nice playtester suggestion that I think might help with this.  He proposed viewing the encounters as semi-cooperative.  Meaning that, at the start of the encounter, each player reveals how much they will commit to the encounter (equipment cards or black cubes).  Then the number of white dice the /group as a whole/ will roll is based on what everyone contributed.  Then, roll all those white dice, and the red dice, all together, and move a /single/ marker along the card's success track.  Each person can decide whether to get out after each roll (receiving rewards for whatever space the group marker is on), and a player who gets out takes their equipment card(s) with them, so the remaining group is a bit weaker. 

To preserve the quick resolution, I think it would be that you don't receive rewards until the challenge has ended.  So the encounter is the action sequence where Indy is frantically trying to grab the headpiece from the burning building, and then in the aftermath he looks at it and actually reads what it says and gets to think about it.  So thematically I think this approach is ok.  And we also thought that these temporary 'alliances' seem basically compatible with the theme -- in one sense we're all participating in the scene to make it exciting, in another sense our characters are perhaps temporarily cooperating to make the challenge a success but you're still trying to get maximal individual advantage out of the challenge.

One other problem was homogeneity -- everyone basically ends up with exactly the same information.  Perhaps it could be that if you get out early, you put a marker on the info you're going to claim, and anyone else who puts a marker on that same card when they get out must pay some cost to access it (or maybe they just flat out can't access it, although that probably won't work -- not enough info to go around in that case).

Small changes are sometimes big changes

In the previous post I proposed what seemed like a simple change -- switch from individual player turns to simultaneous 'group' turns, whilst keeping essentially the same turn logic.  However, as I'm fleshing out this idea, I realize that this actually fundamentally re-wires the entire concept of the game.  Specifically, with individual turns the game is basically an efficiency puzzle with a healthy dose of risk management tossed in:  you want to parse the information puzzle efficiently, and while luck plays a role in that, managing your luck is important to cracking the information nut.  Moving to simultaneous turns inevitably moves the map phase of the game in the direction of being a game of brinksmanship, again with a healthy dose of risk management tossed in.  It has to be this way, in fact, or the simultaneous turns make no sense. 

The encounters in their present form already have an element of press-your-luck tension.  You want to keep rolling because the more checks you get, the more rewards you get.  But you want to stop rolling before the enemy reaches your city lest bad things happen to you.  This will still be the case in a multi-player version, but it can't be the only thing; it seems that there must be some degree of competing not just against the enemy clock but also against the other players who are in the encounter.  One problem is that this has already been done to great effect in Diamant, and there's a real risk of just creating a clone of that.

I think there are three ways around this, and this system should use all of them.  The first is that the rewards are asymmetric.  For each encounter, there are two equipment cards available as rewards, and the ability to look up clues from a subset of the temple cards as determined by the theme card in the city.  Second, taking a clue lookup means you remove a cube from that temple area's 'bin', and either discard it (earning you a green cube, helpful for the final hubris challenge) or add it to the current encounter, making the enemy more powerful -- for everyone who is still in!  Taken together, getting out early means you're probably going to get less stuff, but you are less likely to get closed out of an equipment card and/or can make the challenge harder for the other players. 

The third thing is to add in 'effect' symbols to some of the encounters.  'Effects' are a new feature of the temple, and basically allow you to, essentially, use the machete to destroy the rope bridge.  We could add this kind of effect here as well.  Some encounters could have two tracks, one which is 'easier' (i.e. gets to more rewards more quickly), but has an 'effect' symbol, such that, if another player reveals an equipment card with a matching effect symbol, all the players on that path take some damage.  So in addition to 'stay in/get out', there's a decision about which path to take based on what the other players might be able to do to you.  Presumably, only players still in the challenge can trigger an effect, so if you have the right equipment card, you may want to stay in simply to have the opportunity to whammy the other players.


One nice thing about these additional considerations is that they afford some additional 'handles' with which to differentiate the encounters and maybe make them feel more thematic.  Version 7 was just 'roll a die and that tells you the challenge category' -- quick and effective but quite bland.  Version 10 had encounter cards that showed you the scene you faced, but it was just chrome -- the 'basket game' or 'x marks the spot' weren't functionally different except in the challenge category and difficulty of the challenge.  But now with the success tracks and a couple of simple effects it's possible to make them feel different.  For example, maybe with the 'basket game' there's an 'explosion' symbol late on the track, so staying in too long runs the risk that another player will blow up the truck, costing you a life cube; whereas 'X marks the spot', on one track you can get to the answer with fewer total rolls but must spend a 'time' examining the room, whereas if you charge in you will have a harder time finding the tomb AND might put yourself at risk of another player throwing a 'lever', which lights the oil in the catacombs on fire...


One semi-related and also nice thing about the simultaneous system.  As I might have mentioned in the last post, each player would have several location cards, and each turn, the player who is furthest back on the time track plays one of these cards, indicating where that turn's encounter will take place.  There will only be nine turns total, so this controls the game length and makes it the same for all player counts.  But more importantly, it essentially eviscerates the 'do nothing' strategy.  If you try to sit out too many of the encounters, you'll end up in the back of the time track and will have no choice but to take turns, and this is compulsory -- you can't pass if you're last on the time track.  In light of this it might be worth revisiting the changes to the temple that were intended to kneecap the 'do nothing' strategy, as some of them might not be necessary any more. 


Apart from this it may add an additionally interesting dynamic, whereby you want to participate in a given encounter to potentially get the info that it will provide, but sitting one or two out will leave you in the back of the track, giving you the ability to control where the next encounter will be, which may sometimes be important.  So I think there will be jockeying for position on the time track and jockeying for position on the success tracks, and hopefully these together will create a really nice level of player interaction.

Alleviating tedium

An "influential playtester" had a chance to play the version of the game prior to the changes described in the post before this one.  That playtester reported on enjoying the game but now found the external or "map" phase too long and tedious.  We're awaiting some clarity on those comments but if I had to guess I suspect it's something like this:

In the latest incarnation of the map phase, turns are simple.  You move to a city, reveal an encounter card, read some flavor text about the scene it describes, collect dice (white and red, which are 'good' and 'bad', respectively) and then roll.  'Hits' on white dice move you up on the reward track, 'hits' on red dice move the enemy pawn closer to your city, and when it arrives, further hits start doing damage.  Whenever you want, you end the encounter and take whatever rewards you're entitled to. 

The problem is that although this is simple, it's not always quick:  because you might roll 6 or 7 times before the encounter has resolved, the entire turn could take as much as 3 minutes.  Each player gets seven or eight turns in this phase, so, for four players, that's almost 90 minutes.  And because the encounters are all similar in structure I could see how this could feel tedious after a while.  As this playtester is very influential, we're obligated to take this concern seriously.

There are a few things we could do:

1.  Accelerate the dice:  make the dice stronger so the challenges resolve more quickly.

2.  Accelerate the rewards:  each encounter gives you more stuff (info or equipment), so each player needs fewer turns in total

3.  Accelerate the nature of encounters:  make them more impactful or consequential as the game progresses

4.  Revamp the turn structure completely:  shared turns and simultaneous encounters.

Option 1 is easy to do but it still has the same total number of turns, so it may save some time but won't alleviate the repetitive feel.  Option 2 would require some rebalancing, and one of the nice things about the current system is that luck has time to even out over all those turns; 3 or 4 turns wouldn't be enough for this.  Option 3 might be viable as a way to relieve tedium but I'm not sure it helps with length, plus it probably requires more rules.

I'm therefore somewhat leaning toward option 4.  Instead of individual player turns, players face encounters as a group.  One way that this might work is to give each player a hand of 'city' cards.  When your turn comes up (turns in clockwise order?  Last on time track chooses?) you play one of these cards, which sets the active city for the present turn.  The active player must go there, and other players can either go there or sit the turn out -- but there's some system that caps or effectively caps the total number of turns regardless of whether you sit out or not.

The encounter card that's revealed is faced by ALL players.  Each player rolls their own white dice and moves their marker on the success track accordingly, and then the red dice are rolled for the whole group.

The upside of sitting a turn out is obvious -- it saves you some position on the time track.  But there needs to be some way to make this unfavorable.  And, there should be some way to make jockeying for position in the challenges a bit tense.  One simple thing might be to put some markers on the temple cards.  Each time you receive a card lookup you remove one of its markers, reducing the number of times that card can subsequently be viewed by other players.  So sitting turns out means that the total amount of info available is depleting.  Perhaps also there's an internal timing issue whereby dropping out of a challenge early means you get to receive your reward first, but the longer you stay in the more rewards you're likely to receive.

The real advantage of this system is that you can reduce the total number of encounters to about 8 or 9, which has the other advantage of keeping the game length about the same for all player counts.  You see fewer encounters from game to game so the 'newness' of the flavor text on the encounter deck lasts longer.  Perhaps the nicest thing about this system is that it adds direct player interaction, which the game has pretty much never had in the map phase (except that you could steal a relic in v7).  So I'm inclined to tinker with this a bit and see if it can be made to work. 

Into the temple again

A playtest this week seems to indicate that the latest version of the encounter mechanic outside the temple is working reasonably well.  You flip a card, which describes the scene you face and provides you with a track.  You collect some white dice (good) and red dice (bad), and roll; successes move your marker up the track, bad results move the enemy pawn closer to your city, and when the pawn reaches you, further bad results start to cause damage.  It's straightforward but surprisingly enjoyable and tense -- you can feel like you're safe and then suddenly the enemy closes in on you really quickly!  So it does at least seem to be a way of creating tension and suspense in the encounter phase
.


A more interesting design challenge is shaping up to be in the temple phase.  At present, three are two rows of three cards (representing, in the Grail scenario, the traversal from the start city to the temple, and the temple entrance to the grail room), followed by the grail room itself, and then the final hubris challenge.




Each of those two rows of cards has three cards in it, and I tried something simpler than what the previous post describes:  each card has a single 'peril' on it, of which there are six types.  You must roll four red (bad) dice against the challenge, less one for each symbol you have on an equipment card that matches the card's peril.  And then each 'hit' on the red dice gives a token, and at the end of each card, whoever has the most tokens flips over and resolves a whammy card, and then discards their tokens.




This system turns out to be a bit fragile with respect to two different forms of what I call the "trivial strategy".  The first form is, I forego acquiring information and simply acquire equipment cards, since there's a good chance that any equipment card I pick has a good chance of doing me at least some good.  The second form is, I do almost nothing outside the temple, accept that I'm going to roll poorly in the temple and take damage, and then just try to scoot through the hubris challenge and win by being furthest back on the time track (which increments each time you actually do stuff on a turn outside the temple).




It's hard to address both of these trivial strategies with a single solution.  For example, if we want to make the 'do nothing' strategy a loser, then we could make the whammy cards really bad, such that being totally unprepared means you're really going to get whomped by the temple.  But that doesn't make me want information -- if anything, it pushes you even more into the arms of the 'just get equipment' strategy, because you want as many cards as you can possibly get.  Whereas, any system where I want to get equipment AND info to know what to do with it takes twice as many actions, and so I might feel that just doing nothing is ultimately a better solution.


What makes this even harder to solve is that any modification probably adds steps to the resolution phase, which right now is very quick, and probably adds additional information that you have to record in the map phase as you're getting to look at temple cards.


The next thing I'm planning to try is hopefully an acceptable compromise that also addresses both trivial strategy.  Each slot in the temple will now have a 'peril' card and a 'challenge' card.  You roll 4 'temple dice' on each card, less one for each symbol on your equipment that matches the peril card.  The outcomes on this die are either 'traps' (which cause damage), or 'noise' (which give you 'noise tokens').  On each future card, you also roll red dice, one for each noise token that you hold.  These are the same dice you rolled in encounters and represent the enemy, so it maybe makes thematic sense -- make noise in the temple and it draws the enemy closer to catching you.  The results of the red dice also cause damage.


If you discard an equipment card that matches the 'challenge card' for your current temple slot, you get to cancel the damage you get from traps or from the enemy die.


But!  Prior to revealing the temple card, everyone puts a marker on any one of their equipment cards, and that card's symbols are doubled.  So, this can help you roll fewer temple dice as well as cancel the effects of BOTH traps and enemy dice with a single discard.


But!  You also have a special card with all of the peril symbols and equipment symbols on them, and you can place a marker on any one symbol on that card instead of on one of your equipment cards.  This is useful if, for example, you know the peril but didn't acquire the right equipment for that peril -- at least your info still helps you a bit.


I think that this helps counteract both trivial strategies.  Having equipment and knowledge is strictly better than equipment alone because of the doubling effects.  And both are better than having no equipment, because you're going to take a beating both from the temple dice and the enemy dice.


It might be possible to get a similar effect out a single type of die but I sort of like the feedback loop of noise begetting enemy dice.


The biggest downside seems to be that it's just a bit fiddly for each person to set up for each card.  It would go much faster if each player has their own stash of dice, but including 40 dice in the box doesn't seem very likely.  Maybe dice could be sold separately as a way of accelerating the game, though.  Or maybe there's a way to roll all of the dice together for all players but each player only looks at the number of dice that they're supposed to have rolled.