Caylus Magna Carta Strategy

I’ve taken CMC to the last two game sessions and played. I’ve enjoyed going back and revisiting it (which I don’t think I could do with regular Caylus in a group of new players).

In revisiting the game, I still worry about cube bullying, although I did just fine with a building spam strategy. I got a few castle tokens, mainly via my church, but also got 14 points off of buildings and ~5 gold from my mine. I lost out by a point to the player who built the 14 point blue building. The nice thing is that the strategies seem much more fluid, because of the card flow. It really matters which pink buildings are absent, and which buildings get built early. Despite all my writings about Caylus, I’ve never really gone into CMC strategy.

I don’t think I can get 6,000 words out of it, but a few thoughts:

• Passing and getting der castlegeld is a great deal.
• That being said, there are early buildings that you should consider. In particular, the peddler (buy 1 or 2 cubes for 1/2 dollars, owner can buy one). Getting that out is like a free worker each turn that you don’t have to rush (turn $1 into a cube of your choice whenever someone else places a worker). That’s not as good as one of the wood production buildings early, but it will be great the whole game.
• If you start a turn with no cubes, you better have gotten a great deal for it. It’s a terrible position. Cubes let you threaten to build something and deliver to the castle. If you are cubeless, your opponents should claim production spaces and bully you around. On the other hand, starting a turn with little to no money (and cubes) isn’t horrible. You slap down a worker or two, and can then build buildings or pass and deliver first. (In other words, I’m not convinced that cash is king in CMC)
• Pay attention to what is missing at the start. If the quarry is missing, the game will have a much different flavor because (at the start) only one person can get stone a turn (via peddler). If the quarry and money spot are missing (2-3 player) cash will be incredibly tight.
• Don’t draw cards just because. Particularly if you have a large selection. Dump your hand and replenish if you can.
• Having lots of money isn’t as necessary, except in one situation … jockeying to pass last (to control the provost). Then you can temporize by drawing a card).
• Don’t rush to build the 2nd copy of a wooden production building out (unless you are just going to convert it to a residence). For the stone production buildings, it’s not so bad.
• The Blue buildings are nice, but not overwhelming. The 14 point building costs you 3 gold, 5 stone, and a residence, which is 6 points plus lost income.
• How the non-production buildings come out will change the character of the game. Games can be gold rich or poor (mine/bank), cash rich or poor (depending on church, traders and lawyers). React accordingly. If it’s a cash rich game, get your church out ASAP. If it’s a cash poor game, your trader should be a priority. Etc.
• In a cash rich game, there’s a lot more risk in placing workers near the provost. He moves more.

“Solving” Caylus

There’s a BGG thread about solving Caylus. While I doubt it will be solved anytime soon (Chess shows no signs of giving up its secrets), I do think a reasonable computer opponent could be developed.

My own technique would be to use some basic genetic algorithms for some decisions, and then use some recent advances in computer Go as a jumping off point.

For example, during worker placement you’ll typically have 15 options (Pass, Castle, Gate, Guild, Joust, Stables, Inn, Pass, Spaces 1-n). You mainly can’t take occupied spaces, and you can eliminate some obviously bad moves (spaces you won’t be able to use, or lose money). Now, if your genetic algorithm (or hardcoded rules) point to a clear decision … take it. You can also have a clear evaluation function (money, favor, goods are positive … wasted workers negative, etc etc).

But if you’ve got 2-3 candidate moves, consider each one. Simulate the turn out, then play out rest of the game 100 (or 200, or 1000) times using random moves for all players. (Possibly keeping the smarts that eliminate completely boneheaded moves). Whichever candidate gives you the best average outcome, take it.

Given how well this works for Go, I think it would be generally applicable.

You’d want to avoid randomly moving the provost (that would probably be hardcoded, and possibly genetic).

My Computer Science theory isn’t quite strong enough for me to set up this framework myself (nor do I feel like spending the time) but if a project got started (say, on SourceForge) I may contribute.

Who can take a rainbow … wrap it in a sigh?

Caylus can.
Now that I’m not so busy playing Caylus, I find more time to kibitz games. Things I’ve learned:

• Strong players have definite styles, and often disagree about moves. This is more likely in a 4-5 player game, though, where you can take the safest move, or take a better move that risks you position and hope that your best move waits a turn. [This often happens on a turn there should be a rush on the castle to finish the 2nd section. If everyone waits, you are golden. If the rush starts without you …]
• Don’t take the riskier play when you are clearly winning.
• You can go half the game without a favor. If you set things up correctly, you are behind, but it’s not insurmountable.
• You can also win without scoring in the first section (no buildings, no walls).
• Perfect storms happen. I won a lost game when the leader misjudged the position and waited one round too long to go to the castle. I just witnessed a won game lost when one player grabbed enough cubes to win the castle, but then a pass’s chain reaction left him short of cash.
• Noticing a theme? — “Hey, have you considered that you may get blocked out of the castle?”
• Going last is it’s own reward. I’m getting comfortable with placing workers never intending to use the tiles. (If I can get a cube out of it, though, great!). Just to go last. A player who goes first and grabs the merchant’s quarter often can’t back him up, because you may continue the journey backwards.
• Mo’ Money. Seriously, any turn with the exception of the final turn, the money may be a lifesaver. I’ve only seen one game where anyone took a money favor and I thought “That was a waste”. [It was a 2 player game, and that pushed him over $20 … with an income of $4 or $5].
• I don’t remember anyone starting a turn with $2 and winning without gross negligence on the opponent’s part. In 2er, $4 or less is a problem. (It may be ok with more players, because the spaces will fill up faster).
• That being said, know when you can spend everything during a turn. (To win the castle. Before a sectional scoring where you have a favor or three).
• In 2er, $3 and the merchants quarter are often equivalent.
• I’m fond of building the market or the peddler as my first building (depending on if I’m winning or losing the money situation). The 2 extra VP are nice, and I want to exploit the monetary advantage. I’ll only build the lawyer first if I have a leftover clothe (which happens). If the lawyer appears early, the resource track become viable. Only then.
• Sometimes you take a resource cube during the sectional scoring, but just to build a grey building.
• If you aren’t sure whose winning, see who leads on stone production buildings. They’re that good. However, once you’ve passed the mine, consider taking the extra 3 VP. And don’t forget the church! That favor is usually 3VP, at least. [If you build the church via the mason, you can then build again with the favor.]

Post Groundhog Gaming

Last night we had guests over for brownies and gaming. Lot’s of blasts from the pasts. Power Grid, Puerto Rico, Fearsome Floors and Ricochet Robots.

That means I’ve played Puerto Rico fifty times (face to face).

I’m also played Ticket to Ride with my daughter today. She wins about 1/3rd of the time using the monomaniacal powers of youth, by drawing lots of cards (mystery, especially) playing long routes, and connecting her starting tickets, no more. It works fairly well.

Online, Caylus rules with an iron fist. I’m definitely starting to respect the “VP + Money track” and just give up on blue buildings unless it’s easy. One thing I do know is that if one player builds the mason/lawyer/architect, the players who ignore the blue track and don’t build said buildings can run away with the game. Remember, any sufficiently complex system invites parasites. Damn VP track running parasites. There are several players online, the Jedi of Caylus, and I’m not sure if they’re just using the strategy guide better (fewer mistakes, etc) or if they see something I don’t. I’ve started watching quite a few games. Perhaps I’ll figure it out.

More Caylus Strategy

I’ve uploaded v1.2 of my strategy file, and added in Tom’s concept of Transformational Efficiency, a more detailed look at prestige buildings, and what not. (I actually wrote a bit over the weekend, so I’m not entirely sure what’s new). I also added links to various BGG threads of note.

Now that I’ve played 25 games (finished #25 an hour ago), I may start playing some other games on BSW again.

Update: I could have titled this “Almost 10,000 words about Caylus”.

Winnowing the Minnows

I see more sharks swimming on BSW. I’ve had several close games (wins and losses) and a blowout loss where the winner ignored the building track (reassuring). The basic formula appears to be:

1. Build the Mason & Lawyer
2. Build the architect
3. Run the VP track
4. ????

The first two are to make sure you don’t get shut out of a prestige building. The third is to build up VPs, obviously. That 4th point is “do something clever,” but I’m not sure it stays the same.

But a refreshing loss, nonetheless.

Caylus replies

There’s a provocative post on BGG about Caylus pacing that includes the following: “If you elect the castle building strategy, you prefer a slower game.“

I don’t know if I agree (it certainly flips conventional wisdom, but everyone has played so few games.) I can see several arguements in its favor.

Tom mentions the following points (paraphrased):

1. Mathematically, a game is 12 turns average (not 13-14), and will likely be fewer.
2. I need to stress transformational efficiency.
3. I need to discuss prestige buildings in more detail.

Taking the last point first, “I kaenna change the laws of typing!” But I’ll probably add onto it, later.

Mathematically, yes, I messed up. On the other hand. In my games (against different opponents, so a reasonable sample for anecdotes) I’ve seen one game where the provost moved one every turn except ~2 or 3, I’ve seen one where it moved two almost every turn. Groupthink.

But most games the provost seems to take one step for the first few turns.

As for transformational efficiency, that deserves more detail. Tom says:

Over the course of a game (depending on length and groupthink), some 30-40 $ (and a few starting cubes) turn into 60-100 VPs. That means you want to be looking for good $->cube->VP transformations. Anything with multipliers of more than 2x along the way is quite nice.

The neutral buildings present us with mostly 1:1 transformations. For a $, I get a pink cube here, a brown cube there, and a wooden building (2 or 4 VPs) there. 3 $ in; and, on average, 3 VPs out.

This runs counter to the building as VP investments and opportunity cost ways of thinking about Caylus. Yes, my Trattoria yields me 1 VP if you use it; but you get 2 purple cubes instead of 1 from a neutral building. This is mostly a wash (just both of us climbing from 1x to 2x transformations) and, if everyone is using each other’s buildings to the same extent (which if you’re all competent in choosing decent buildings to buy, should occur), a complete wash.

A few rebuttal points:

• If we assume I’m building more often, it isn’t a complete wash.
• The 1:1 only matters if you spend $1 per action. This is probably true for most actions, but not all. Buildings provide some flexibility.

But I’ll concede this are both minor points. The major points are that the extra VP you get from building no longer cost you an action. We’re costing actions at $1, but that discounts the timing impact. My opponents get their ‘extra’ VP by taking their action. I get it in addition to my action, and can presumably get an extra VP somewhere, too.

In any case, I don’t think a pure strategy (build and ignore the castle) would work. It may be that building is worse than the castle, but if I dont’ go first, the option isn’t to win the castle on turn one. It’s to build the first building or hoard (cubes or money). You take what you can.

By the strict transformational efficiency (I like that phrase), building 1 section at the castle costs $4 and earns 3-5 VP (1:1) plus a favor, so it’s a winner. [And it earns 1/2 a favor, approximately, for when you score. And it avoids the VP penalties].

Jousting can certainly compete well with that ($3 for 5 VP if you max the favor track, or $4 for 6 VP via a stone building).

However, building Residences early (so that they have time to return) as $ investments can be a viable strategy. Pumping your income stream from $2 to $5 quickly by building Residences on turns 3-5 introduces a 2.5x multiplier right off the bat…

Yes, that should be tough to beat. (I’ve won down 3 residences, but it didn’t occur that early. A residence costs $3 for $1/turn. $2 for 6 now (via the wooden marketplace) is worse, but not horribly so.

But I think that, in general, I’m leaning towards the castle myself. Transformational efficiency does clarify quite a bit. It means I should strongly consider hoarding on turn one, and letting others rush the buildings.

Snatching the pebble from Caylus

Ok. I am now going to enter a Caylus hiatus. Despite being a chanceless game, I’ve had good luck with this. And I mean that literally. I just won a game (by one point) because the ‘perfect storm’ hit on the last turn. The distant 3rd opponent managed to get enough cubes to build the huge 25 point prestige building, preventing the ‘close second’ player from building it. [I had actually thought that #3 couldn’t do it, missing that he’d upped the resource track once and could use two favors to a) get a stone, b) build out].

Even with all of that, it was 105-104-101. I’ve won another game when I lost the thread (got cash poor) and then everyone else seemed to have an attack of niceness. In fact, I suspect know that many of my wins are due to opponent’s mistakes, and not any particular brilliance on my part.

After 16 games, I firmly believe:

• The construction favor is most powerful. My default plan is “Construction early, catch up VPs in midgame. Take money or resources only when desperate.”
• Build a minimum at the castle to get the turns favor. Exception — on the scoring turn, building to get to a favor breakpoint.
• You need to know where your money will be at the start of the next turn.
• Getting lots of residences is not an overwhelming advantage, assuming that every player gets one. [I’ve frequently fallen behind by two or more.]
• Proper play varies with the number of players. It’s even quite dramatic. [I’ve finally played several two player games. I think I prefer 3/4, but 2/5 are interesting, too.]

I believe somewhat strongly —

• In a 5 player game (with competent players), the last player is at a significant disadvantage. This might be true in a four player game, but I’m not as confident.
• People should pass early more often. [In a 2 player game, this is definitely true].
• The player who passes last (on the bridge) should spend $1 to make the bailiff move two much more often than I’m seeing. [Apart from myself, I’m barely seeing this].

Caylus still dominates my gaming thoughts, despite the brisk pace. But I’m taking off a night or two to play real games.

2,500 more words about Caylus, and “Strategy”

Rather than post it (which is a pain), I’ve just started adding to a document. New sections on timing, more about the opening/midgame/endgame, number of players, ‘building chains’, flexibility, resources. And other thoughts that I’ve had.

Is Caylus Strategic? More so than PR. There are “positional” affects (things that aren’t immediately tactical, but that affect play). Larry brought up an example of “If I’m doing the shipping strategy, I’m going to want a harbor, or If I’m doing the factory, it will change my builds.” Well, that’s certainly true, and you can see “I want this, then this, then this.” But I’m not sure it’s positional. You could model the same effects as ‘evaluation of future values’. If I have the factory, it’s better if I have lots of good types. Ditto Harbor, but the Wharf wants specialization. My current position makes those better (or worse), and I don’t have to look ahead to see it. If I get them, they make future plays better (or worse). Also, other players can’t really affect this (except by buying buildings before I can). These buildings have effects in the future, but they local to me. There are second order effects, to be sure. If I know that other players are building or shipping, I can expect them to pick the appropriate roles more often.

On the other hand, if everyone builds resource buildings (or doesn’t) then the Caylus landscape changes for everyone. I think it’s strategic (over my last several games, I’ve thought about ‘board position’ issues, and these form quite a bit of my new verbiage). I still consider PR mainly evaluational + tactical. I think about the future in both games, but my thoughts are more definite, and more complex in Caylus. In PR, I just come up with a plan (I’ll buy A, intending B, C & D) and then look at the tactics.

I think this may be due to the competition for buildings (actions), unlike PR where the roles (actions) are shared by all player. But I’m not sure.

2,500 more words about Caylus, and “Strategy”

Rather than post it (which is a pain), I’ve just started adding to a document. New sections on timing, more about the opening/midgame/endgame, number of players, ‘building chains’, flexibility, resources. And other thoughts that I’ve had.

Is Caylus Strategic? More so than PR. There are “positional” affects (things that aren’t immediately tactical, but that affect play). Larry brought up an example of “If I’m doing the shipping strategy, I’m going to want a harbor, or If I’m doing the factory, it will change my builds.” Well, that’s certainly true, and you can see “I want this, then this, then this.” But I’m not sure it’s positional. You could model the same effects as ‘evaluation of future values’. If I have the factory, it’s better if I have lots of good types. Ditto Harbor, but the Wharf wants specialization. My current position makes those better (or worse), and I don’t have to look ahead to see it. If I get them, they make future plays better (or worse). Also, other players can’t really affect this (except by buying buildings before I can). These buildings have effects in the future, but they local to me. There are second order effects, to be sure. If I know that other players are building or shipping, I can expect them to pick the appropriate roles more often.

On the other hand, if everyone builds resource buildings (or doesn’t) then the Caylus landscape changes for everyone. I think it’s strategic (over my last several games, I’ve thought about ‘board position’ issues, and these form quite a bit of my new verbiage). I still consider PR mainly evaluational + tactical. I think about the future in both games, but my thoughts are more definite, and more complex in Caylus. In PR, I just come up with a plan (I’ll buy A, intending B, C & D) and then look at the tactics.

I think this may be due to the competition for buildings (actions), unlike PR where the roles (actions) are shared by all player. But I’m not sure.