A lot of card games this week (Nyet, No Thanks, Mamma Mia, Frank’s Zoo), which… is entirely fine by me. It even convinced me to finally pick up a copy of Mü. I have also finally reckoned with letting go of my copy of San Juan - happy enough to play it if someone else is in the mood, but not so happy that I’ll suggest it myself.
Collection Played in 2016: 32%. 98 games to go. Added this week: Mü (from Mü & Lots More), Aton. Trade Pile: Mr. Jack Pocket, San Juan.
Pandemic - A few 2p games earlier in the week, when we decided to go after it legacy style - adding expansions back in, one by one. The Contingency Planner is just mediocre with only the basic five events in the deck.
La Citta - This held up nicely to a second play, and it was a tight race for first place - there are real decisions to make, and you don’t always have all the information you would like. It’s a good feeling.
Nyet(new to me) - The composition of the deck has changed in every version of this game that has ever been published, presumably to accommodate 3/5 players. For unknown reasons, the bidding board has also changed (there’s no notrump option in the new printing).
No Thanks - I don’t win No Thanks often, but I had a good feeling when I was struggling to hold all the chips in my hand.
Mamma Mia - A new record for me - four completed recipes in round one. And only one completed recipe in rounds 2-3, but not for lack of trying. Of course I lost.
Kitty Paw(new to me) - Speed games are fine, but this feels off… point spreads each round vary from 5-0-0-(-1) to 0-0-(-1)-(-1). If there was more room in the box, I’d wager it would include a pile of 1st/2nd/3rd tokens to place in the middle of the table.
Frank's Zoo - Without partnership rules, with the hedgehog/lion scoring. I’m surprised this isn’t an official variant, it works great.
Automania(new to me) - I may be crazy - it feels like this game started with Traumfabrik. I can imagine a Hollywood retheme, with studios instead of assembly lines.
Mission: Red Planet(new to me) - If this keeps up, I may have to admit that I enjoy majority games. This was quite good.
Aton(new to me) - Good initial impression, I appreciate the variety in victory conditions (two possible instant wins, and one victory on points).
Couldn’t find enough card gamers for a game of Mü this week, but I did manage to convince a couple of eurogamers to play Frank’s Zoo at Friday game night. They may have even enjoyed it!
Collection Played in 2016: 33%. 97 games to go. Added this week: Witness.
Scoville - One of Anthony’s games I’m usually happy to play. The victory came down to who could plant a phantom pepper first on the final turn - and it was not me.
Frank's Zoo - In the first two games I didn’t note that there are exactly five cards of each animal in the deck, which makes counting important cards (‘important’ could be any number of things depending on your hand) much easier.
Witness(new to me) - Technically Anthony’s game, but one I’m probably going to suggest pretty often. I read a lot of Clue mystery books in elementary school, and Witness taps into the same kind of enjoyment (with the extra wrinkle of having important information degrade on its way around the table).
Russian Railroads - I pushed hard on Vladivostok, and ended with a toe in engineers and the industry track. First and second place were separated by five points - and second and third by nearly 100 - in a game with a winning score of 480.
Rome: City of Marble - I made it my goal to construct at least one colosseum, at great expense to myself… which actually paid off.
Lords of Vegas(new to me) - Reminds me (and only me) of Fast Food Franchise, given the expansion mechanism, hostile takeovers, and game length. I only wish the casinos were named Dark & Gooey, Family Style, etc.
504 - Game 716 (Majorities, Pick up & Deliver, Connections). A better game than our first, although the five-turn game clock meant upgrading the delivery cart was a big mistake. (So was starting in the corner of the map)
Witness(new to me) - Technically Anthony’s game, but one I’m probably going to suggest pretty often. I read a lot of Clue mystery books in elementary school, and Witness taps into the same kind of enjoyment (with the extra wrinkle of having important information degrade on its way around the table).
Lords of Vegas(new to me) - Reminds me (and only me) of Fast Food Franchise, given the expansion mechanism, hostile takeovers, and game length. I only wish the casinos were named Dark & Gooey, Family Style, etc.
504 - Game 716 (Majorities, Pick up & Deliver, Connections). A better game than our first, although the five-turn game clock meant upgrading the delivery cart was a big mistake. (So was starting in the corner of the map)
Witness sounds interesting. Even though I somehow came in second Lords of Vegas seemed way too random for my taste. As for 504, upgrading the delivery cart's speed was probably better than it's capacity. The Connections aspect played a much bigger part in seizing terrain. Although it really helped that the game ended as quickly as it did.
This paper studies a cooperative card game called Hanabi from an algorithmic combinatorial game theory viewpoint. The aim of the game is to play cards from 1 to n in increasing order (this has to be done independently in c different colors). Cards are drawn from a deck one by one. Drawn cards are either immediately played, discarded or stored for future use (overall each player can store up to h cards).
We introduce a simplified mathematical model of a single-player version of the game, and show several complexity results: the game is intractable in a general setting, but becomes tractable (and even linear) in some interesting restricted cases (i.e., for small values of h and c).
This paper studies a cooperative card game called Hanabi from an algorithmic combinatorial game theory viewpoint. The aim of the game is to play cards from 1 to n in increasing order (this has to be done independently in c different colors). Cards are drawn from a deck one by one. Drawn cards are either immediately played, discarded or stored for future use (overall each player can store up to h cards).
We introduce a simplified mathematical model of a single-player version of the game, and show several complexity results: the game is intractable in a general setting, but becomes tractable (and even linear) in some interesting restricted cases (i.e., for small values of h and c).
Reduction from 3-SAT.I always assumed if you can see everyone's hand, it should be pretty easy to always get a perfect score using this algorithm:
On your turn choose the first valid option in this list:
1) Play valid card if possible. If given a choice, try to work towards playing 5s earlier. 2) If it does not exceed the number of discards that will result in a non-perfect game, discard a card that will not ruin a perfect game if discarded. 3) Use a hint token. Sure, we are cheating, but those are still a part of the game. 4) Play an invalid card and expend one of the two fuses that won't end the game.
If that algorithm really won't beat the game when cheating that's incredible. I saved the PDF for reading later.
I always assumed if you can see everyone's hand, it should be pretty easy to always get a perfect score using this algorithm:
On your turn choose the first valid option in this list:
1) Play valid card if possible. If given a choice, try to work towards playing 5s earlier. 2) If it does not exceed the number of discards that will result in a non-perfect game, discard a card that will not ruin a perfect game if discarded. 3) Use a hint token. Sure, we are cheating, but those are still a part of the game. 4) Play an invalid card and expend one of the two fuses that won't end the game.
If that algorithm really won't beat the game when cheating that's incredible. I saved the PDF for reading later.
It's definitely not guaranteed; it's quite possible that you will fail simply due to bad luck with regards to the order in which the cards come out. For example, all of the ones might be at the very end of the deck.
As for the paper, it's interesting but the subject of their paper isn't Hanabi per se; their NP-complete single-player version assumes you know the order of the entire deck. Although in general games are harder with uncertainty, it may be that optimal Hanabi play is algorithmically easier if you don't know what's coming than if you do.
The reality is that there is a lot of advanced signaling among skilled players. Not pre-arranged signals or outright cheating, but familar players assuming the logic other players will use, and then tuning their own play to fit that logic and thus predict their behavior.
the game is intractable in a general setting, but becomes tractable (and even linear) in some interesting restricted cases (i.e., for small values of h and c).
So if you have to play cards 1-100 million, intractable. But if you only have to play up to 5, possible.
Explore a lot because you want discovery tiles, attack ancients early (but not too early), and play as a human in your first game.
So, I placed third in a five player game. I didn't get an opportunity to read the rules due to short notice and our teacher admittedly stated he'd only played once before and hadn't read the rules himself. Grrrrrrrr. Turned out I knew as much or more about the rules just based on overhearing conversations of people who had played. So the game was looooong as I was turned into the rules go to guy for no reason other than I was willing to look up rules and no one else was.
As to the actual game, my early explores revealed nothing but material worlds and a discovery tile that got me neutron bombs so I had both of my dreadnaughts constructed on turn 3. Next door on one flank was the alien race that can influence ancient sectors (among other things). Turn 4 I acquired the two damage gun and equipped it on the aforementioned dreadnaughts and rolled them into an ancient sector occupied by those aliens.
He instantly poops out all 4 of his star bases in that hex, at the expense of going to zero money the next turn to do it and I'm forced to retreat. On turn 5, I attempt another invasion to my other neighbor and the exact thing happens except she turns out 4 cruisers that have targeting computers up the ying yang.
The other players opposite them each capitalized on their all ins and ate up their backsides. I chill for two turns making minor improvements to my dreadnaughts and building a small escort fleet of interceptors, languishing in my pathetic 4 science a turn due to shitty explorations for me.
Turn 8 I attempt to push on the ancients hex again and he again goes all in on defense (with another player actively already eating up his back side) and the result is the same. I lose no vessels but he overcommits opening the back door further for the other players and I go nowhere.
Turn 9 I'm finally able to crack into my other neighbors systems taking two of them and making her rage quit but it's not enough to do anything more than take third since the two empires what got feed easily outstripped me.
Tl;dr - I did not win the vote who wins game.
Don't get me wrong, I enjoy the occasional 4X game and even in a spite of all the rules ignorance I had to deal with, I think eclipse is a good one that I'd play again, however it's games like this that just pinch my exposed nerve to king making in games. I like games that are true tests of skill where there is minimal King making as an ideal.
My first couple games were a shit-show. I'm at ~100 plays though at this point. The kingmaking is one of those always possible things that should be mitigated a bit with some experience. That said, it still happens. I'd love to play a slightly shorter eclipse... like I think 6 rounds is good... but that's also possibly related to how much I focus on "rushing" in this game. Like, turn 3 first alien tile is what I consider late. Like the latest it should ever happen unless you make a mistake. Round 1 is pretty possible with some board configurations, technologies, and races... but you have to actually know all those different count by numbers things. "Oh, Orion are open, improved hull is out, rush tier 1 hope for aliens (even double), 2 cruisers up their ass." It's a slight gamble, but it works. Same for rho indi (expansion race) with the rare tech computer/hull part. Having a good idea about what the different opening games are (especially considering discovery tile possibilities and hex draws) helps. The expansion turns this up to 11 with some of the development tiles.
I just wish the IOS app was actually fully implemented properly and had the expansion.
This Saturday was one of the better game days I’ve had in months - seven different games in seven hours. Usually, gaming at that pace only happens at conventions.
Collection Played in 2016: 36%. 93 games to go. Added this week: Tally Ho!
Tally Ho!(new to me) - Tally Ho is ahead of its time (1973) - although Anthony isn’t really into the theme of chasing foxes around a forest with guns. Still a very clever game, an abstract where a tile flip can change everything about the state of the board in an exciting way.
Swinging Jivecat Voodoo Lounge(new to me) - I classify this alongside Mage Knight and Lewis & Clark - games where I can’t plan my entire turn in my head. I enjoy both of those games, and this one is pretty good too, albeit a bit more abstract.
Keltis: Der Weg der Steine(new to me) - One of the games I threw into an Amazon.de order at the end of 2015. Perfectly fine for a 20 minute game, and a likely candidate for lunchtime games at work.
Acquire - Maybe someday I’ll get my hands on a midcentury copy of Acquire - but until then, the cheapo Hasbro edition from 2008 actually works fine.
Eminent Domain: Battlecruisers(new to me) - The first game of ‘don’t play the same card as everyone else’ this week. Almost works for me, but a player’s position doesn’t give you enough information to narrow down the possibilities for which card they will play.
Happy Pigs(new to me) - The second game of ‘don’t play the same card as everyone else,’ this time in service of an economic game.
Mü(new to me) - I’ve played this trick-taking game dozens of times on iOS, but this is the first time we played in person. It was much easier than I expected to explain and play, and holding an actual hand of cards led to some insights that multiple games against AI did not.
Witness - None of the easy cases have been quite as interesting as the first, but more play has convinced me one dull case was the exception, not the rule.
Starship Merchants - The prices of the ships seem quite balanced - you would think it’s a mistake to pay $50 for a permanent ship when $100 is a winning score, but that single action can rust entire fleets of your opponents’ ships, then pay for itself in one trip around the board.
I played Swinging Jivecat Voodoo Lounge in prototype form over 4 years ago. Damn that was a long time ago. I remember really liking it, but don't remember aaaaanything about it. Glad to hear it held up in final form.
I played some Tammany Hall over the weekend - didn't actually finish out the game because I had to catch a train, but I had fun. It's an area-control game with some nice catch-up mechanisms that prevent one player from running away with the game, and it's a theme I quite enjoy.
The lunch board game club has been playing Mafia de Cuba the last couple of days. It's pretty good; like a Werewolf-type game with a bit of information to help it from degenerating into "Everyone vote to kill Rym because he's a filthy liar".
Comments
Collection Played in 2016: 32%. 98 games to go. Added this week: Mü (from Mü & Lots More), Aton. Trade Pile: Mr. Jack Pocket, San Juan.
Pandemic - A few 2p games earlier in the week, when we decided to go after it legacy style - adding expansions back in, one by one. The Contingency Planner is just mediocre with only the basic five events in the deck.
La Citta - This held up nicely to a second play, and it was a tight race for first place - there are real decisions to make, and you don’t always have all the information you would like. It’s a good feeling.
Nyet (new to me) - The composition of the deck has changed in every version of this game that has ever been published, presumably to accommodate 3/5 players. For unknown reasons, the bidding board has also changed (there’s no notrump option in the new printing).
No Thanks - I don’t win No Thanks often, but I had a good feeling when I was struggling to hold all the chips in my hand.
Mamma Mia - A new record for me - four completed recipes in round one. And only one completed recipe in rounds 2-3, but not for lack of trying. Of course I lost.
Kitty Paw (new to me) - Speed games are fine, but this feels off… point spreads each round vary from 5-0-0-(-1) to 0-0-(-1)-(-1). If there was more room in the box, I’d wager it would include a pile of 1st/2nd/3rd tokens to place in the middle of the table.
Frank's Zoo - Without partnership rules, with the hedgehog/lion scoring. I’m surprised this isn’t an official variant, it works great.
Automania (new to me) - I may be crazy - it feels like this game started with Traumfabrik. I can imagine a Hollywood retheme, with studios instead of assembly lines.
Mission: Red Planet (new to me) - If this keeps up, I may have to admit that I enjoy majority games. This was quite good.
Aton (new to me) - Good initial impression, I appreciate the variety in victory conditions (two possible instant wins, and one victory on points).
http://www.polygon.com/2016/3/16/11249004/bloodborne-tabletop-card-game
Collection Played in 2016: 33%. 97 games to go. Added this week: Witness.
Scoville - One of Anthony’s games I’m usually happy to play. The victory came down to who could plant a phantom pepper first on the final turn - and it was not me.
Frank's Zoo - In the first two games I didn’t note that there are exactly five cards of each animal in the deck, which makes counting important cards (‘important’ could be any number of things depending on your hand) much easier.
Witness (new to me) - Technically Anthony’s game, but one I’m probably going to suggest pretty often. I read a lot of Clue mystery books in elementary school, and Witness taps into the same kind of enjoyment (with the extra wrinkle of having important information degrade on its way around the table).
Russian Railroads - I pushed hard on Vladivostok, and ended with a toe in engineers and the industry track. First and second place were separated by five points - and second and third by nearly 100 - in a game with a winning score of 480.
Rome: City of Marble - I made it my goal to construct at least one colosseum, at great expense to myself… which actually paid off.
Lords of Vegas (new to me) - Reminds me (and only me) of Fast Food Franchise, given the expansion mechanism, hostile takeovers, and game length. I only wish the casinos were named Dark & Gooey, Family Style, etc.
504 - Game 716 (Majorities, Pick up & Deliver, Connections). A better game than our first, although the five-turn game clock meant upgrading the delivery cart was a big mistake. (So was starting in the corner of the map)
I'll spare you the plot of the show but naturally curious on how they are going to pull off Mouse Guard.
On your turn choose the first valid option in this list:
1) Play valid card if possible. If given a choice, try to work towards playing 5s earlier.
2) If it does not exceed the number of discards that will result in a non-perfect game, discard a card that will not ruin a perfect game if discarded.
3) Use a hint token. Sure, we are cheating, but those are still a part of the game.
4) Play an invalid card and expend one of the two fuses that won't end the game.
If that algorithm really won't beat the game when cheating that's incredible. I saved the PDF for reading later.
As for the paper, it's interesting but the subject of their paper isn't Hanabi per se; their NP-complete single-player version assumes you know the order of the entire deck. Although in general games are harder with uncertainty, it may be that optimal Hanabi play is algorithmically easier if you don't know what's coming than if you do.
I didn't read the paper, though.
So, I placed third in a five player game. I didn't get an opportunity to read the rules due to short notice and our teacher admittedly stated he'd only played once before and hadn't read the rules himself. Grrrrrrrr. Turned out I knew as much or more about the rules just based on overhearing conversations of people who had played. So the game was looooong as I was turned into the rules go to guy for no reason other than I was willing to look up rules and no one else was.
As to the actual game, my early explores revealed nothing but material worlds and a discovery tile that got me neutron bombs so I had both of my dreadnaughts constructed on turn 3. Next door on one flank was the alien race that can influence ancient sectors (among other things). Turn 4 I acquired the two damage gun and equipped it on the aforementioned dreadnaughts and rolled them into an ancient sector occupied by those aliens.
He instantly poops out all 4 of his star bases in that hex, at the expense of going to zero money the next turn to do it and I'm forced to retreat. On turn 5, I attempt another invasion to my other neighbor and the exact thing happens except she turns out 4 cruisers that have targeting computers up the ying yang.
The other players opposite them each capitalized on their all ins and ate up their backsides. I chill for two turns making minor improvements to my dreadnaughts and building a small escort fleet of interceptors, languishing in my pathetic 4 science a turn due to shitty explorations for me.
Turn 8 I attempt to push on the ancients hex again and he again goes all in on defense (with another player actively already eating up his back side) and the result is the same. I lose no vessels but he overcommits opening the back door further for the other players and I go nowhere.
Turn 9 I'm finally able to crack into my other neighbors systems taking two of them and making her rage quit but it's not enough to do anything more than take third since the two empires what got feed easily outstripped me.
Tl;dr - I did not win the vote who wins game.
Don't get me wrong, I enjoy the occasional 4X game and even in a spite of all the rules ignorance I had to deal with, I think eclipse is a good one that I'd play again, however it's games like this that just pinch my exposed nerve to king making in games. I like games that are true tests of skill where there is minimal King making as an ideal.
I just wish the IOS app was actually fully implemented properly and had the expansion.
Collection Played in 2016: 36%. 93 games to go. Added this week: Tally Ho!
Tally Ho! (new to me) - Tally Ho is ahead of its time (1973) - although Anthony isn’t really into the theme of chasing foxes around a forest with guns. Still a very clever game, an abstract where a tile flip can change everything about the state of the board in an exciting way.
Swinging Jivecat Voodoo Lounge (new to me) - I classify this alongside Mage Knight and Lewis & Clark - games where I can’t plan my entire turn in my head. I enjoy both of those games, and this one is pretty good too, albeit a bit more abstract.
Keltis: Der Weg der Steine (new to me) - One of the games I threw into an Amazon.de order at the end of 2015. Perfectly fine for a 20 minute game, and a likely candidate for lunchtime games at work.
Acquire - Maybe someday I’ll get my hands on a midcentury copy of Acquire - but until then, the cheapo Hasbro edition from 2008 actually works fine.
Eminent Domain: Battlecruisers (new to me) - The first game of ‘don’t play the same card as everyone else’ this week. Almost works for me, but a player’s position doesn’t give you enough information to narrow down the possibilities for which card they will play.
Happy Pigs (new to me) - The second game of ‘don’t play the same card as everyone else,’ this time in service of an economic game.
Mü (new to me) - I’ve played this trick-taking game dozens of times on iOS, but this is the first time we played in person. It was much easier than I expected to explain and play, and holding an actual hand of cards led to some insights that multiple games against AI did not.
Witness - None of the easy cases have been quite as interesting as the first, but more play has convinced me one dull case was the exception, not the rule.
Starship Merchants - The prices of the ships seem quite balanced - you would think it’s a mistake to pay $50 for a permanent ship when $100 is a winning score, but that single action can rust entire fleets of your opponents’ ships, then pay for itself in one trip around the board.