Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[theory]

in this case chess isn't even a bad example. a computer can play chess. doesn't this mean you consider playing chess an insult of your intelligence? what if there was an AI that played dominion really well?

That's a good point and my statement was definitely sprayed with a flair of hyperbole, but the fact that computers play games like Chess and Backgammon better than humans certainly has caused those games to lose a lot of their charm to me. If there was an AI that played Dominion really well I'd play it all day at first, expecting to learn a ton. Whether I'd get dissatisfied in the long run is something I can't tell without actually going through the process. I don't expect that to happen any time soon, though, because while it's straightforward to write a bot that plays a not too complex given kingdom fairly well, teaching it to play a random kingdom is more like teaching it to play Chess960.

I feel the same way. I was really into chess in middle school but not so much anymore and I don't have a strong desire to get back into it because I know I'll never be as good as a computer program.

By similar reasoning, I don't bother doing crossword puzzles, playing Sudoku, or even working out.

[AdamH]

By similar reasoning, I don't bother doing crossword puzzles, playing Sudoku, or even working out.

I know you're being sarcastic, but when I did the Project Euler problem that involved writing a computer program to solve sudokus, I no longer had fun doing them myself.

[Awaclus]

By similar reasoning, I don't bother doing crossword puzzles, playing Sudoku, or even working out.

I know you're being sarcastic, but when I did the Project Euler problem that involved writing a computer program to solve sudokus, I no longer had fun doing them myself.

You have exactly 1337 respect!

[mpsprs]

By similar reasoning, I don't bother doing crossword puzzles, playing Sudoku, or even working out.

I know you're being sarcastic, but when I did the Project Euler problem that involved writing a computer program to solve sudokus, I no longer had fun doing them myself.

I had a similar experience (though it's probably more accurate to say that I no longer had as much fun doing  sudokus when I realized that I could write such a program easily).  I have no such illusions about my ability to program a crossword solver or a dominion AI, so crosswords and dominion retain their 'fun' value, despite the existence of programs that perform these tasks. 

As far as the salvager feature, I think my (weak) preference would be along the lines of SCSN's thoughts.  When I play, I often play in a distracted setting, and like being able to (attempt to) make good decisions with the information I could have if I were fully focused.  I'll admit to suspecting that, if I played IRL more often (at all?) I'd be much more sympathetic with AdamH's view.

And finally, when I do want to solve a sudoku, I solve these: http://www.amazon.com/Sudoku-Masterpieces-Elegant-Challenges-Lovers/dp/1402771924.  The puzzles are excellent and have different (and unusual) topologies which hold my interest.

[soulnet]

I know you're being sarcastic, but when I did the Project Euler problem that involved writing a computer program to solve sudokus, I no longer had fun doing them myself.

Many of the online judges that pose a "solve Sudoku" problem do not have strong enough test cases and are thus simple to get accepted with the straightforward 10-line backtracking code. Solving a set of strong test cases is (surprisingly?) harder than that (me and my team did it on our university's internal selection, and got "time limit" on the first try for doing the straightforward thing that worked in lots of other places).

To get more on topic, I don't think a computer program will play better Dominion than humans just because it has access to a better tracking and probability calculation. After all, Poker is a much simpler and less variable game than Dominion and machines are still not clearly better than the best humans at it (AFAIK) despite having more than 10 years of research put on them. Of course, Dominion being less popular than poker (only god knows why) makes it even less likely that a computer will be more proficient than MicQ or Stef at Dominion any time soon, even if it were reasonably possible.

[Awaclus]

Dominion being less popular than poker (only god knows why)

Everyone already has a Poker deck, and anyone who doesn't, can buy one for a relatively small price, while Dominion costs hundreds of dollars if you buy all of the expansions. That might have something to do with it.

[silverspawn]

poker is way simpler and has a higher skill gap anyway. I don't find it surprising at all.

[liopoil]

in this case chess isn't even a bad example. a computer can play chess. doesn't this mean you consider playing chess an insult of your intelligence? what if there was an AI that played dominion really well?

That's a good point and my statement was definitely sprayed with a flair of hyperbole, but the fact that computers play games like Chess and Backgammon better than humans certainly has caused those games to lose a lot of their charm to me. If there was an AI that played Dominion really well I'd play it all day at first, expecting to learn a ton. Whether I'd get dissatisfied in the long run is something I can't tell without actually going through the process. I don't expect that to happen any time soon, though, because while it's straightforward to write a bot that plays a not too complex given kingdom fairly well, teaching it to play a random kingdom is more like teaching it to play Chess960.

I feel the same way. I was really into chess in middle school but not so much anymore and I don't have a strong desire to get back into it because I know I'll never be as good as a computer program.

I'm still really into chess... I don't see why I should care that a computer will always be better than me. I know the world's best grand masters will always be better than me too. All I have to do is not play those computers or grandmasters (although really, I would like to play GMs as a fun learning experience).

[LastFootnote]

This talk of banning "skill-reducing cards" turns my stomach. That is all.

well, given that you designed gambler, you obviously don't care about it, but it's a legitimate concern to some people.

The reason I hate it is that it's such a can of worms (as evidenced by e.g. Smogon). It means having endless goddamn arguments over what should and should not be banned and bad feelings all around. And for what? So the more-skilled player can win even more often? Not worth it.

[theblankman]

And for what? So the more-skilled player can win even more often?

"For what" is the easy part.  For games where you make more decisions that affect the outcome, because having your decisions matter is what makes the game fun.  The last several turns of a game in which you badly lost an important split are typically not fun; you just hope your opponent makes a mistake or gets some bad luck. 

if it's blatantly obvious that the best strategy is "buy lots of Cultists/Familiars/Hunting Parties/etc," and the winner is whoever gets the most plays of their good card, that's an un-fun kingdom.  That's the legitimate gripe with Pro as it is now: too many kingdoms with few or no interesting decisions.  On the other hand, the legitimate gripe with constructed kingdoms, i.e. Casual, is that any leaderboard based on such games is much too easily abused, as evidenced by what people did with Masq pins on Iso. 

So neither Pro nor Casual as currently constructed is "good enough" for the competitive crowd here on f.ds, which brings us to the hard part: What would be better?  I'm not in favor of blanket bans.  Iso's veto mode might be okay, but might also lead to de-facto bans.  Last time this topic came up, DonaldX had a proposal that I can't remember at the moment, I think it might have been veto but only if both players agree to it?  Hopefully one of these, or something else that hasn't been proposed yet, finds the happy medium. 

[Beyond Awesome]

A veto where everyone agrees might work because chances were I would not ban any cards except maybe tournament.

[theblankman]

A veto where everyone agrees might work because chances were I would not ban any cards except maybe tournament.

Whether veto mode works depends on what the goal is, and we seem to have two competing goals: exciting games, to use SCSN's words, and a leaderboard on which we can compete and feel like we earned our places, cause we seem to have a bunch of competitive sorts here.  Veto mode isn't so great for the competitive balance aspect, in either variant.  With unilateral veto, why would SCSN ever play an obvious IGG game against me?  With mutual veto, I have an incentive to make him play that game. 

There are other ways to give players limited kingdom control besides veto.  Maybe come from the opposite direction: have players pick a few of the cards, but not so many that they can totally control the game.  Say one card per player, the rest are random, and the players choose after seeing the random cards (so if I think it's an IGG game and don't like that, maybe I add Ambassador).  I'm not sure how much better that is for competitive balance, but at least it keeps cards from getting widely thrown out.

[Donald X.]

So neither Pro nor Casual as currently constructed is "good enough" for the competitive crowd here on f.ds, which brings us to the hard part: What would be better?  I'm not in favor of blanket bans.  Iso's veto mode might be okay, but might also lead to de-facto bans.  Last time this topic came up, DonaldX had a proposal that I can't remember at the moment, I think it might have been veto but only if both players agree to it?  Hopefully one of these, or something else that hasn't been proposed yet, finds the happy medium.

I looked it up. The proposal (a "how matchmaking should work" proposal I sent to MF in March) has it as, each player can have a hate-list of up to 3 cards, and in casual-rated games cards anyone in the game hate-listed are excluded. For pro-rated games, nothing is excluded (except unowned cards). You can hate-list cards despite not owning them.

There was talk back then of, if both players hate-list a card, can't that be excluded in pro-rated games, but there were people who somehow couldn't tolerate that.

[shark_bait]

And for what? So the more-skilled player can win even more often?

if it's blatantly obvious that the best strategy is "buy lots of Cultists/Familiars/Hunting Parties/etc," and the winner is whoever gets the most plays of their good card, that's an un-fun kingdom.  That's the legitimate gripe with Pro as it is now: too many kingdoms with few or no interesting decisions.

I have to politely disagree with you on that point.  But I guess I don't know what your metric of "too many" is.  Out of my past 20 games the following 3 are the only 3 that weren't too interesting to me.

http://gokologs.drunkensailor.org/static/logprettifier.html?20140904/log.50ea2963e4b0429cfe091cec.1409874353363.txt

Spice Merchant Open into mass FG buy a Merchant Guild at $5 and green.

http://gokologs.drunkensailor.org/static/logprettifier.html?20140903/log.50ea2963e4b0429cfe091cec.1409781587265.txt

5/2 against 4/3 with Mountebank.  No +Card or trashing defined game. 

http://gokologs.drunkensailor.org/static/logprettifier.html?20140904/log.50ea2963e4b0429cfe091cec.1409867616629.txt

Trading Post into Duchy/Duke.  KC present but no +Card and only +Buy Ruined Market.

So out of 20 games I had 3 that weren't too exciting.  IMO, 15% is definitely not "too many."  Especially when those other 85% make for exciting games wherein after over 5000 games I'm still encountering new situations.

[Mic Qsenoch]

Those 3 boards all look pretty cool to me, I guess the Mountebank one is the worst. The Duke board has a guaranteed (basically) Province a turn option with KC-KC-Navigator-Scheme-Scheme, and those can usually compete with Duke. The FG board has the potential of Advisors going for double Provinces. This deck could go with Merchant Ship and Fool's Gold, or I guess Soothsayer/Counterfeit. It is probably worse than straight FG, but it looks possible at least. The Mountebank one at least has Island (adds question: when do I get Islands?), and of course it's a messy game so you have the chance for nontrivial decisions about when to take Gold over Mountebank and when to go for Duchies.

So I'd say the percentage (for me) is even higher than 85%

My strong preference for sets of kingdoms is not "exciting" individual games, but a series with a mix of all the different kinds of games one can encounter in Dominion. Random has always done this very well in my experience. I think people are waaay to quick to reject games as being "obvious/boring" when there are actually lots of important decisions to make in the way the game plays out, even if you know for sure you want a particular strong card. It's easier to say "I lost this because it was a dumb IGG/Cultist/Rebuild game and it's a low skill card" instead of recognizing and figuring out the mistakes you made.

People mistake playing actions for decisions, sorry playing 5 Labs or your Village/Smithy chain does not constitute a significant decision.

[silverspawn]

yea there may very well be more skill in a rebuild game than in a basic village/smithy engine.

i still think there are low skill cards though.

[theblankman]

yea there may very well be more skill in a rebuild game than in a basic village/smithy engine.

i still think there are low skill cards though.

Notice rebuild wasn't one of my examples   It might be a "one-card strategy," but it's got a lot of tactics mixed in. 

People mistake playing actions for decisions, sorry playing 5 Labs or your Village/Smithy chain does not constitute a significant decision.

I'm not saying the only interesting boards are engines.  In fact I find Lab to be one of those annoying "open two silver and whoever wins the split wins the game" cards in a significant portion of the games it appears in.  I do find, though, that villages tend to make the other cards around them more useful; they let you have more different cards in your deck, which leads to decisions about which cards to buy, and how and when to play them.

I have to politely disagree with you on that point.  But I guess I don't know what your metric of "too many" is.  Out of my past 20 games the following 3 are the only 3 that weren't too interesting to me.

In a perfect world one is too many, but okay, I'll do the same exercise for my last 20 games...

http://gokosalvager.com/static/logprettifier.html?20140904/log.516d33dbe4b082c74d7b39fa.1409888181462.txt
We're gonna three-pile on Curse, Duchy and Duke.  Yawn.  The only decisions that felt less than automatic were "Second Sea Hag?" (if I happen to hit exactly $4 a second time early, then yes), and how many Vaults before Duchy. 

http://gokosalvager.com/static/logprettifier.html?20140904/log.516d33dbe4b082c74d7b39fa.1409870770980.txt
Masq is the only draw but it's still gotta be Goons engine, right? 

http://gokosalvager.com/static/logprettifier.html?20140904/log.508ae4c20cf264ac914cc446.1409868943740.txt
Jack.  Nuff said.

http://gokosalvager.com/static/logprettifier.html?20140904/log.516d33dbe4b082c74d7b39fa.1409864586924.txt
Win the Amb fight, win the game.

http://gokosalvager.com/static/logprettifier.html?20140904/log.516d33dbe4b082c74d7b39fa.1409863785973.txt
Bishop/Fortress.  Too bad too, cause there are some neat cards otherwise.  There might even be a more complicated Bishop engine (trashing Gold gained with Explorer) if the village was anything other than Fortress.  Probably Grand Markets would be better than Bishop/Explorer, but even that's more fun than just seeing who trashes down to the golden deck faster.

http://gokosalvager.com/static/logprettifier.html?20140904/log.516d33dbe4b082c74d7b39fa.1409861865384.txt
Herald is still shiny and new so the jury's out on it, but I'm starting to think if there's even a decent cantrip for it to hit (here Baker), then the high chance of being +2 cards, +2 actions makes it dominant for a $4-cost.  I'm afraid it could become another "win the split, win the game" card on a lot of its boards.  Let's count this game as half. 

http://gokosalvager.com/static/logprettifier.html?20140904/log.516d33dbe4b082c74d7b39fa.1409861147658.txt
Nothing to mitigate the Cultist-shaft here.  Lots of interesting cards but it doesn't matter if 7 ruins = game over, and Cultist makes it much easier to have a 7-3 junk split than any other attack.  If I had to pick one card I never want to see again, Cultist would be it. 

6.5/20 = 32.5% (counting the Herald game as half).  I agree that your 15% is not too many, but I think my one-third is.  Maybe I've had a bad run of random boards lately, maybe you've had a good one, hard to tell with the small samples we have here.

[silverspawn]

one third sounds right to me, which is why I sympathize a lot with this idea, and even more with a hatelist. thanks for making the point I was far too lazy to make myself 

[theblankman]

one third sounds right to me, which is why I sympathize a lot with this idea, and even more with a hatelist. thanks for making the point I was far too lazy to make myself 

Before I started my guess was a quarter, I was a little surprised to be near a third.  What's especially frustrating is that two of them were so close to being really good.  That Bishop-Fortress game has multiple good choices that just aren't as reliable or quick-to-build as the golden deck.  Replace Fortress with vanilla Village and I'd be saving that one for the next kingdom-design challenge.  The Cultist game isn't quite on that level, but I think that board gets a ton better if Cultist is replaced with even another draw/junk like Witch or Torturer. 

[Holger]

http://gokosalvager.com/static/logprettifier.html?20140904/log.516d33dbe4b082c74d7b39fa.1409863785973.txt
Bishop/Fortress.  Too bad too, cause there are some neat cards otherwise.  There might even be a more complicated Bishop engine (trashing Gold gained with Explorer) if the village was anything other than Fortress.  Probably Grand Markets would be better than Bishop/Explorer, but even that's more fun than just seeing who trashes down to the golden deck faster.

I think your opponent "killed himself" by needlessly trashing two of his Bishops. If he had continued to build the same golden deck as you had a few turns later, it would have been difficult for you to win the game since buying any card to three-pile would have risked breaking your golden deck (you couldn't afford the $5/$6 cantrips) and losing your lead. There might have been an interesting endgame...

[Mic Qsenoch]

I prefer the metric of interesting decisions per unit time, not per game. Simple games often have much shorter turns. Complex games offer lots of nice decisions at the price of taking forever. I find playing games along the whole continuum of these extremes helps me stay interested over a series of games.

blankman, I didn't mean to imply you only find engines interesting. I was just trying to point out that there's a superficial complexity about playing engines which doesn't correspond to real complexity of decisions.

I am also trying to suggest that people take a closer look at their games, what choices did they actually make. It is too easy to write "Jack mirror nothing to see", but when I look at the game logs there are often serious mistakes by both players. The 2-3 word summaries (FV-Wharf, Jack BM, etc.) are just a communication aid, it's a terrible mistake to think that it describes a strategy. It's a shame that the competitive Dominion community (myself included) is so willing to gloss over the important strategic decisions by relying so much on brief summaries because it's easier to talk/think about.

[blueblimp]

My beef with Rebuild is not really that it's low-skill (since that's arguable) but that, because it's both pretty strong and doesn't interact much with most other cards, the games with it tend to feel pretty similar to each other. That said, I still wouldn't choose to ban it yet because I haven't been particularly active post-DA, so I still have more to learn.

[Mic Qsenoch]

http://gokosalvager.com/static/logprettifier.html?20140904/log.516d33dbe4b082c74d7b39fa.1409888181462.txt
We're gonna three-pile on Curse, Duchy and Duke.  Yawn.  The only decisions that felt less than automatic were "Second Sea Hag?" (if I happen to hit exactly $4 a second time early, then yes), and how many Vaults before Duchy. 

http://gokosalvager.com/static/logprettifier.html?20140904/log.516d33dbe4b082c74d7b39fa.1409870770980.txt
Masq is the only draw but it's still gotta be Goons engine, right? 

http://gokosalvager.com/static/logprettifier.html?20140904/log.508ae4c20cf264ac914cc446.1409868943740.txt
Jack.  Nuff said.

http://gokosalvager.com/static/logprettifier.html?20140904/log.516d33dbe4b082c74d7b39fa.1409864586924.txt
Win the Amb fight, win the game.

http://gokosalvager.com/static/logprettifier.html?20140904/log.516d33dbe4b082c74d7b39fa.1409863785973.txt
Bishop/Fortress.  Too bad too, cause there are some neat cards otherwise.  There might even be a more complicated Bishop engine (trashing Gold gained with Explorer) if the village was anything other than Fortress.  Probably Grand Markets would be better than Bishop/Explorer, but even that's more fun than just seeing who trashes down to the golden deck faster.

http://gokosalvager.com/static/logprettifier.html?20140904/log.516d33dbe4b082c74d7b39fa.1409861865384.txt
Herald is still shiny and new so the jury's out on it, but I'm starting to think if there's even a decent cantrip for it to hit (here Baker), then the high chance of being +2 cards, +2 actions makes it dominant for a $4-cost.  I'm afraid it could become another "win the split, win the game" card on a lot of its boards.  Let's count this game as half. 

http://gokosalvager.com/static/logprettifier.html?20140904/log.516d33dbe4b082c74d7b39fa.1409861147658.txt
Nothing to mitigate the Cultist-shaft here.  Lots of interesting cards but it doesn't matter if 7 ruins = game over, and Cultist makes it much easier to have a 7-3 junk split than any other attack.  If I had to pick one card I never want to see again, Cultist would be it. 

I want to say Vault/Alchemist stack can beat Dukes if you compete with the Curses. But I could be kidding myself, it's something to consider. I still think there are interesting decisions around when to get Remodel, and obviously mistakes were still made (Procession for your opp?). Is any of his Walled Village and Inn stuff better than just Vaults and Silvers?

The Goons game is full of significant decisions: Remake/Masq? How many times to activate Market Square? When to add the second Masq? What to keep in hand after a Goons play? General deck balance between terminals and villages? How do you setup the 3 pile (I think you could have done it sooner)? Why take Pawn first over Hamlet? I mean it's obviously Goons, but just saying "Goons" doesn't tell you what the actual plays to make are. Nobody is going to play this game anywhere near perfectly on the first go.

The Jack game: Why no Courtyards? I am sure opening Courtyard/Jack here. Is there a place for a Rogue in your deck when your opponent is already buying Duchies on Turn 7? Besides, this game probably took 5 minutes (which may not matter to you, but it is a strong point in it's favor for me)

Ambassador game: It's not at all clear that deck thinning needs to be prioritized here. Which means single Ambassador is also an option. Amb/Moneylender is a reasonable opening. Swindler could be extremely useful in this game, but nobody touches one ever. Your opponent buys lots of Spies over something more useful like Silver/Swindler. You draw both Ambassadors on turns 12 and 15, but keep both when you should definitely get rid of one. You both keep feeding each other Coppers late into the game where it's not clear whether that helps or hurts your opponent.

I am pretty sure the Bishop/Fortress game can incorporate more of the board to produce more points per turn than you are getting here.

The Herald game just looks like you played very nicely, I don't know what the issue is. You went for Ironworks and cycling (Warehouse) earlier. You targeted the Heralds as important. Your choices made you win the game.

Cultist one sure is rough.

I am not trying to convince you to enjoy these example games, enjoy whatever you like, I am trying to point out that they aren't as straightforward as your summaries indicate, and maybe there are some nice things to think about while playing these games. You say you don't like when winning a card split is important, but that's a LOT of Dominion games (if nothing else it's Province/Duchy split), the skill is in figuring out how to win the important splits and how to squeak out wins if you don't.

[SirD]

I think Salvager already found a near to perfect implementation for the point tracker. Whenever some additional tracking capabilities will be included, they most likely will be implemented the same way, so that every player can choose freely to disable them:

#troff
#tron
#tr?

I like this very much.

[silverspawn]

well I have to admit that I was simplifying the problem. If you just look at the games that have only trivial decisions, that's less than 1/3. But if you look at the games that have "uninteresting" decisions, that's more.

Another thing is the importance of luck and skill for certain cards. for rebuild, I think that while it does take a lot of skill to play it perfectly, doing so still only amounts to a very small edge over playing it decently. in other words, even though rebuild is not low skill, you will lose a lot of matches against people that are significantly worse than you. On most engine boards though, skill matters a lot more. Generally, I tend to appreciate decisions/skill more if it also has more impact, and I think a lot of people do.