Print

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

[timchen]

How about a house rule stating that
"You may reshuffle all the cards from your discard into your deck in your clean up phase, before you draw the cards for your next turn, if you have shuffled in this turn."

This rule will help on those unfortunate situations where the victory is determined by unlucky reshuffle misses, such as drawing the witch at the bottom of a deck.

For a casual game this rule might be too bothersome to implement, but I think in a competitive environment it is worthwhile.

Thoughts?

[aaron0013]

Not sure that I am completely following you, but it seem like this would make certain cards too powerful.  First one that comes to mind is stash.  You can see where that would lead...

[Lekkit]

The way I interprete it is that if you trigger a reshuffle during your turn, you may reshuffle again at the clean up. So if you play a Witch on turn 4 you can get to play it on turn 5 again.

[TrojH]

Here's a house rule to reduce luck factors:

"Don't play Dominion."

Face it. Horrible luck is a part of the game. 

[aaron0013]

The way I interprete it is that if you trigger a reshuffle during your turn, you may reshuffle again at the clean up. So if you play a Witch on turn 4 you can get to play it on turn 5 again.

But that is one of my favorite aspects of the game: know what cards are in your deck, and know when to trigger reshuffles.  The better you get at that, the better you become at Dominion.

[Watno]

Different fix: play more dominion

[timchen]

Play more fixes nothing....

[michaeljb]

Play more fixes nothing....

Play more and (statistically speaking) you will see good luck even out the bad luck.

[Morgrim7]

No house rules i've come up with have made dominion better.

Horrible luck is a part of the game.

[rod-]

Play more fixes nothing....

Play more and (statistically speaking) you will see good luck even out the bad luck.

Variance being evenly distributed about the mean has no effect on the fact that the variance is high. (Statistically speaking)

[Tables]

Play more fixes nothing....

Play more and (statistically speaking) you will see good luck even out the bad luck.

Variance being evenly distributed about the mean has no effect on the fact that the variance is high. (Statistically speaking)

No, but your luck compared to the best and worst possible cases will very quickly tend towards something roughly in the middle.

[rod-]

Suggestion for mitigating unequal tax burden:

Spend time being both poor and rich.  Sometimes you will pay taxes, sometimes you will get a refund.

Maybe that's too subtle.  The point is:  The proposed "solution" is just ignoring the underlying problem. 

[rod-]

I have actually put a decent amount of thought into solving the reshuffle problem, and when it comes to online implementations, it is quite trivial:

Each card has a 'has been drawn X times' counter associated with it.  Every bought/gained card is assigned the highest value of that counter. (EG: On turn 1, your starting 3 coppers and 2 estates have a counter = 1.  The silver you buy also has a counter = 1).  You will always draw cards that have counter = 0 before counter = 1, so you'll get your 4/1 hand before that silver.  However, you will *never* get your starting estate a fourth time before you get that silver a second.

Effectively, nothing ever 'misses' a reshuffle - every card will get drawn exactly one time per deck cycle. 

Mostly impossible to do offline, though, so i can't really fault Don for leaving it out of Dom

[SirPeebles]

I really like having some luck though.  I don't want Dominion to be a game where you can plan your moves several turns ahead of time.  I like that as you construct your deck, you need to prepare for some randomness and adapt accordingly.  In the long run, the better player really does win.  This is why I'm not at the top of the leader board.  But it's also nice that sometimes the weaker player wins.

[DStu]

The proposed "solution" is just ignoring the underlying problem.

I would guess the proposed "solution" was mostly a joke, which of course partly reduces the problem, as the more you play the more likely the better player wins.

What's true is that it is not very efficient, quadratic runtime and such, and also probably not applicable, because of course one should assume that everybody already plays as much Dominion as time permits, so the possibility to increase that time does not exists.

Your solution next post is of course nearly the optimal solution to this problem, or at least tries to be it with some correctelable edgecases.  When reducing variance is the goal, going to deterministic is optimal.
Edge-cases: If you discard cards with say Warehouses, and redraw them with Lab the same turn, and iterate this, you can increase the counter of some cards by more than 1 per turn. I don't think this is a desired behaviour.  Would be a nice way to get rid of some Victory cards from your draws.  Don't want to think how to solve this, but should be easy.

[rod-]

The proposed "solution" is just ignoring the underlying problem.

I would guess the proposed "solution" was mostly a joke, which of course partly reduces the problem, as the more you play the more likely the better player wins.

None of the four different people who said "deal with it" gave any indication that they were joking.  Also, if four people make the same joke, shouldn't it at least be funny?  I don't think any of them were joking.  They're all coming into this topic effectively just to say "I don't have a problem with that, you shouldn't either".

Peebles' comment is another of the same.

Play regular dominion all you want, a house rule for reducing luck is no threat to that. 

I don't know why I'm all up in arms trying to keep posts on topic for this topic in particular, i know the internet's full of failure, and i know i've posted plenty of things off-topic, but at least i don't go into the "what's missing?" threads and post "an actual game of dominion" as the answer to every one.  If you don't see how those two are the same, maybe someone else can clarify for you.

[Piemaster]

Why bother with shuffling at all?  Why not just make a rule where every time you would shuffle, you just reorder your deck in whatever way you like instead?

[DStu]

The proposed "solution" is just ignoring the underlying problem.

I would guess the proposed "solution" was mostly a joke, which of course partly reduces the problem, as the more you play the more likely the better player wins.

I was thinking of Watno, but anyway...

[DStu]

Why bother with shuffling at all?  Why not just make a rule where every time you would shuffle, you just reorder your deck in whatever way you like instead?

Because that is an entirely different game with different power levels of cards (->Chancellor).

[Drab Emordnilap]

Why bother with shuffling at all?  Why not just make a rule where every time you would shuffle, you just reorder your deck in whatever way you like instead?

"Left hand opponent reorders your deck" is more fun.

[SirPeebles]

To give a more reasonable response, I agree that there is some randomness/luck which is positive, and some which is negative.  In particular, randomness which contributes to the strategic depth is, in my opinion, quite positive.  I like knowing that I need to be cautious while building my village/smithy engine for instance.  That is, there's a fundamental and strategic notion of reliability which randomness thrusts upon the game.  Another example is the importance of managing your reshuffles.

On the other hand, it is quite frustrating when someone utterly disregards reliability, and yet gets a lucky draw and megaturns anyhow.  Similarly, it sucks when your opponent's Sea Hag hits your Sea Hag.  Was it wrong to buy that Sea Hag?  Probably not.  Was there something you could have done to protect yourself?  You likely didn't have the time.

And of course, many cards are balanced with these factors in mind. 

"Play more games" was mostly likely suggested tongue in cheek, since the spirit of this topic is clearly to decrease the amount of luck within a game.  But at least what I'm suggesting is that the randomness does not necessarily diminish the role of skill in the game, and in fact contributes to the strategy.

I would be interested in a house rule which decreases the "bad" effects of randomness while preserving the "good".  The suggestion in the original post seems to do more to remove the strategic concept of shuffle management than it does to mitigate the impact of shuffle luck.

[WanderingWinder]

The total re-order of your deck allows for weird best/worst case luck (who is doing the ordering), which is usually really off from the average case. Which mostly makes it into an entirely different game, but one which among other things has ENORMOUS AP - if you get rid of the randomness of the shuffling, the game should be quite directly 'solvable' by pretty direct backward induction, but this will take FOREVER. Of course, you can do this with the probabilistic game we actually have too, in terms of probabilities, but it's so complicated that, like chess, you just aren't going to even try to calculate the game to the end from the start position.

So "Play more games" or "Play a different game" are sort of tongue in cheek, but I think at some point, you do actually have to acknowledge that the luck thing is an inevitable issue with dominion. I mean, we can say the same thing about poker, no? But even something as modest (and complicated) as rod_ suggests in reply#12, you actually get some issues - what happens is that the cards that miss the shuffle all get clumped at the start not only of the shuffle immediately after they miss, but also in the next shuffle. Which lets me plan ahead somewhat for certain collisions. Is it tricky and sorta small? Yes, but especially with software aid, it's doable. And this doesn't solve things like sea hagging a sea hag. The problem is that any way of forcing some regularity in is going to fundamentally change the nature of the game by changing either the percentage of times that particular sets of cards collide downwards or upwards, and you're looking at a quite different game.

[LastFootnote]

I am vehemently against any rule that 'reduces luck' in a game. I can see the reasoning behind wanting reduced luck; you're a competitive player and it sucks losing games to weaker players because you get unlucky. You want that to happen less often.

But when the community surrounding a game starts to embrace that kind of rule, it can easily become a slippery slope, and I don't use that expression lightly. If you want proof, look at the competitive online Pokemon community. Pokemon (the video game, not the TCG) is the game I played competitively before Dominion. It could be a great game, with deep strategy and lots of potential variety, two of the things I also love about Dominion. The problem with Pokemon is that it's built to be unbalanced. Little kids just love their legendary Pogeymons! So in order to make the game interesting, certain things had to be banned. Among those things were some moves that the community felt "promoted luck over strategy".

I won't get too far into the details, but suffice it to say that this sort of rule pretty much just breeds discontent and frustration in the community once you have a bunch of player-created rules for minimizing luck. Is Stash too powerful using this new rule? Let's just ban Stash. No big loss, right? What, 30% of the players don't like that Stash is banned? Etc., etc.

The beauty of Dominion is that it's a great, balanced game where every card can shine given the right circumstances. Although I understand the desire to "improve" it, I'd prefer to play it as is and leave the metagame design for other, less-balanced games.

[blueblimp]

I agree that luck mattering more than skill becomes a problem at higher levels of Dominion. I'm not that good a player (level 35-ish), yet I'd guess more than half the games I play are decided more by luck than by skill. That's boring. (The games that are decided by skill are still quite interesting though!)

I think the solution here is NOT to try to reduce luck, but try to increase skill. Witch+BM with zero luck would still be boring, because then it'd just come down to analysis and calculation, and hey if I wanted that then I'd play Chess or Go. So the problem is that mirror strategies reduce the importance of skill. Skill can still be more important than luck in mirrors; for example, if Amb is available and only one player plays it well. But generally there isn't enough opportunity for skill to influence the outcome in mirrors.

I like one idea that was posted here about increasing the number of kingdom cards. Instead of 10, have 12 or more. That increases the number of viable strategies, leading to fewer mirrors, and it also increases the number of card interactions, which favours skill.

[Drab Emordnilap]

I like one idea that was posted here about increasing the number of kingdom cards. Instead of 10, have 12 or more. That increases the number of viable strategies, leading to fewer mirrors, and it also increases the number of card interactions, which favours skill.

I would argue the opposite; the more card in the kingdom, the more likely that some combination of two or three is the clear best strategy. At least, it's not as clear cut as "more options means more viable strategies".

[timchen]

Regarding to rod-'s solution, if I understand it correctly, will become pretty weird in some cases-- for example, suppose that silver is the 11th card in the second shuffle, and if I discard it with a warehouse, do I immediately reinsert it into the deck? Another situation is that when one can occasionally draw their deck. In this case, they can expect always to start with those cards that were drawn fewer times first. This actually sounds good for the engine.

Really, my original intention is pretty simple. In dominion, it is horrible (not only for you, but for the strategy depth of the game) for a key card to miss a reshuffle. It is not only late, but can only be played one less time. It is almost game determining between similarly skilled opponents if it happens in the first reshuffle. Think about the turn 5 chapel. Or turn 5 witch when both players start 5-2. A game determined by this kind of swing are unsatisfactory for both players, assuming that the luckier player has enough skill to realize what happens. Indeed I think those who blindly suggested to play more games are the ones who need to play more.

Why is stash buffed with this suggestion? I don't quite see it.

[dondon151]

But when the community surrounding a game starts to embrace that kind of rule, it can easily become a slippery slope, and I don't use that expression lightly. If you want proof, look at the competitive online Pokemon community. Pokemon (the video game, not the TCG) is the game I played competitively before Dominion. It could be a great game, with deep strategy and lots of potential variety, two of the things I also love about Dominion. The problem with Pokemon is that it's built to be unbalanced. Little kids just love their legendary Pogeymons! So in order to make the game interesting, certain things had to be banned. Among those things were some moves that the community felt "promoted luck over strategy".

I won't get too far into the details, but suffice it to say that this sort of rule pretty much just breeds discontent and frustration in the community once you have a bunch of player-created rules for minimizing luck.

Evasion and OHKO clause were part of the competitive scene probably long before you started playing, and the existence of those clauses did not detract a single bit from the game. The policy review process has become a bureaucracy since I was last active on Smogon, but you simply cannot argue that the enforced bans have not been healthy for the metagame. Just because a game was "built" to be unbalanced (which, by the way, is not entirely true according to an interview with the developers) does not mean that players who would like to play it in a competitive fashion have to look elsewhere.

Indeed I think those who blindly suggested to play more games are the ones who need to play more.

Or maybe you need a better sense of humor?

[TWoos]

Why is stash buffed with this suggestion? I don't quite see it.

As I understood the suggestion, Stash is buffed because when you trigger that reshuffle to draw during your turn, you can put Stash on top and draw it.  Then you allow another reshuffle, putting it on top again.  This is more obviously powerful with multiple Stash.

[rod-]

even something as modest (and complicated) as rod_ suggests in reply#12, you actually get some issues - what happens is that the cards that miss the shuffle all get clumped at the start not only of the shuffle immediately after they miss, but also in the next shuffle.

Technically, nothing would ever miss a shuffle in my suggestion, as there would be no such thing as a shuffle.  Supposing that you opened silver/silver, and both were randomly selected to be drawn 11th/12th, they woudn't be drawn 13th-15th (you still can't draw the same card more than once per turn), but they could be drawn anywhere thereafter.  Essentially, you reshuffle your deck every turn, but nothing would get "stuck" at the top / bottom and played disproportionately as a result.

Re: Warehouse, yes, you could expect to draw a card that you just discarded, assuming it was among the cards previously drawn the fewest times.

Re: Drawing your deck, starting with cards that were drawn fewer times is not a priori better for an engine.  It's often worse, as your engine is less likely to "go off" if you drew all of your coppers and estates first.  I'd imagine that, when it comes to something like hunting party stacks that "draw" a lot of bad cards over and over again, there would have to be some sort of qualification (no card can have a counter more than 1 greater than the lowest?) to prevent over-priming your draw.

Edit:  Maybe a better way to implement the OT suggestion (and which may potentially be physically possible) would be to simply mark / note one card left out of the current shuffle, then randomly insert it in the deck during the following cleanup.  You can still have cards missing shuffles, but you can't just throw your action chains together every turn without proper trashing.

[timchen]

Or maybe you need a better sense of humor?

I really do.

[aaron0013]

This is terribly changing the subject, and I am sorry, but does this not look like a smiley face eating something?

= 0 before counter = 1

BTW, Wandering Winder, how long have you been back?  Just wanted to put in a plug that your videos caused me to become a fan of Dominion.

[Donald X.]

In dominion, it is horrible (not only for you, but for the strategy depth of the game) for a key card to miss a reshuffle.

You could mean, that luck reduces strategy. If so that's just wrong, they are different axes. Or you could mean, you personally don't like the combination of strategy and luck. That's fine but not so typical. If you have a certain amount of strategy in a game then some people will not like it if you have a certain amount of luck; you don't want to feel like the thinking you're doing is pointless. I think obv. Dominion does not have that issue for most people.

When I showed the game to RGG, a card at the bottom didn't just miss out on a shuffle - you weren't guaranteed to ever draw it. If you had to draw n cards and there weren't n left, you shuffled before drawing (just once since there might still not be n left). That turn one Witch that ended up on the bottom was not guaranteed to show up turn 5+ either.

Valerie wanted the more conventional "draw then shuffle" (just to be more conventional, not for any other reason). And that sounded fine, except, would it reduce luck too much? As it turned out, it didn't.

[timchen]

In dominion, it is horrible (not only for you, but for the strategy depth of the game) for a key card to miss a reshuffle.

You could mean, that luck reduces strategy. If so that's just wrong, they are different axes. Or you could mean, you personally don't like the combination of strategy and luck. That's fine but not so typical. If you have a certain amount of strategy in a game then some people will not like it if you have a certain amount of luck; you don't want to feel like the thinking you're doing is pointless. I think obv. Dominion does not have that issue for most people.

Sure, I do not imply in general luck reduces strategy. I only mean that in specific cases they do. And for some of those cases we can actually eliminate those "bad" luck.

Say for example why do we limit ourselves to identical openings in a tournament? It is just because on certain boards it may be hard to compete if you get an unfavorable opening split. For a typical example, on a large number of boards there is no good play with a 4-3 opening after your opponent opened witch/mountebank.

Here the situation is similar to some extent. When a key card misses a reshuffle early it can be hard to recover. And in those situations there is not that much strategy to worry about after that happens. You pretty much just wait and see if opponent blunders or if miracle happens.

One analogy you made that I like is the modified chess with coin flip game. In a sense the game is of identical strategy depth with the original chess, but in reality it is not. Say if I were to play 1,000 games of this modified chess in my life time, since in those 1,000 games in average less than 1 game will end up longer than 10 moves, it only takes me the skill to not lose in 10 moves, to be as good as a chess master, in those 1000 games I'll ever play. However, it only takes a single game of the original chess for the master to teach me a lesson! In other words, in this modified game of chess the coin flip part is really introducing luck factors which reduces the strategy depth of the game and the game can benefit by reducing this luck factor. (At least, for those who want to compete at the game.)

[SirPeebles]

... after your opponent opened witch/mountebank.

Ouch.  Damn you, zaps.

[Donald X.]

Sure, I do not imply in general luck reduces strategy. I only mean that in specific cases they do.

The zero cases?

Perhaps by "strategy" you mean "the tendency of strategy to determine the winner." In that case the word you are looking for is actually "luck."

One analogy you made that I like is the modified chess with coin flip game. In a sense the game is of identical strategy depth with the original chess, but in reality it is not. Say if I were to play 1,000 games of this modified chess in my life time, since in those 1,000 games in average less than 1 game will end up longer than 10 moves, it only takes me the skill to not lose in 10 moves, to be as good as a chess master, in those 1000 games I'll ever play. However, it only takes a single game of the original chess for the master to teach me a lesson! In other words, in this modified game of chess the coin flip part is really introducing luck factors which reduces the strategy depth of the game and the game can benefit by reducing this luck factor. (At least, for those who want to compete at the game.)

Chess-with-dice (or in this case a coin flip) does not have less strategy than chess. It just has more luck.

They are not the same axis. That is in fact the whole point of the chess-with-dice example.

[SirPeebles]

I'm not familiar with this "chess with coin flip" example, but from context I gather that you play a game of chess and at the start of each of your turns you flip a coin:  If heads you win, if tails you continue playing.  Your opponent has no such coin flip.

In this case, your strategy really ought to be to just stay alive.  Playing offensive would be foolish.  I would argue that while this variant offers as many choices as chess, it lacks the strategic depth.

[timchen]

@SirPeebles: almost that, except your opponent has the coin flip as well. Still I would say practically a noob's strategy should still be trying to stay alive as long as possible, and he really does not need to think past that.

But I guess the point is starting to be missed here; okay we can say luck and strategy are two axes. The point is just that (in your terms) when the luck is independent of the strategy, in a competitive setting one should reduce that kind of luck as much as possible.

[Donald X.]

But I guess the point is starting to be missed here; okay we can say luck and strategy are two axes. The point is just that (in your terms) when the luck is independent of the strategy, in a competitive setting one should reduce that kind of luck as much as possible.

I'm not sure how you're going to define "independent" there, but chess with dice (at the start of your turn, roll two dice, on boxcars you win) is not a game people in general would like as much as chess, even though in general people prefer luck to the lack of it.

[Titandrake]

The reason I'm not a big fan of rules about limiting reshuffle luck is that I feel that the strategy in managing reshuffles is more than worth it.

In a lot of games, "strategy" usually includes setting up safety margins so that luck doesn't horribly destroy you. Dominion is no different. Really, a lot of Dominion strategy is about adding in redundancies and accounting for as many factors of luck as possible.

Consider the situation you describe about a Witch that misses the reshuffle. Yes, in this instance, it sucks. But what about a game where you have an engine, you've played most of your action cards, and you're just before a reshuffle. Do you keep going, or no? In the Dominion we play now, this is an interesting choice because you have to balance a better reward now + worse hands later with decent reward now + consistent hands later. And it's not at all clear which is better, and that rewards skill with the game. However, if you know that you can simply reshuffle, you can just keep going, and shuffle all your engine pieces in. This removes the risk in "risk and reward."

Still, in a tournament setting there do need to be some limits on luck. Identical hands is useful on some boards (but arguably detracts strategy space on others). But overall, the tendency for reshuffle-skipping is to create interesting decisions, and trying to stop reshuffle luck removes some of those decisions that help separate skill levels.

[rod-]

A talk (about which games are chosen to be featured on Wil Wheaton's Tabletop) at BGG.con last week would tend to agree that strategy and luck are independent axes where games are concerned.  Everyone in the audience bought that supposition.

The producer's rationale for choosing the games he chose was that high strategy + high luck games give nubs the feeling that they could've won if only X+Y, while still giving experienced players the feeling that their actions weren't meaningless.  (These games are also better "television").  Coin flip chess goes too far along the luck axis for any experienced player.  Regular chess goes too far down on the luck axis for any nub.

I believe the argument of some (many? - certainly the OP of this thread) here is that in Dominion games between two experienced players, the luck rating is still unacceptably high (if only, perhaps, by a slight amount).

[Donald X.]

One analogy you made that I like is the modified chess with coin flip game. In a sense the game is of identical strategy depth with the original chess, but in reality it is not.

Here is one you may find clearer.

Play a game of chess. At the end flip a coin. If you get heads and your opponent doesn't, you win regardless of who won the game of chess. Otherwise the winner of the game of chess wins.

It should be clear that this game has *exactly* the strategic depth of chess; all decisions are made exactly as in chess (well, not the one about whether or not to play it). Yet it has far more luck than chess. Adding luck here did not reduce strategy.

[SirPeebles]

I still see a distinction between luck and randomness.  Randomness does tend to open the door to luck, but it can also expand strategy, as has been mentioned with shuffle luck.  Another example I've seen mentioned elsewhere has to do with making the most of variance. 

Suppose a card give you the following option.  Choose one:  Gain 2 Silvers; Gain 1 Copper and 1 Gold.

Either option has the same impact on money density and therefore on the average hand.  But in response to your environment, one may be better than another.  The Copper and Gold gives more variance.  You'll get more trashy hands, but you're more likely to hit spikes of $8 even with a money density below $8/5cards.  This variance may also be leveraged if you are using sifters like Warehouse or Embassy, or if you are the target of handsize decreasers like Militia or Margrave.

On the other hand, gaining the Silver made be more valuable if you're looking to consistently hit $4, perhaps in a Gardens or Silk Road rush.

Without variance, we would not have these strategic decisions.  So randomness is not independent of strategy.  However, randomness does not necessarily equate to luck either.

[ftl]

Oh hey, I like that coin-flip-at-the-end-of-chess analogy. I think I get the two-axis thing now.

Thing is, I think Dominion games are typically short enough that you can reduce the effect of luck by just playing more of them. One dominion game between may well come down to shuffle luck, best-of-three will give the better player a chance to catch up despite some bad luck, and if you move it up to, say, best-of-seven, then it's pretty unlikely that shuffle luck will decide the match.

Even a best-of-nine Dominion game is going to take less time than a serious chess match (or a 'serious' competitive match of lots of different games and sports.) and is basically guaranteed to come down to player skill.

[Saucery]

Randomness can shape the landscape of the game, but it shouldn't bury one player (and not others) in the process.

[rod-]

Thing is, I think Dominion games are typically short enough that you can reduce the effect of luck by just playing more of them. One dominion game between may well come down to shuffle luck, best-of-three will give the better player a chance to catch up despite some bad luck, and if you move it up to, say, best-of-seven, then it's pretty unlikely that shuffle luck will decide the match.

As long as the margin of difference between "better" and "worse" players is 10 levels, i agree.   As soon as it gets down to +/- 5, variance kicks in and you can't expect to get the same result.  When you go back to the council room data (i know, i know...how dare i remember any of these stats) you have to remember that "only" 5 levels translates to a pretty slim marginal edge.  It's (relatively) easy to judge a weighted coin in 7 flips if it's a 90%/10% flip, but more or less impossible when it's a 55/45, and no point trying when it's 52/48.

I don't know of any data showing how many games are dictated by 'shuffle luck', but given that a bot mirror match "should" end in a P1 victory barring any significant shuffle luck effects, it might be as high as 40%, although 20% may be more palatable to most in the crowd.

[blueblimp]

I like one idea that was posted here about increasing the number of kingdom cards. Instead of 10, have 12 or more. That increases the number of viable strategies, leading to fewer mirrors, and it also increases the number of card interactions, which favours skill.

I would argue the opposite; the more card in the kingdom, the more likely that some combination of two or three is the clear best strategy. At least, it's not as clear cut as "more options means more viable strategies".

I agree that in principle it's possible for adding more kingdom cards to reduce strategy, but taken to an extreme, if this were always true then 9 kingdom cards would have more strategy than 10. I feel that for inexperienced players, 10 feels like a lot of kingdom cards, but for experienced players it really doesn't, since some of them are immediately dismissed as irrelevant. (Sometimes, 9 are dismissed and the best strategy becomes BM+X!)

[Donald X.]

I still see a distinction between luck and randomness.

In fact the Richard Garfield essay that chess-with-dice comes from also argues that chess itself, straight chess, has luck.

Consider the case where I make my moves randomly. It is really unlikely I'll win - but the chance isn't 0%. Chess is very low on luck, but does not have zero luck. It doesn't have any randomness though - or, you could put it as, the randomness is confined to the players.

[Donald X.]

Oh hey, I like that coin-flip-at-the-end-of-chess analogy. I think I get the two-axis thing now.

Thing is, I think Dominion games are typically short enough that you can reduce the effect of luck by just playing more of them.

I am pleased it worked for someone, and yes, just play best 4 out of 7 or whatever if you want skill to shine through.

[Kirian]

Even a best-of-nine Dominion game is going to take less time than a serious chess match (or a 'serious' competitive match of lots of different games and sports.) and is basically guaranteed to come down to player skill.

This is only true online, of course.  In person, depending on the board, a single game takes at least five minutes to set up (most of which is spent finding the kingdom cards) and ten minutes to play (for two players).  That's a minimum--with crazier cards that force lots of shuffling (Hunting Party, I'm looking at you), or with lots of attacks, that time can go up to as much as half an hour.  A best-of-nine set of 2P Dominion at a table could take as long as 4 hours or so, as long as many of the longest regulation chess matches, and much longer than most sport matches (cricket excepted, but then cricket's wacky in so many ways).

[GigaKnight]

This is really interesting, Donald.  Can I pick your brain a bit on the role of luck?

Valerie wanted the more conventional "draw then shuffle" (just to be more conventional, not for any other reason). And that sounded fine, except, would it reduce luck too much? As it turned out, it didn't.

From this, it sounds like you're content with the amount of luck in Dominion.  If all other things were equal, would you prefer a different amount of luck in Dominion?  More or less?

Related, for the shuffling mechanic, do you prefer the original (pre-Valerie) or the current (more conventional) approach?  Would you prefer a game where a card could potentially never be drawn (trapped on the bottom)?  Or is it a wash to you?

[GigaKnight]

Also, FWIW, I was curious about Richard Garfield's thoughts on luck vs skill and found this, which is about an hour long



I think this basically has the same content as what DXV was referencing.

The biggest thing that helps clarify DXV's statement in this thread is that Garfield defines luck as simply as "uncertainty of outcome".  By that definition, literally all games have some amount of luck (if nothing else, a meteorite could kill a player, after all) but, IMO, it's more interesting to talk about where they fall on the luck/skill axes.

[Donald X.]

From this, it sounds like you're content with the amount of luck in Dominion.  If all other things were equal, would you prefer a different amount of luck in Dominion?  More or less?

Related, for the shuffling mechanic, do you prefer the original (pre-Valerie) or the current (more conventional) approach?  Would you prefer a game where a card could potentially never be drawn (trapped on the bottom)?  Or is it a wash to you?

I am happy with the amount of luck in Dominion. Obv. it varies from game to game; some cards are swingier and some reward skill more. I enjoy the range; I like the feeling of the pressure being off that comes with the swingier cards, and the crazy moments when they swing especially far, but I also like the skill-intensive cards.

I prefer the shuffling how it is. I automatically did it the other way, without thinking, because I have done that kind of thing in multiple games, as a way to lesson the value of card-counting. In a game with one shared deck that we tend to go through in a single game, it can be good to track what cards are left, and some people really don't enjoy that. If for example you shuffle five from the bottom, then everyone can give up on card counting without feeling like they're playing poorly. It's a great trick for certain games.

So then for Dominion I did that automatically and then kept it just because I liked the amount of randomness the game had. Like I said, when the change was proposed, I worried about whether or not that was getting rid of too much luck, and as you know of course it was not. Shuffling at the bottom is more natural and that's worth something.

Dominion is an unusual case here, since you go through the deck many times and it's your own deck and stuff. For certain games with shared decks and an incentive to count cards, I do prefer not going through the entire deck. Sometimes it doesn't do anything; in San Juan I will not know what cards you drew and discarded (discards are face down and plentiful), so I don't have to worry about what's left when we get near the bottom.

[Donald X.]

Also, FWIW, I was curious about Richard Garfield's thoughts on luck vs skill and found this, which is about an hour long



I think this basically has the same content as what DXV was referencing.

The biggest thing that helps clarify DXV's statement in this thread is that Garfield defines luck as simply as "uncertainty of outcome".  By that definition, literally all games have some amount of luck (if nothing else, a meteorite could kill a player, after all) but, IMO, it's more interesting to talk about where they fall on the luck/skill axes.

I have seen some other version of that speech, probably hitting on some of the same things. I remember in particular him giving "guess the nth digit of pi" as an example of a high luck game that has no random element. Anyway the version I saw was good.

Chess-with-dice I first read about in the Duelist, and that article is online now, so here it is: http://www.wizards.com/magic/Magazine/Article.aspx?x=mtg/daily/feature/119

[Piemaster]

Say for example why do we limit ourselves to identical openings in a tournament?

IIRC wasn't it because the person running the tournament (WW?) decided he liked that rule and basically enforced it upon everyone else?  As far as I am aware there are no official tournament rules to Dominion.

[yudantaiteki]

Chess-with-dice (or in this case a coin flip) does not have less strategy than chess. It just has more luck.

They are not the same axis. That is in fact the whole point of the chess-with-dice example.

I think I prefer the Dualist article's assertion that the game has the same strategy but that strategy is not rewarded as much as regular chess.

That just gets into what constitutes a game "having strategy".  Someone else used a different example from the dualist with even higher luck -- every turn after you make your move, you flip a coin, and if you flip heads you automatically win.  Personally I think this game has little or no strategy.  There are strategic choices to make, but in the end they don't matter because statistically, no game is ever going to get out of even the opening phases.  The chance of anyone actually winning the game by checkmate rather than by the coin flip is close to zero.  For all practical purposes, endgame strategy does not exist in this modified game.  So while there may technically be the same strategic choices as in regular chess, the choices make no difference -- to me, this means the game does not have strategy.

But I don't think this is a perfect analogy because the coin flip is something external to the game.  You can introduce luck that affects the strategy as well.  For instance, let's say that you play chess, but when you make a capture, you have to flip a coin.  If you flip heads, your piece gets taken instead.  To me this increases luck *and* reduces strategy because you can no longer count on your strategic decisions to produce the result you expected.  To me, this is a closer analogy to a card missing the reshuffle in Dominion.  And in that sense, I think Garfield is wrong to say that the amount of skill and luck are unrelated.

[DStu]

But I don't think this is a perfect analogy because the coin flip is something external to the game.  You can introduce luck that affects the strategy as well.  For instance, let's say that you play chess, but when you make a capture, you have to flip a coin.  If you flip heads, your piece gets taken instead.  To me this increases luck *and* reduces strategy because you can no longer count on your strategic decisions to produce the result you expected.  To me, this is a closer analogy to a card missing the reshuffle in Dominion.  And in that sense, I think Garfield is wrong to say that the amount of skill and luck are unrelated.

Na, that is too far on the other side now.  This would somehow mean that when a card misses a reshuffle, your opponent get to play it for free...

[Piemaster]

It's always difficult when you start trying to grade games according to how much luck they have relative to the amount of skill.  The thing is, with any game, the better players get at it the more luck it seems to have. 

If I played with a full set of Dominion against a brand new player (and I tried my hardest) then it isn't inconceivable that I would win the first 20-25 games quite easily.  And that's not because I'm great at the game, just because it would take that many games for an average Joe to get close to even competent at the game to be able to compete with a run-of-the-mill 20-something on Isotropic.  The game would essentially have a near-zero amount of luck in terms of the outcome. 

But then if we kept on playing he would improve enough to win the odd game here and there, occasionally through good play, but mostly due to luck.  Further on still, we could get to a point where we are quite evenly matched in terms of skill.  Some games would be determined by one of us picking the better strategy, but a large number would come down to luck.

Then if we kept on playing on and on into infinity we would get better and better at picking the optimal strategy until eventually there would be little to decide our games except for luck.

I'm not saying anybody has got to the point where they routinely choose the perfect strategy yet, but when a player of isotropic level 40+ plays against another player of level 40+ then they are both going to have a very deep understanding of the game and will often both decide on the same optimal strategy.  Or if they decide on different ones they are likely both of similar value.  Occasionally you may get a curveball kingdom which allows one of them to make a decisive decision to win the game, but ultimately most of their games are going to come down to luck to a large degree because, well, what else is there?

[yudantaiteki]

Also, to me we're talking about two different kinds of luck when we talk about (a) a situation where the rules of the game intentionally put luck into the game via coin flips, dice rolls, e.g. and (b) luck via external things like "I might accidentally come across a great strategy" or "the grandmaster might be sleepy".

The rules of chess have no luck built into them.  The rules of Dominion do.  Both games can be affected by the external luck of the better player making a mistake or being drunk/sleepy, or the novice accidentally using a good strategy.

In addition, the chess example can be complicated by considering the following two situations:
1. Play a normal game of chess.  At the end, flip a coin, if the coin comes up heads the person who won actually loses instead.
2. Play a normal game of chess.  At the end, roll a 6 sided die.  If the die comes up 1 or 2, the person who won actually loses instead.

These are different situations because in #1 it makes no difference whether you checkmate the other person or not because your chance of winning is 50% regardless.  But in #2, checkmating gives you a 2/3 chance of winning vs. 1/3 for the person who got checkmated.  So in #1, your chess strategy has no effect on the outcome but in #2 it does.  I have no problem saying that #1 has less strategy than chess, while #2 has the same strategy (but with the strategy having less effect on the outcome than in normal chess).

EDIT: I should make it clear that I'm happy with the luck/strategy quotient of Dominion and I don't see the need for a house rule.

EDIT 2: You can add a third situation to come up with luck reducing strategy but not completely eliminating it: Play a normal game of chess, but if you try to take a piece with a Queen or Rook, you have to flip a coin, and if it comes up heads you have to skip your turn instead of taking the piece.  There's still strategy in this game but it's not the same strategy as regular chess.

[Davio]

Enforcing a house rule is trying to make Dominion into something that's not.

It's a strategic game with luck.

If you want to take away the luck factor, you will inevitably break other things.

For example let's say you want to diminish first player advantage by giving the first player 3 Curses instead of 3 Estates so second player starts with a 6 VP lead, that's a whole Province, woohoo! Now there will be a lot of games where this doesn't matter at all (Curses and Estates get trashed early) or a lot (no trashing, rush games), so what have you done exactly? Replaced one luck aspect (first player advantage) by another: the layout of the kingdom.

Or you're tired of being on the wrong end of 5/2 starts? Well, suck it up and play another game.
Dominion is made to play under an hour and often under 30 minutes, so you can easily move on to the next game.

[WanderingWinder]

Say for example why do we limit ourselves to identical openings in a tournament?

IIRC wasn't it because the person running the tournament (WW?) decided he liked that rule and basically enforced it upon everyone else?  As far as I am aware there are no official tournament rules to Dominion.

I'm mostly staying away from this discussion, but I'll jump in on this. First, I haven't run any (playing) tournaments (I ran a sim tourney). You are right though, that it's just because whoever has run the tournament has said 'make it so', and they've done this with varying amounts of player input.
This argument doesn't seem like it's a convincing one, because it's not like there's a rule somewhere that says 'in tournament play, use identical start hands), and so of course you can argue 'well people LIKE this and use it', but then you would have to acquiesce to people agreeing to the amount of randomness in the game, inherent in the luck of the draw.

[Piemaster]

I'm mostly staying away from this discussion, but I'll jump in on this. First, I haven't run any (playing) tournaments (I ran a sim tourney). You are right though, that it's just because whoever has run the tournament has said 'make it so', and they've done this with varying amounts of player input.

Sorry about that, I couldn't remember who ran the large tournament in question.  I thought it might have been you, but I knew I could easily have been wrong, hence the question mark.

This argument doesn't seem like it's a convincing one, because it's not like there's a rule somewhere that says 'in tournament play, use identical start hands), and so of course you can argue 'well people LIKE this and use it', but then you would have to acquiesce to people agreeing to the amount of randomness in the game, inherent in the luck of the draw.

That's why I am very much against this as a house rule.  The better players in the tournament will want to enforce it because it takes some of the variance out of the game, which in turn gives them a greater chance of winning.  And because the better players are often the ones with the most clout or held in the highest regard, their opinions will often be given the most weight, even if their suggestion is entirely selfishly motivated.

[DStu]

I ran a sim tourney

... which solved the randomness problem via "play more games". ;P

[Eevee]

I don't think dominion has a lot of variance. In fact, I rather dislike playing the top 5 or so players, because I feel the only way I ever beat them is if some kind of a jack-big money (or something else super simple) is the dominant strategy on board.

Sadly I don't have council room to back me up here, but my feeling is that I, a level 40 player, win way less than 30% of the games against the level 50 guys and that's with me using veto as a tool to make the board as easy as possible (compared to Stef who iirc vetoes random at least against me). So idk maybe I suck, but to me dominion feels really fair and has pretty much the amount of variance I'd like it to have.

[aaron0013]

I highly doubt timchen intended this thread to be an international debate on luck factors, but this is cool

Thanks for the interesting input Donald X.

[greatexpectations]

Sorry about that, I couldn't remember who ran the large tournament in question.  I thought it might have been you, but I knew I could easily have been wrong, hence the question mark.

the tournament in question was likely the 2011 DS.com championships. someone proposed using identical starting hands and theory was on board with it. there was a whole discussion about it in that linked thread. shockingly, some like it, some don't, and most opinions are heavily influenced by how it might affect them in the tournament.

[dondon151]

And it is not uncommon for variants that reduce the impact of luck to be adopted for tournament play, even if no such variant is detailed in the official rules.

[Piemaster]

Sorry about that, I couldn't remember who ran the large tournament in question.  I thought it might have been you, but I knew I could easily have been wrong, hence the question mark.

the tournament in question was likely the 2011 DS.com championships. someone proposed using identical starting hands and theory was on board with it. there was a whole discussion about it in that linked thread. shockingly, some like it, some don't, and most opinions are heavily influenced by how it might affect them in the tournament.

Ah yes that's the one, it was theory not WW (apologies WW).  While there was a 'discussion' about it, the discussion was after the decision had already been made.  I was new to the community at the time so I didn't say anything, but it left something of a bitter taste in my mouth that one player was basically changing the rules of Dominion (after the tournament had started no less) in order to favour the stronger players.

[Davio]

And it is not uncommon for variants that reduce the impact of luck to be adopted for tournament play, even if no such variant is detailed in the official rules.

Well, the thing is, how are you going to compare results?

Comparing results doesn't matter at all for single matches, but a lot for tournament play. Otherwise, it basically invalidates any results even more.

If you're going to break ties based on points (which is a bad idea, but still) and you have one table where players are going heavy Gardens (without rushing) and some Salvager board where one guy keeps the lead by trashing Province for Province every turn, you can't really compare this.

So you want the same set on every table so the results have more meaning. Otherwise you might end up with a "bad" player winning with 3 straight BMU games (because the kingdoms favored it) and a "good" player crafting 3 beautiful engines in the finals.

What I'm saying is that tournaments shouldn't really focus on reducing luck, but on providing comparable results, which often comes down to trying to remove said luck. Can you compare a guy with three 5/2 starts on Mountebank/Witch boards vs. a guy with three 4/3's? Keeping as much as you can the same for all players is the best way I feel.

Of course real life tournaments may have limitations, a limited number of physical sets for instance, so you just have to deal with whatever the organization has conjured up.

[Drab Emordnilap]

I have seen some other version of that speech, probably hitting on some of the same things. I remember in particular him giving "guess the nth digit of pi" as an example of a high luck game that has no random element. Anyway the version I saw was good.

Not to be pedantic, but isn't "Guess the Nth digit of pi" the same measure of skill as "Keep track of the cards remaining in a deck so that you know what the last card is"? Either way, it's only random unless you put the work in to memorize/calculate it, and then it's a skill? Or in the "Guess the Nth" game, are you not told what N is?

[DStu]

I have seen some other version of that speech, probably hitting on some of the same things. I remember in particular him giving "guess the nth digit of pi" as an example of a high luck game that has no random element. Anyway the version I saw was good.

Not to be pedantic, but isn't "Guess the Nth digit of pi" the same measure of skill as "Keep track of the cards remaining in a deck so that you know what the last card is"? Either way, it's only random unless you put the work in to memorize/calculate it, and then it's a skill? Or in the "Guess the Nth" game, are you not told what N is?

It's not about skill, it's about luck. And for example for N=9^(9^(9^(9^9))), good luck with calculating this.  And the problem with luck in Dominion was not about knowing what the last two cards in your shuffle are, but about what they are and that you can't play them twice in two shuffles, no matter how good you know them.

[Drab Emordnilap]

It's not about skill, it's about luck. And for example for N=9^(9^(9^(9^9))), good luck with calculating this.

I would still say that this is a challenge of skill. Just because someone (or anyone ever, obviously) doesn't have the skill necessary for the challenge doesn't mean it's no longer a skill. And sure, you can just guess and then it's purely luck based, but you can just play random squares in Tic-Tac-Toe too, it's still not a game based on luck. I know there's a pretty big difference of scale there, but I feel like it's the same principle.

[DStu]

It's not about skill, it's about luck. And for example for N=9^(9^(9^(9^9))), good luck with calculating this.

I would still say that this is a challenge of skill. Just because someone (or anyone ever, obviously) doesn't have the skill necessary for the challenge doesn't mean it's no longer a skill. And sure, you can just guess and then it's purely luck based, but you can just play random squares in Tic-Tac-Toe too, it's still not a game based on luck. I know there's a pretty big difference of scale there, but I feel like it's the same principle.

We can go back and play "Who can name the higher number?", and construct a number which by construction is not computable by the (assumed) computational capacity of the universe. Take this number as N, and name the Nth digit of Pi.  Of course, given your skill to compute this number exceeds the one of the universe, you could solve this, but for any less extreme notion of "skill" the best you can do is a random guess.

Edit: When it exceeds the (assumed) laws of nature, it somewhere stops to be a skill for me.  Can as well say it is a skill to know which number the dice that is thrown now will show, just because the (assumed) laws of nature don't let us view in the future doesn't make it a random guess.

[rrenaud]

I think you are basically recovering the notion of bounded rationality.

http://en.wikipedia.org/wiki/Bounded_rationality

[DStu]

I think you are basically recovering the notion of bounded rationality.

http://en.wikipedia.org/wiki/Bounded_rationality

Yepp, sounds like it. But especially I propose to call everything that is not covered by ths bounded rationality "random".  Or at least "luck".

[dondon151]

Well, the thing is, how are you going to compare results?

Comparing results doesn't matter at all for single matches, but a lot for tournament play. Otherwise, it basically invalidates any results even more.

I'm not sure how this is relevant. You don't need to compare skill levels between different matches. Tournaments are bad at that anyway with poor or non-existent seeding.

[Donald X.]

Not to be pedantic, but isn't "Guess the Nth digit of pi" the same measure of skill as "Keep track of the cards remaining in a deck so that you know what the last card is"? Either way, it's only random unless you put the work in to memorize/calculate it, and then it's a skill? Or in the "Guess the Nth" game, are you not told what N is?

Guess the Nth digit of pi, for large N, is always going to end up with a completely random winner. You won't possibly have memorized the digit you need, and there's no way to calculate it fast enough. If the winner is completely random, most people would call that "luck." This is utterly different from counting cards, which many people can manage.

The entire point of the "guess the Nth digit of pi" game, in Richard's speech, is in fact, "here is an example of something that's luck that you might not have considered."

[Drab Emordnilap]

This is utterly different from counting cards, which many people can manage.

I feel like it's only a difference of scale, though. You can't compare counting a deck of 52 cards to knowing the 1,002nd digit of pi. If you add enough decks of cards, you hit the same issue; neither you nor I have the calculatory capacity to derive the answer, even though it's fully mathematically derivable.

I think we're just defining words differently, though; I don't look at "I can't figure out the Nth digit of pi fast enough" as me being unlucky; I look at it as being unskillful in that particular aptitude test. The new test that I'm giving myself (Guess a digit) is completely luck, but I don't think that makes the original question a luck-based one.

[Toskk]

Personally, I try to split the luck-based factors into 'kingdom selection factors' and then 'game design factors'. Ultimately, Dominion has a fair amount of luck/randomness to it, and addressing all of those factors would totally remake the game (it wouldn't be Dominion anymore, and it wouldn't be a deck-building game with discrete hands, either). That aside, I feel like there are a few definite ways to minimize the effect of luck/randomness on how the game plays out by artificially selecting kingdoms. The first and biggest one, in my opinion, is the 5/2 vs 4/3 split opening strength difference in some kingdoms. For example a Witch/Chapel opening vs. a Silver/Chapel opening.. that's very likely going to go to the player opening 5/2, *especially* if they're in first seat anyway. I've tried a variety of methods for minimizing this with kingdom selection, but the one I ultimately implemented was to grab the councilroom.com best-and-worst opening buys database (while it was available), and to use it to determine the relative strength of each possible opening. The system implemented in my card picker program takes that data and will reject any kingdom where the strength difference between the best 5/2 opening and the best 4/3 opening is outside of 1.5 levels. Yes, it does rely on a whole lot of averages (i.e. councilroom.com data), and doesn't work at all for Dark Ages (no data). But it seemed to work reasonably well for rejecting the really really broken kingdom sets out there where the 5/2 opening absolutely dominates the game.

[Guy Srinivasan]

Luck versus skill, I love that topic. I hadn't seen the comparison between chess and tic-tac-toe before, that's quite a nice illustration that chess has luck of the sort that Garfield is talking about. Thanks for the pointer!

[Toskk]

The kind of 'luck' Garfield is talking about isn't really what people here are talking about, though. I think a better continuum line would be 'control'.. i.e. how much control does the player have over the outcome of game events by the 'decisions' they make? You have everything from activities with zero control (e.g. Chutes and Ladders), to activities with complete control (e.g. Chess). Basically, what's the correlation factor between making the 'correct' decisions throughout the game and winning the game. In chutes and ladders, players make zero decisions. In Chess, making the 'correct' decisions is 100% correlated with winning the game.

[GigaKnight]

This is utterly different from counting cards, which many people can manage.

I feel like it's only a difference of scale, though. You can't compare counting a deck of 52 cards to knowing the 1,002nd digit of pi. If you add enough decks of cards, you hit the same issue; neither you nor I have the calculatory capacity to derive the answer, even though it's fully mathematically derivable.

I think we're just defining words differently, though; I don't look at "I can't figure out the Nth digit of pi fast enough" as me being unlucky; I look at it as being unskillful in that particular aptitude test. The new test that I'm giving myself (Guess a digit) is completely luck, but I don't think that makes the original question a luck-based one.

Rather than say the original question is a luck-based one, I'd say it's at a specific place on a continuum of luck.  I'd also argue that, as you consider all factors affecting a game, you quickly arrive at the conclusion that literally everything has some element of luck (even if it's a lame definition like "were you lucky enough to have the aptitude for the skill this game requires").

Looking at it another way, you have to separate calculating pi from the hypothetical game itself.  The core task of calculating pi digits is mechanical.  But the game "Guess the Nth digit of pi" puts constraints on that such that you cannot predict the outcome (which is Garfield's definition of luck).