Messages

There should be a thing to click on to see all of the cards being used in the game - full card images with both art and text. This seems straightforward, like a trivial task.

Do you think this is doable with Travelers/Castles/Knights in the same Kingdom? In this case this would be 23 different cards from 3 Kingdom supply pile alone.

No, I meant the 10-11 kingdom cards currently buyable. The other cards can be seen by clicking on the piles as in the other part of what I said. "I do not think you have to be able to see all kingdom cards and also all Castles etc. on the same one screen."

The problem is: Not everybody will get his dream solution. We need to find a solution that is doable and has the most "bang for the buck", where the buck is the limited amount of time that can be spent on this feature.

You see your feature as good for many players, I see my feature as better for many new players. Of course we should discuss that and hopefully get a solution that is even better that what we individually proposed.

I can break this tie.

There should be a thing to click on to see all of the cards being used in the game - full card images with both art and text. This seems straightforward, like a trivial task.

You should also be able to click on an individual pile to see a full version of that card. If that pile is two different cards, or 8 or 10, you should instead see all however many - you click on the pile, instead of showing you 1 card, oops it was Patrician / Emporium so it shows you both. If the pile has a card paired with it that isn't in the pile - the Bane (even though it's a kingdom card also), Madman, Spoils, Prizes, etc. - it should show the main card and also that other card or sets of cards.

I do not think you have to be able to see all kingdom cards and also all Castles etc. on the same one screen.

[So does the opponent know the bot has stepped in?]

Generally I agree with your idea, but I believe that it is important for the system to have different standards for beginners and experienced players. Beginners may use more time for making decisions before they will be labelled as "super slow". Players with higher ratings should be able to make decisions in a shorter time, so they are subject to a stricter standard.

New players are already separated out by rating. So this situation is just, experienced player with a new player's rating vs. a new player.

I dunno, it has to be that new players can get in games, but of course you can get faster by playing against bots, and if someone doesn't want to play against a slow new player, it seems unfriendly to make them.

(But additionally, maybe it would be good to allow friends or players to set a different time limitation on a case-by-case basis.)

I think obv. a game where you picked your opponent has no need for any time limit. Similarly games (started) against bots don't need time limits.

1. I play a game with human player.
2. After some turns, I am very sure that I will win. I click the button and let the bot do the calculation.
3. The bot also think that I will win. I can now close the game or do something else because the bot will now replace me. The bot decides what to do, what cards to play and buy, etc. (So does the opponent know the bot has stepped in?) Goto 5.
or
4. The bot is unsure whether I will win. I have to continue to play the game. For the rest of the game, I cannot click the button again.
5. The game ends and the result is announced.

Yes that's it. You can go play another game even. I expect we tell the opponent; they will know when the bot gets a turn, but uh I don't know. We want to be nice; if it's not a troll I'd rather be honest with them. I don't get much out of not telling the troll.

Again this is a thought experiment for the moment and also would utterly depend on actual good bots.

My general plan for a good bot is to constantly be simulating the rest of the game (rather than using heuristics); I am not worried about a mere 100 rest-of-the-games being simulated, especially when we stop after any game isn't won by the right player, and expect the game to be nearly over.

The bot would not try to imitate either human in any way. It would just play out the game like a bot. Presumably in the troll situation, the troll normally quits after that, as there's no point to sticking around. In a non-troll situation, the other player sees whatever more of the game they wanted to (that they can before the game ends). Again I don't think there's any way to actually identify them as a troll; this would be possible against non-trolls. And it's hard to guess what people would think of that.

I was going to post in the old thread but it vanished. Here is a new thread focused on the important part. No complaining about specific games or players here please, just suggestions for how to make the world a better place, and discussions of suggestions.

* Time-out rule *

I think that an honest, reasonable player should never be in a situation where they have to play a Copper to extend how much time they have to think. I mean that's obviously bad. If there's a time-out at all, it shouldn't be based on when you last played a card. I think I would base it on how much total time the program has spent waiting for you compared to your opponent, with a threshold. For example, just an example, maybe you get to take twice as much time as your opponent, or 10 minutes, whichever is more. Probably with a warning before the first time it gives them the chance to eject you.

Obv. existing game logs could be looked at to see how often a candidate rule would kick in, and tweak it to be something that mostly would not happen except when slow-rolled or that phone call lasts a while. If existing game logs don't store enough time information to do that, they could be changed to store the extra information, and then data could accumulate. There are a lot of games per day, so, the next day you've got your data and can pick the best rule.

Both players get to stare at the board at once on turn one, but it's someone's turn. The rule should be generous enough that typically no-one worries about that. You could potentially try to special-case this - players click "ready" at the start, and have 3 minutes to do it (or are automatically "ready"), and this time doesn't count for anyone. But in the end some people will want to stare for a while and sometimes they will go first. It's nicer not to have to click "ready" at the start. I dunno.

* Matchmaking *

There is a separate but related issue of just not wanting slow opponents, even non-slow-rolling ones. I think that's reasonable. I would add speed as a matchmaking criteria (or, a thing reported to you when a match is made, that you have to okay to actually play the match). You could split it up into speed for the first and last two turns, and speed for other turns.

I think people who are not super slow will appreciate not being paired against super slow players. I think super slow players are not entitled to make opponents suffer through slow games or resign. Net happiness goes up if you can just avoid the match. If there are too many matchmaking settings, it may take longer to find matches; that seems moot here. If all you can get is pain, you can accept it or be happy not to experience it. If you are not getting matched you can lower your standards; if you see an offer and it looks bad you can decide to do it or not.

* Bots step in *

I don't think there's any way to stop someone from creating a new account, playing what appears to be a normal game, and then, if they're losing, waiting out the clock to give the finger to their opponent. Or, being a seemingly reasonable player for a while and then one day losing their mind and letting the clock run out on a late turn.

I do have a thing to try here though, or at least, a thought experiment. Once there are good bots, it could be that you could let a bot take over for you, and get the win (or loss), provided that it looks like you have the game locked up. Bots do not mind getting slow-rolled, they can sit there and wait patiently.

We can see if you're winning by having you click a button, "I think I've locked this one up," and the program simulates 100 games from this point on, with all bots (again, this requires that the bots be good). If you win all 100 then it lets you replace yourself with a bot and count the result for yourself (and if not the button is gone for this game, you just get one shot at it). However then you could just have bots finish all your locked-in wins for you and uh I dunno. Maybe it's okay - if they haven't resigned because they want to see what their deck does, well they can see that vs. a bot. I dunno what people would think of it though, and if they didn't like it, well it's impossible to test "is someone slow-rolling you" in order to confine it to that situation. So. A thing to think about.

And does Donald count as a "popular culture figure?"

I'm just a red link in wikipedia (well, the English one).

So the same would apply to Artisan - but it wouldn't work with Trader, because "would gain" is before I actually have Trader in hand?

"Would-gain" is before you gain the card and so you don't have it yet and can't reveal it.

When I use Transmogrify to gain a Watchtower to my hand, I get the option to reveal this newly gained WT to trash/topdeck itself. In game 4836503 I could actually trash it.

Not a bug. Transmogrify gains the card directly to your hand (rather than gaining it to your discard pile and then moving it; note the program has the new phrasing that hasn't made it to print yet). Watchtower is used directly after gaining a card (thus it's already in your hand), and can move the card from wherever it was gained to.

Game #4711125 amoffett11 turn 10

amoffett plays Dame Sylvia, calls RC on it, it is then trashed by another knight, and he was still able to call a second RC to play it again despite it no longer being in play.

You are correct that this is a bug.

Royal Carriage: When you play this, you get +1 Action and put it on your Tavern mat. It stays on your mat until you call it, directly after resolving a played Action card that is still in play. Royal Carriage cannot respond to Actions that are no longer in play, such as a Reserve card that was put on the Tavern mat, or a card that trashed itself (like a Raze used to trash itself). When called, Royal Carriage causes you to replay the card you just played. You can call multiple Royal Carriages to replay the same Action multiple times (provided the Action is still in play). You completely resolve the Action before deciding whether or not to use Royal Carriage on it. If you use Royal Carriage to replay a Duration card, Royal Carriage will stay in play until the Duration card is discarded from play, to track the fact that the Duration card has been played twice.

How about this:

* On the first of each month, ban a small (3-10) list of cards from all rated games.
* The size of the list and the particular cards banned would change each month.
* The ban list would be chosen by DonaldX each month based on whatever criteria or theme he favors at the time.  If some months he's too busy to make a list, Stef would do it.
* As a matter of policy, no card would be banned in consecutive months.
* The ban list could be hard (card never appears) or soft (card only appears 20% as often as normal cards.

I am not seeing the beauty of this plan. If someone hates Possession, they don't magically unhate it for alternating month-long periods. They'd be happier never seeing it.

If you want to propose an algorithm to auto-deny some undo requests, feel free to do so. But it has to be a pretty concrete algorithm before I can even consider implementing it... not just "a limit on what undos can do". And my gut feeling is that defining a good algorithm here is an impossible task.

As jsh noted, the key thing is new information. IRL if someone has no new information I always let them undo, and if they have new information I rarely do. Sometimes they are all, oh I was going to play Village first, I had it in my hand, and well I'm not made of stone.

I think it's fine if the system is made of stone, for rated games. I would automatically allow undos with no new information, and automatically reject other ones. For unrated games I would let you ask for more generosity like you can now.

The stingy version of "no new information" is, you haven't accessed your deck or uncovered a Knight/Ruins, and the other player hasn't done anything or been given the option to do something. The generous more-programming version is the same except you track how many known cards there are for each player on each deck, and accessing those doesn't count against you.

I confirm that the player should gain a Curse if they have no card to reveal.

After reading your posts, though, I'm actually unclear now on what your stance is with regard to the leaderboard. I think everyone agrees that the guy who resigned should lose his matchups against the other two players who didn't resign, but what about the matchup between the two remaining players? What leaderboard result should be recorded for that matchup? Some options I've seen suggested are:

1. The matchup is completely discarded (this is what I think should happen)

I would record it as a win for each remaining player over the resigning player, and for the two remaining players, not record a result.

When a player out of contention swings the game, well that's what you get for playing with 3 players (which is not me knocking 3-player games; it's my preferred number of players). When a replacement bot swings the game, that's what you get for... that guy resigning. It's bad to feel punished for that.

Regarding ratings, I can absolutely see an argument for allowing the complete game outcome after replacement with a bot count. Currently, the bot is very likely to buy Provinces/points when able pushing the game to a conclusion. It's not unreasonable to expect far behind players in multiplayer games to do this to speed things along, too. Would ignoring the final outcome of games with resignations that are finished with a bot give better rankings than not counting it? I don't think that's an obvious "yes", at all.

The idea isn't to do the best job of ranking the players in this situation; it's to make the players the happiest they can be.

Could it possible if you included the cards from first edition Intrigue and First edition Dominion for gold subscription players.

It is unlikely to ever happen, as I personally don't want it to. Player demand would have to be significant.

I don't mind that you still play with them IRL, but to me they're just inferior cards that made it out because I didn't know any better. I don't think the nostalgia value is enough to actually put them into games.

Some people may want to say "3 is not enough" or "3 is too many". Well, they are free to express that opinion in their personal settings. If they get to play an opponent who agrees with them, we will use the shorter/longer list.

To me "preferred number of respected disliked cards" just seems like extra complexity that isn't getting us much. People who think N is too many can just not ban cards; they are getting the same experience without the extra setting, since their opponent can still ban N. People who think N is not enough, well I disagree with them, I think if you ban more cards eventually you are gaming the system and we want to avoid people feeling like the system is being gamed, even in games they aren't playing. You know, two other people think sure ban 30 cards, and play against each other, and you are missing out because you didn't do that, and you don't want to do it because it's less fun and yet hate that their games count as fair games. And it's not like anyone actually hates 30 cards and yet still plays Dominion.

So I would make N say 5, and then it's so simple, you can ban 5 cards. We can always see, do people say "I hate that I can't ban 10 cards" or what. And this is just to make rated games more fun, there is always whatever you want in unrated games.