Print

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

[Donald X.]

I think we mostly agree. I was aiming for "if you manually create the board, it's automatically an unrated game" because that's very easy to remember and understand.

Part of my bias is also that I don't see unrated games as being such a penalty.

Well that may be so, it's hard for me to evaluate personally. Okay let's take the case of, you just bought Guilds, you want to play with Guilds. Everyone was like that when Guilds came out. So they all force 4 Guilds and none of the games are rated and I would think plenty of people wouldn't like that. Whereas if today I feel like playing with Develop and that's automatically unrated that doesn't seem so bad.

So, maybe "pick a card" forces you to be unrated, but "pick an expansion for this slot," I bet people would prefer to be able to play rated games of that.

I guess we'll have to agree to disagree that repeatedly playing the same (perfectly valid, non-trap) board a hundred times [is/is not] OK to rate. If you don't see a problem with it, I don't think I'm going to convince you.

Well it gives you an edge, no question. Doing it by forcing 10 cards from a small set seems just so unlikely to be an issue. No-one is thinking, "I want to cheat. I know, I will play all-Cornucopia games with 3 cards banned and get great at those 10 cards." I might force it to be unrated anyway just to close a loophole, but the special-case seems weird/complex. A game is rated unless you pick a card, or want at least N cards from one small set. The non-special-case - if you force an expansion, it's unrated - puts me back to, I just bought Guilds, my games are all unrated, this sucks.

I didn't think of a button for "repeat that last set of 10," I bet some people would like that enough to want to make it not as much trouble as it would otherwise be.

Right, but again, that means that players who pick more than 1 are just waiting around, wondering why nobody is joining their game.

Let's say the option is more like you have it - on the list of possible ways to generate the card lists there's "pick 'em," and it's never what happens for two normal matched players (you have to host to make it happen), but checking it means you are okay with it.

Let me be more clear. What I'm suggesting is the abolishment of "minimum # of sets" setting and that players are matched without taking that variable into consideration at all. If you bought all the sets, well then all the games you play will have access to all the cards because the player with more cards hosts. If you bought nothing, sometimes you get matched to someone with bought cards, sometimes not.

I have to be able to play with my friends. So, I can host a game, call it "all cards," invite someone (who is not actually my friend it turns out), play a game with my cards. They have no cards, I've got everything, hooray for them. This works regardless of how matchmaking works. So... why not be friendly with the matchmaking?

Maybe if the matchmaking is unfriendly, people with expansions still use it, because it's easy and hey they have the expansions. Making Fun might like the idea, like I said, I dunno.

Is that what most people will prefer? I think it's probably a huge pain in the neck to manually modify your settings as your rating goes up/down. Having it automatically slide so that you're playing opponents about at your skill level is arguably the most of the point of having a rating, and I'm not sure why anybody would prefer the manual system.

I have never done any ratings-related stuff so I don't know how it goes, but I would imagine your rating doesn't endlessly change giant amounts? You build it up and then it's actually telling us about how good you are.

The question is, do you want to enforce seeing cards with approximately equal frequency? If not, great. You can just pick two expansions randomly and you'll see each Hinterlands card about half as often as each Alchemy card, and each Dark Ages card with even less frequency, etc.

Uh, whatever, it would be some good way. Maybe we pick Alchemy less often because it's smaller but still take 5 cards from it when we pick it. You don't need to see all cards with precisely equal frequency. This is a particular special mode for a particular experience where set themes are reinforced via having multiple cards from a set; you can get flat random from the pure random option.

EDIT: Sorry, misunderstood. You meant how I took it off of the list of options. Yeah, either the hate list should apply to Pro or it shouldn't. Period. I don't care which way it goes. Obviously it should apply to Casual games or what's the point? Having a setting of "Union of hated cards/Intersection of hated cards" is just another needless option that complicates the automatch interface and makes it harder to find games.

But, whatever modes it applies to, "intersection of hated cards" offends no-one except with regards to leaderboard accuracy; union of hated cards also may mean not getting to play with cards you like that lots of people don't. In casual, I may be willing to let my opponent veto cards that I don't veto myself; I may not.

I understand that it's a feature. I had assumed that the feature was primarily there for groups of people who knew each other in real life to jump online and play with the cards without everybody needing to buy them. If it was actually intended to (also) enable about half the random-match players to not buy into the system, well color me surprised.

Well I can't ask Ted what he intended. To me it's mimicking the situation IRL where only one player needs to buy a game. You can bring Dominion to a public game night and play against random people who don't own it. The online game can be greedier but I can let Making Fun worry about that.

[jl8e]

I understand that it's a feature. I had assumed that the feature was primarily there for groups of people who knew each other in real life to jump online and play with the cards without everybody needing to buy them. If it was actually intended to (also) enable about half the random-match players to not buy into the system, well color me surprised.

Well I can't ask Ted what he intended. To me it's mimicking the situation IRL where only one player needs to buy a game. You can bring Dominion to a public game night and play against random people who don't own it. The online game can be greedier but I can let Making Fun worry about that.

Freeloading will help get online-only players to buy cards. It lets them see what they’re missing.

[LastFootnote]

Okay let's take the case of, you just bought Guilds, you want to play with Guilds. Everyone was like that when Guilds came out. So they all force 4 Guilds and none of the games are rated and I would think plenty of people wouldn't like that. Whereas if today I feel like playing with Develop and that's automatically unrated that doesn't seem so bad.

So, maybe "pick a card" forces you to be unrated, but "pick an expansion for this slot," I bet people would prefer to be able to play rated games of that.

In my original proposal, the idea is that when you choose "Classic" mode (equal cards from [up to] 2 sets), you'd be able to force one or both expansions if you chose and still have it ranked casual. This takes care of "I just bought Guilds and want to play it" and has the further advantage that you can match together multiple such requests that don't conflict. Like one player says, "half Guilds", and another says, "half Hinterlands" (or even "half Guilds and half Hinterlands"). You can match those two together as long as one of them owns both. That's complex, but it's all behind-the-scenes complexity.

If you want to be able to have matches ranked even if they're a specifically hosted game that says, "3 from this set, 2 from this set, 5 from this set", that's fine. Just make sure that the stipulations about what can and cannot be ranked are clear and relatively easy to understand.

Let's say the option is more like you have it - on the list of possible ways to generate the card lists there's "pick 'em," and it's never what happens for two normal matched players (you have to host to make it happen), but checking it means you are okay with it.

I think lumping that together with "suggested sets of 10", like you had it earlier, is probably fine. In retrospect, I only split them out because I thought that was the way you'd reach the hosting screen. But I guess you'd first click either "host" or "automatch" and then you set automatch parameters.

Maybe if the matchmaking is unfriendly, people with expansions still use it, because it's easy and hey they have the expansions. Making Fun might like the idea, like I said, I dunno.

I guess I don't consider ownership-agnostic matchmaking to be "unfriendly". It's not like it's malevolently matching up haves with haves and have-nots with have-nots. It's just matching based on rank or whatever other variables. But if Making Fun would prefer it that you can specify number of sets in automatch, that's fine.

Is that what most people will prefer? I think it's probably a huge pain in the neck to manually modify your settings as your rating goes up/down. Having it automatically slide so that you're playing opponents about at your skill level is arguably the most of the point of having a rating, and I'm not sure why anybody would prefer the manual system.

I have never done any ratings-related stuff so I don't know how it goes, but I would imagine your rating doesn't endlessly change giant amounts? You build it up and then it's actually telling us about how good you are.

I suppose it's fine to make it manual if you display your current rating prominently in the automatch window so that you can set it easily. I still think at least being able to set it as a range and forget it is preferable. I'm guessing most player here would agree, but I can't speak for them or for non-f.DS players.

The advantage of having it scale automatically and default to some reasonable range like ±1000 is that players never even have to touch it and their opponents' skills will automatically adjust to them.

The question is, do you want to enforce seeing cards with approximately equal frequency? If not, great. You can just pick two expansions randomly and you'll see each Hinterlands card about half as often as each Alchemy card, and each Dark Ages card with even less frequency, etc.

Uh, whatever, it would be some good way. Maybe we pick Alchemy less often because it's smaller but still take 5 cards from it when we pick it. You don't need to see all cards with precisely equal frequency. This is a particular special mode for a particular experience where set themes are reinforced via having multiple cards from a set; you can get flat random from the pure random option.

Please do not wave your hand and say, "It'll be some good way." My whole point is that there likely isn't "some good way". Take it from a guy who has spent a fair bit of time on this exact problem of trying to pick exactly two sets from an arbitrary pool. The "good way" is to occasionally allow games that are all one large set. What's the problem with that?

But, whatever modes it applies to, "intersection of hated cards" offends no-one except with regards to leaderboard accuracy; union of hated cards also may mean not getting to play with cards you like that lots of people don't. In casual, I may be willing to let my opponent veto cards that I don't veto myself; I may not.

I'm sorry, I was under the impression that you weren't concerned about the people who wanted to see oft-hated cards the usual amount. Again, I am just trying to eliminate options from the screen here. Because the more preferences you can stipulate, the less easy it is to find matches. You could argue that it's the player's responsibility not to shoot himself in the foot this way, but it's nice if you don't hand them the gun. I still think the easiest way to do hate-list is have it be the union in Casual and either the intersection or not used at all in Pro. That way you can leave it off the automatch options screen.

[Donald X.]

In my original proposal, the idea is that when you choose "Classic" mode (equal cards from [up to] 2 sets), you'd be able to force one or both expansions if you chose and still have it ranked casual. This takes care of "I just bought Guilds and want to play it" and has the further advantage that you can match together multiple such requests that don't conflict. Like one player says, "half Guilds", and another says, "half Hinterlands" (or even "half Guilds and half Hinterlands"). You can match those two together as long as one of them owns both. That's complex, but it's all behind-the-scenes complexity.

I see. Well it's more options in this window - optionally pick the expansions for "classic." That does handle it though, I think it's more likely that someone wants to play with their new set than it is that they want to pick some specific weird cocktail. OTOH I don't need to punish cocktails for casual; I remain unscared of the worst case for picking expansions for slots.

I guess I don't consider ownership-agnostic matchmaking to be "unfriendly". It's not like it's malevolently matching up haves with haves and have-nots with have-nots.

If matchmaking is appealing for haves over hosting, then the new system would reduce how often have-nots see expansions vs. the current system. The matchmaking could just give you what you get now, but in this scenario would not. That is the unfriendly thing - taking something away (via how the system gets used).

The advantage of having it scale automatically and default to some reasonable range like ±1000 is that players never even have to touch it and their opponents' skills will automatically adjust to them.

I think/agree that for most players the best thing would just be an on/off button for "match me with someone +/- 1000" (or whatever is good). Obv. some players would want more control, and then the question is, how bad is it for them not to have it, what % of players is that.

Please do not wave your hand and say, "It'll be some good way." My whole point is that there likely isn't "some good way".

Your point is that there's no good way to get the cards to show up precisely even amounts when picking to play 5 cards each from two expansions. I don't value that goal though, I do not remotely need them to show up precisely even amounts.

I don't want it to be "sometimes it's all one set" because that is not "5 cards each from two expansions." The mode becomes "sometimes 5 cards each from two sets, sometimes 10 cards from one set."

And if you want 10 cards from one set, it's there for you, via hosting.

I'm sorry, I was under the impression that you weren't concerned about the people who wanted to see oft-hated cards the usual amount.

I am concerned about everybody, and am there for them.

I still think the easiest way to do hate-list is have it be the union in Casual and either the intersection or not used at all in Pro. That way you can leave it off the automatch options screen.

It still seems like some pros would actually hate the intersection, would feel like other pros either had an unfair advantage over them in games they weren't in (if they didn't hate-list stuff) or else like they weren't getting to play with all the cards (if they did). So uh dunno there.

The intersection for casual wouldn't need to be an option, because you could just not hate-list. The union for casual, it sounds reasonable but I dunno.

If the system had the intersection for casual and nothing for pro, I think the only complaint could be "I want more veto-ing than this dammit." It's a straight improvement over the current "no-vetoing at all" system. If someone is going to quit any game with Tournament though, we are confining them to hosting games, picking the expansions, and not buying Cornucopia. If casual has either the union or it's an option, we are confining them to casual but maybe they can feel like "if you want to be a pro you have to be able to stomach Tournament." It's a harsh world out there.

So then, if union is mandatory for casual and pro has no hate-list, then if you want a hate-list you play casual and if you want no hate-list you play pro, those options are there for you without an uh option. Because you hate Tournament, you never get to play with Possession; Possession will be there for you if you can stomach some Tournaments. I dunno it doesn't seem so unreasonable.

The intersection is very friendly. If pros mostly liked it then for sure I would have it there. Again there are the noted potential problems. I don't know how many pros care that much but obv. a nonzero number. For casual, the point of an option rather than always-union would be, to see hated cards more often, which would include your own. Maybe it's not that great.

Any hate-list I think should be small. If it were intersection only then man it could be as many cards as you wanted. Once it's union I don't want it to be, "wow I hate Possession but if I play casual I will never see a single attack."

[LastFootnote]

I think/agree that for most players the best thing would just be an on/off button for "match me with someone +/- 1000" (or whatever is good). Obv. some players would want more control, and then the question is, how bad is it for them not to have it, what % of players is that.

That sounds good. Default option is "close to my rating", but you can set the min and max manually if you want to.

I don't want it to be "sometimes it's all one set" because that is not "5 cards each from two expansions." The mode becomes "sometimes 5 cards each from two sets, sometimes 10 cards from one set."

Yes, that is why I gave it a name, "Classic Mode". It sounds better than "sometimes 5 cards each from two sets, sometimes 10 cards from one set" or "Playtesting-style Mode". Heck, even "5 cards each from two expansions" isn't real concise and what's more it's still misleading if you're sprinkling promos in there.

Your point is that there's no good way to get the cards to show up precisely even amounts when picking to play 5 cards each from two expansions. I don't value that goal though, I do not remotely need them to show up precisely even amounts.

"Precisely even amounts" is not crucial. If you always force exactly 2 sets, you can't even get "approximately even amounts" for players that e.g. just own one small set. You yourself said, "it'll just pick Alchemy less often but still have 5 cards." But if you JUST have Alchemy, you can't pick it less. All your games are Base/Alchemy and you see your Alchemy cards twice as often as your Base cards. But if you allow 1/3 of your games to be Base only and 2/3 of them to be Base/Alchemy, then bam: you get a pretty even distribution.

Even with a bunch of sets, an algorithm that tries to approximately even this out while maintaining exactly 2 sets is going to be nutty-complex and have a bunch of hard-coded special cases. Or you could just allow the occasional 10-cards-from-one-set game. Even if you own everything, that only happens 1 out of 15 games. And it's not like those games aren't fun to play. 3/8 of those games are Dark Ages only, which is terrific fun! You get to see all those great combos that you will pretty much never have in full random.

I think you and I just see this differently. I'm getting the feeling that you see this mode as "novel thing to try once and then go back to full random". Whereas I want to use it to play all my games, so an approximately even distribution is important to me. If this mode doesn't have that, I'll just never use it.

And if you want 10 cards from one set, it's there for you, via hosting.

Man, all this other stuff is also "there for me via hosting". It all just depends on whether people leave the checkbox for "I'm OK with others choosing the exact table" checked. I have no idea how that will go.

[LibraryAdventurer]

I still think the easiest way to do hate-list is have it be the union in Casual and either the intersection or not used at all in Pro. That way you can leave it off the automatch options screen.

It still seems like some pros would actually hate the intersection, would feel like other pros either had an unfair advantage over them in games they weren't in (if they didn't hate-list stuff) or else like they weren't getting to play with all the cards (if they did). So uh dunno there.

The intersection for casual wouldn't need to be an option, because you could just not hate-list. The union for casual, it sounds reasonable but I dunno.

If the system had the intersection for casual and nothing for pro, I think the only complaint could be "I want more veto-ing than this dammit." It's a straight improvement over the current "no-vetoing at all" system. If someone is going to quit any game with Tournament though, we are confining them to hosting games, picking the expansions, and not buying Cornucopia. If casual has either the union or it's an option, we are confining them to casual but maybe they can feel like "if you want to be a pro you have to be able to stomach Tournament." It's a harsh world out there.

So then, if union is mandatory for casual and pro has no hate-list, then if you want a hate-list you play casual and if you want no hate-list you play pro, those options are there for you without an uh option. Because you hate Tournament, you never get to play with Possession; Possession will be there for you if you can stomach some Tournaments. I dunno it doesn't seem so unreasonable.

Any hate-list I think should be small. If it were intersection only then man it could be as many cards as you wanted. Once it's union I don't want it to be, "wow I hate Possession but if I play casual I will never see a single attack."

Seems to me, it would be good to have the union (or option of union/intersection) of ban-lists for casual and no ban list for pro. I'd be fine with playing only casual as long as I can find people to play with. And AFAIC a three-card ban list sounds just right.
Also, I think different people have different enough tastes that if you use the ban list, you won't be regularly going without any cards that you like (unless you play the same other player or two every time).

[Donald X.]

I think you and I just see this differently. I'm getting the feeling that you see this mode as "novel thing to try once and then go back to full random". Whereas I want to use it to play all my games, so an approximately even distribution is important to me. If this mode doesn't have that, I'll just never use it.

No, I see it as something some people might often do. IRL a significant impetus is not having to cart everything around, but by playing with multiple cards from a set, you get more of a functional theme to your game, and some cards play better. It might seem less intimidating too.

I also like having an option for generating the set-of-10 in some pre-established way that isn't pure random. I wouldn't want two but one seems good. I don't want it to have options, it's quick and easy. Sure it would have some novelty but you might stick with it.

I'm not trying to pick something that maximizes the pleasure of one individual, you or me or whoever. It seems easier to grok as "5 cards from each of two sets," even with a catchy name. It doesn't need to include "maybe 10 from the same set" in the same way it doesn't need "sometimes it's 5 treasures or 5 attacks or 5 whatever."

The case where someone just has base and Alchemy isn't too interesting here. Yes this mode won't be so exciting for them (when they host / provide the sets); they are already always () seeing a random mix from those two sets.

Man, all this other stuff is also "there for me via hosting". It all just depends on whether people leave the checkbox for "I'm OK with others choosing the exact table" checked. I have no idea how that will go.

Again I would flag "they picked a card for the 1st slot" but not "they wanted the 1st slot to be from a particular expansion."

[LastFootnote]

I think you and I just see this differently. I'm getting the feeling that you see this mode as "novel thing to try once and then go back to full random". Whereas I want to use it to play all my games, so an approximately even distribution is important to me. If this mode doesn't have that, I'll just never use it.

No, I see it as something some people might often do. IRL a significant impetus is not having to cart everything around, but by playing with multiple cards from a set, you get more of a functional theme to your game, and some cards play better. It might seem less intimidating too.

I also like having an option for generating the set-of-10 in some pre-established way that isn't pure random. I wouldn't want two but one seems good. I don't want it to have options, it's quick and easy. Sure it would have some novelty but you might stick with it.

I'm not trying to pick something that maximizes the pleasure of one individual, you or me or whoever. It seems easier to grok as "5 cards from each of two sets," even with a catchy name. It doesn't need to include "maybe 10 from the same set" in the same way it doesn't need "sometimes it's 5 treasures or 5 attacks or 5 whatever."

The case where someone just has base and Alchemy isn't too interesting here. Yes this mode won't be so exciting for them (when they host / provide the sets); they are already always () seeing a random mix from those two sets.

The main thrust of the Alchemy example was to show that there isn't an algorithm that can be generalized that picks exactly 2 sets and keeps an approximately equal frequency of cards. But despite the fact that you used to be a computer programmer, you seem to be studiously avoiding considering any specific set-picking method. For some reason.

Fine, how about this as a compromise? 2 sets are picked from an unweighted random list of sets you own. But instead of exactly 5 cards from each, the number of cards is slightly weighted by set size. So if the 2 sets are the same order of magnitude, 5 cards from each. If one is larger than the other, 6 cards from the larger set and 4 from the smaller set. Finally, if one is all of Dark Ages and the other is small (like Guilds or the half of Intrigue you own or whatever), 7 cards from Dark Ages and 3 from the small set.

This way you just label the mode "2 Sets" or something. You get Alchemy flavor as often as any other flavor, just not quite as much of it at a time (unless it gets paired up with another small set). It has the added advantage of allowing promos without violating a "5 cards each" stipulation. Using this method, you still see your cards from small sets a bit more often, but not twice as often, which is fine. See, I am not such an ultra-stickler for equal frequency after all.

Again I would flag "they picked a card for the 1st slot" but not "they wanted the 1st slot to be from a particular expansion."

That works great except that I just realized that I can't add promos to it.

EDIT: Unless "Promo" counts as a set for these purposes. Then I'm all good.

[Donald X.]

The main thrust of the Alchemy example was to show that there isn't an algorithm that can be generalized that picks exactly 2 sets and keeps an approximately equal frequency of cards. But despite the fact that you used to be a computer programmer, you seem to be studiously avoiding considering any specific set-picking method. For some reason.

I haven't devoted any work to it because it's an unimpressive problem and working on it isn't getting me anything. I have games to work on; this is something for a programmer on this project to do. I have zero worries that something good is impossible, and you are not making me write code.

Fine, how about this as a compromise? 2 sets are picked from an unweighted random list of sets you own. But instead of exactly 5 cards from each, the number of cards is slightly weighted by set size. So if the 2 sets are the same order of magnitude, 5 cards from each. If one is larger than the other, 6 cards from the larger set and 4 from the smaller set. Finally, if one is all of Dark Ages and the other is small (like Guilds or the half of Intrigue you own or whatever), 7 cards from Dark Ages and 3 from the small set.

That sounds fine.

EDIT: Unless "Promo" counts as a set for these purposes. Then I'm all good.

Or if promos can show up in any expansion slot. Counting promos as a set seems problematic for people who have fewer than I don't know five of them.

[LastFootnote]

Or if promos can show up in any expansion slot. Counting promos as a set seems problematic for people who have fewer than I don't know five of them.

Yes, I realized that it was problematic for small numbers of Promos a bit after posting. Good call. Your solution of "Promos can show up in any expansion slot" is good, and furthermore it's the obvious solution in retrospect.

The main thrust of the Alchemy example was to show that there isn't an algorithm that can be generalized that picks exactly 2 sets and keeps an approximately equal frequency of cards. But despite the fact that you used to be a computer programmer, you seem to be studiously avoiding considering any specific set-picking method. For some reason.

I haven't devoted any work to it because it's an unimpressive problem and working on it isn't getting me anything. I have games to work on; this is something for a programmer on this project to do. I have zero worries that something good is impossible, and you are not making me write code.

OK, sorry about that. It's just a bit irksome when I say, "I have tried to code this already and it is not as trivial as it first appears," and you're all, "Pfft, whatever. It totally is." Apologies for getting bent out of shape about it.

Although I am consciously aware that Making Fun's programmers are not all the same as Goko's programmers, you will hopefully forgive me if I worry that, when confronted with making a non-trivial algorithm like this, they will botch it. Hence my desire to nail an actual algorithm down.

Fine, how about this as a compromise? 2 sets are picked from an unweighted random list of sets you own. But instead of exactly 5 cards from each, the number of cards is slightly weighted by set size. So if the 2 sets are the same order of magnitude, 5 cards from each. If one is larger than the other, 6 cards from the larger set and 4 from the smaller set. Finally, if one is all of Dark Ages and the other is small (like Guilds or the half of Intrigue you own or whatever), 7 cards from Dark Ages and 3 from the small set.

That sounds fine.

Cool.

Thanks as always for being willing to discuss these things endlessly on the internet. This topic is (obviously) important to me, so I appreciate the dialogue.

[LastFootnote]

So it's time for me to eat crow here. I decided to run the numbers on a chooser that chose exactly two sets with simple weighting for the case of somebody who owns all the sets. By "simple weighting", I mean that you're twice as likely to pick a full-sized set as a small set and three times as likely to pick (all of) Dark Ages than a small set.

To the surprise of nobody except me, the discrepancies in card frequency between this and full-random is very small. You see small-set cards about 4% more often than you would in full random and Dark Ages cards about 4% less. All in all, that's pretty negligible. I thought it was going to be on the order of 15% to 20%, but well estimation was always my weak point. These percentages will be higher the fewer sets you own, but well whatever.

I guess I just fell into the same trap as Wei-Hwa where I came up with this set-choosing algorithm and clearly it's the best thing ever what is everybody else thinking. Sorry for making a federal case out of it. Mea culpa.

So Donald, do you prefer the "5-from-2 with weighted set selection" or the "weighted card ratio with even set selection"? I think they both seem fine.

[Donald X.]

So Donald, do you prefer the "5-from-2 with weighted set selection" or the "weighted card ratio with even set selection"? I think they both seem fine.

Possibly it's more fun to play with 4 Alchemy cards than 5. There's a certain something to keeping it even too though.

The promos could possibly be associated with expansions for this, just treated like cards in a certain expansion. Yes you might own the promos but not the expansions and be using this mode. Something like

Envoy - Intrigue
Black Market - Seaside
Stash - Prosperity
Walled Village - Hinterlands
Governor - Dark Ages

[LastFootnote]

So Donald, do you prefer the "5-from-2 with weighted set selection" or the "weighted card ratio with even set selection"? I think they both seem fine.

Possibly it's more fun to play with 4 Alchemy cards than 5. There's a certain something to keeping it even too though.

The promos could possibly be associated with expansions for this, just treated like cards in a certain expansion. Yes you might own the promos but not the expansions and be using this mode. Something like

Envoy - Intrigue
Black Market - Seaside
Stash - Prosperity
Walled Village - Hinterlands
Governor - Dark Ages

Y'know, I actually slotted the promos into sets for a while, too. Envoy in Intrigue is the obvious choice. I actually put Governor in Prosperity instead of Seaside. Governor has the game-acceleration thing that is nice for Colony games, whereas Stash seems less useful when you're trying to reach $11. I put Stash back in Seaside instead. I put Walled Village in the Base Set, thinking that it would help encourage more interesting decks there. There are a lot of terminal Actions in Base Set and not too many cantrips, and when I use Walled Village it's mostly in that kind of deck so that its special ability shines. Maybe it isn't the best spot for it with Village right there, though. Hinterlands is a nice place for it cost-distribution-wise.

But that's just an anecdote. I think it would honestly be better to just keep promos separate, sprinkling them proportionally into games. It eliminates the "I have this Promo but not the associated set" issue and it's what people will expect.

For what it's worth, I really like that Black Market is a promo instead of in a set. In my perfect world, when you used "2 Sets" mode, Black Market would be one of each card from a large set that wasn't selected, like it was in playtesting. I've started doing this in my real-life games, and I really like how it creates a manageable Black Market deck where everybody knows its contents without having to look through it. It's also convenient that…

• There are never Potion-cost cards in the Black Market deck, which are almost always dead in a game without Potion cards since you're unlikely to buy a Potion just for the Black Market
• Young Witch and Baker are never in the Black Market deck, eliminating some setup.

The fact that Black Market a promo instead of in a large set means that the Black Market deck could be any large set, which is neat.

[Polk5440]

For what it's worth, I really like that Black Market is a promo instead of in a set. In my perfect world, when you used "2 Sets" mode, Black Market would be one of each card from a large set that wasn't selected, like it was in playtesting.

The one time I've really enjoyed playing with Black Market was a few IRL games where I set up the Black Market to have one of every base Dominion card. It really reduced the time people spent reading the cards.

[Donald X.]

I sent a proposal to Jeff last night. I erred on the sides of simplicity over flexibility, less hate over more love. So, pro mode with no hate-list, casual with union 3-card expansions-only hate-list. I didn't mention promos in classic. I strongly suspect that if this thing is actually implemented someday, there will be another chance to discuss specifics. I put quit% on the blocked-list page rather than in matchmaking. And I explained all the terms and the way the lobby works but the overview is:

Matchmaking options (* is default):
- leaderboard: pro / *casual / unrated / don't care [if neither cares, uses casual]
- card selection type: *random, random from two sets, random recommended set, don't care [if neither cares, uses random]
- minimum number of players: *2 / 3 / 4 / 5 / 6
- maximum number of players: 2 / 3 / *4 / 5 / 6
- find an opponent of similar skill (+/-1000 rating): *yes / no
- find opponents from friend-list only: yes / *no
- ignore games where host knows any cards in advance: yes / *no [can't know in pro]

Special options for hosting:
- request particular opponents (they get a window inviting them) and indicate who goes first (or leave unchosen)
- for the 10 card slots, each slot is either: random / random from chosen expansion / specific card
- select a specific recommended set (filled into the 10 slots)
- visible card list: yes / *no

I also sent him the "medium bot" proposal - no PPR, overbuy one terminal, prefer spending all money when buying actions.

[LastFootnote]

Quick follow-up here. I ran the numbers on choosing two sets from an unweighted list and then weighting the number of cards in the game with a static ratio. The best I could get was this:

Small/Small	5 cards from each
Small/Normal	3 cards from Small and 7 cards from Normal
Small/Large	2 cards from Small and 8 cards from Large
Medium/Medium	5 cards from each
Medium/Large	3 cards from Medium and 7 cards from Large

Even with these lopsided pairings, cards in your Small sets will still appear almost 19% more often than they would in full random. Maybe that's not a dealbreaker, it's a matter of opinion. Just thought I'd report on the data.

[Holger]

Quick follow-up here. I ran the numbers on choosing two sets from an unweighted list and then weighting the number of cards in the game with a static ratio. The best I could get was this:

Small/Small	5 cards from each
Small/Normal	3 cards from Small and 7 cards from Normal
Small/Large	2 cards from Small and 8 cards from Large
Medium/Medium	5 cards from each
Medium/Large	3 cards from Medium and 7 cards from Large

Even with these lopsided pairings, cards in your Small sets will still appear almost 19% more often than they would in full random. Maybe that's not a dealbreaker, it's a matter of opinion. Just thought I'd report on the data.

So "Large"=Dark Ages and "Medium"=any other non-small set? If so, Medium/Large should mathematically be 4:6 rather than 3:7, since 25:35 is close to (even slightly larger than) 4:6. The other distributions seem good.
To get a better randomization, you should probably weight the chosen sets slightly (not proportionally) according to their sizes, I think.