Dominion Strategy Forum

Please login or register.

Login with username, password and session length
Pages: 1 ... 5 6 [7] 8 9 10  All

Author Topic: Announcing Dominion Set Generator  (Read 94887 times)

0 Members and 1 Guest are viewing this topic.

WanderingWinder

  • Adventurer
  • ******
  • Offline Offline
  • Posts: 5275
  • ...doesn't really matter to me
  • Respect: +4381
    • View Profile
    • WanderingWinder YouTube Page
Re: Announcing Dominion Set Generator
« Reply #150 on: January 14, 2013, 02:05:14 pm »
0

Which is extremally important especially when those generators will be used for casual play - I look at the table, see which generator was used and can instantly decide if I want to play this game, without having to spend time on analyzing the whole set.

Well, what I had in mind is that while you do get the see what generator was used, in casual play the table generation should try to occur before you decide if you want to play the game.  This not only mimics real life but is what isotropic does:  Here's the kingdom you're going to play with, do you want to play it?

And if you can see the kingdom, it is really that important which generator was used to make it?  (My generator, unlike any other custom generator out there that I've seen, is completely probabilistic and so it is theoretically possible that every set will be generated by it.)
How many custom generators are out there that you've seen that don't fit this description? How widely used are they?

hsiale

  • Duke
  • *****
  • Offline Offline
  • Posts: 383
  • Respect: +244
    • View Profile
Re: Announcing Dominion Set Generator
« Reply #151 on: January 14, 2013, 02:11:16 pm »
0

Well, what I had in mind is that while you do get the see what generator was used, in casual play the table generation should try to occur before you decide if you want to play the game.  This not only mimics real life but is what isotropic does:  Here's the kingdom you're going to play with, do you want to play it?
Of course the set can, and probably should, be generated before the game. But still I prefer to base my decision if I want to join on what generator was used - just because there will be just a few of them so I can look and instantly decide. Analyzing the kingdom itself takes way more time and for more casual players is completely impossible (because they don't know all the cards by heart).
Logged

onigame

  • Thief
  • ****
  • Offline Offline
  • Posts: 96
  • Respect: +73
    • View Profile
    • Dominion Set Generator
Re: Announcing Dominion Set Generator
« Reply #152 on: January 14, 2013, 02:13:17 pm »
+1

You could have had basically a similar good effect with much less bad effect by announcing "this generator will be the default for casual games. Please try it out" instead of just implementing it first, making it the default by surprise, and asking questions later...

A good start when finding out what players want would probably be, well to ask them. Put out a survey (to your target audience, I guess not to us ;)) , or maybe get goko to add a "like/dislike this kingdom" button and data-mine the results of that. I suspect that would be a much better starting point for designing something that casual players want, rather than trying to come up with something super-complicated on your own based on what your hunches are as to what people would find fun.

"Making it the default" should be the *last* step in designing something, not the first one! ;)

While I completely agree with the overall gist of what you're saying, I think you may be making assumptions that doing things the "right" way is easy.  Addressing these assumptions:

* Making my generator the default for casual games on the live Goko site was just an error in judgment that I didn't catch in time.  If I had been consulted on it, I would have been more cautious -- "Let's try it on alpha first, make sure there aren't any bugs, then let's try it on beta and see what our beta users think."  The problem is there's a common (not everywhere, but common) mentality that wants to pushing out new features as fast as possible.   Some of it is due to investor pressure (not me), and some of it is due to (perceived) community pressure.  (Think about it, if your pool of customers is a pool of disgruntled opinionated fans that say "You won't see a dime of my money unless you do X!", you're going to feel that pressure too.)  Had I been more prescient, I would have more explicitly said, "Hey dudes, I know that you feel it's important to get Alchemy running and the iOS standalone app working, but my set generator is not in the same category as that.  I really want to push it out more slowly and get some user feedback."

* This is an offshoot of the aforementioned problem, which is that Goko's Customer Support / User Feedback / QA / Testing, which really should be a team of people considering the rate at which features are getting pushed out, is currently being handled by 1.5 people (one full-timer and one part-timer).  So while yes, it would be nice to have regular user surveys and more feedback UI, it's not something the company has manpower for right now.  I've been coming to the office part-time for the last month and have been doing a lot more QA / Testing than I was expecting to do.  (Feel free to thank me when Alchemy comes out and the rules are correctly implemented when you have Alchemists, Herbalists, Schemes, Hermits, and Outposts in play.  [EDIT: Except that my proposal isn't implemented yet on Alpha, so it's entirely possible that things will still be broken when they release it.  :P ] )

* Putting out a survey to your existing users is tricky.  Putting out a survey to your target audience is *hard*.

* Yes, "like / dislike this kingdom" is important.  I proposed it last month, and it's still in the massive UI re-design proposal that I'm writing.  This is a more general programming thing instead of being Goko-specific: coming up with ideas is easy.  Programming them takes effort.  One of the reasons my set generator went through so quickly is because I provided Goko an actual API:  "Here's my code, here's a sample program that calls it."  Putting aside the question of whether it was the right thing to do, it was technically easy to replace the existing set generator with a call to my code.  Adding a "like / dislike this kingdom" thing, though, would require adding a system for storing that feedback, getting a graphic artist to design what the buttons look like, and so on.
« Last Edit: January 14, 2013, 05:40:42 pm by onigame »
Logged

onigame

  • Thief
  • ****
  • Offline Offline
  • Posts: 96
  • Respect: +73
    • View Profile
    • Dominion Set Generator
Re: Announcing Dominion Set Generator
« Reply #153 on: January 14, 2013, 02:27:18 pm »
0

And if you can see the kingdom, it is really that important which generator was used to make it?  (My generator, unlike any other custom generator out there that I've seen, is completely probabilistic and so it is theoretically possible that every set will be generated by it.)
How many custom generators are out there that you've seen that don't fit this description? How widely used are they?

I will answer the questions you asked, instead of trying to guess at what point you were trying to make (it's not clear to me).

All of them (I think I've seen probably around the neighborhood of 6-10?).  Every custom generator out there usually has some flag such as "have at least 9 Treasure cards" or "have at most 2 cards from Prosperity" or something like that.  My generator doesn't let you do that.  You can say "I'd like to have at most 2 cards from Prosperity", but if you have other likes and dislikes that make it impossible to have at most 2 cards from Prosperity (maybe you also said "I'd like to have at least 9 Treasure cards"), the generator will try to do its best to compromise between your preferences.  You can also indicate how strong your individual preferences are.  You can say "At Least 9 Treasure cards" is more important than "At most 2 cards from Prosperity", or vice versa.

I really enjoy the flexibility of this approach (I do want to see the occasional screwy set, just less frequently than the pure random generator makes them), but I have definitely realized that it is a pain to explain it.  In my experience, most people really like having absolutes when they're in control.  (But as the Bandit Camp example shows, people really hate even near-absolutes when they're not in control.)

I don't have statistics on how widely used other generators are.
Logged

WanderingWinder

  • Adventurer
  • ******
  • Offline Offline
  • Posts: 5275
  • ...doesn't really matter to me
  • Respect: +4381
    • View Profile
    • WanderingWinder YouTube Page
Re: Announcing Dominion Set Generator
« Reply #154 on: January 14, 2013, 02:32:40 pm »
0

And if you can see the kingdom, it is really that important which generator was used to make it?  (My generator, unlike any other custom generator out there that I've seen, is completely probabilistic and so it is theoretically possible that every set will be generated by it.)
How many custom generators are out there that you've seen that don't fit this description? How widely used are they?

I will answer the questions you asked, instead of trying to guess at what point you were trying to make (it's not clear to me).

All of them (I think I've seen probably around the neighborhood of 6-10?).  Every custom generator out there usually has some flag such as "have at least 9 Treasure cards" or "have at most 2 cards from Prosperity" or something like that.  My generator doesn't let you do that.  You can say "I'd like to have at most 2 cards from Prosperity", but if you have other likes and dislikes that make it impossible to have at most 2 cards from Prosperity (maybe you also said "I'd like to have at least 9 Treasure cards"), the generator will try to do its best to compromise between your preferences.  You can also indicate how strong your individual preferences are.  You can say "At Least 9 Treasure cards" is more important than "At most 2 cards from Prosperity", or vice versa.

I really enjoy the flexibility of this approach (I do want to see the occasional screwy set, just less frequently than the pure random generator makes them), but I have definitely realized that it is a pain to explain it.  In my experience, most people really like having absolutes when they're in control.  (But as the Bandit Camp example shows, people really hate even near-absolutes when they're not in control.)

I don't have statistics on how widely used other generators are.
I wasn't trying to make a point. It's just, I think I've only seen like 1 other generator that wasn't just a pure-random or pure-random-from-these-sets-which-I-own. So when you made your comparison, I was like 'man, is that a thing?' I was just curious.



Having said that, are you advocating that your generator is better here because it doesn't LET you have options like 'including at least X cards with attribute Y'? Or are you saying that the other ones FORCE you to have something like that? Because I would think the option to do that is good - more options is always good - but forcing you to have some flag like that would be really weird.

LastFootnote

  • Adventurer
  • ******
  • Offline Offline
  • Posts: 7495
  • Shuffle iT Username: LastFootnote
  • Respect: +10721
    • View Profile
Re: Announcing Dominion Set Generator
« Reply #155 on: January 14, 2013, 02:37:49 pm »
+2

Onigame, please answer me this question: why are you convinced this convoluted algorithm is necessary, especially for casual players? Have you played Dominion with new players a bunch of times using full random and had them complain that too many kingdoms weren't fun? Because I've never had that happen.

If you want it so that new players can get the hang of the game with some good kingdoms, Adventure mode already takes care of that. Hell, the fact that they can only play with the Base Set takes care of that. I've said it before: Dominion has become very popular without your fancy algorithm that reduces variety.

Why are you so convinced that there is a problem and that an algorithm of your design is the solution? You yourself have said that there are a bunch of other things you'd like to be working on. Why is this such a high priority? Every time you post, it seems like you're operating under the assumption that this randomizer is necessary for Goko. Why? Why can't you just tell them, "Hey, turns out this was a bad idea. Nevermind"?

Again, the reason I'm so vehemently against your algorithm is that it only serves to reduce variety. In my eyes, Dominion's main draw for new players is that variety. I'm guessing your algorithm is more likely to make the experience of playing several games less fun for new players, rather than more.
Logged

onigame

  • Thief
  • ****
  • Offline Offline
  • Posts: 96
  • Respect: +73
    • View Profile
    • Dominion Set Generator
Re: Announcing Dominion Set Generator
« Reply #156 on: January 14, 2013, 02:41:44 pm »
0

So if the issue is that a larger card pool creates more crazy Kingdoms, here's my solution: reduce the size of the cardpool before choosing your Kingdom cards.

Doing this is simple. Just choose 2 or 3 expansions that you're going to use and then only choose cards from them. Specifically, here's how my homemade randomizer does it:

Your homemade randomizer is pretty close to how I randomized games when using real randomizer cards.
I certainly fully support having this as an option for Goko, assuming that you don't feel that you have some sort of ownership over the basic concept.

As an independent question, would you have any objections if I replicated your generator's behavior in my set generator?  Not as the "default" settings, but effectively a separate generation option.
Logged

GendoIkari

  • Adventurer
  • ******
  • Offline Offline
  • Posts: 9701
  • Respect: +10741
    • View Profile
Re: Announcing Dominion Set Generator
« Reply #157 on: January 14, 2013, 03:02:21 pm »
0

Mm, that checkbox is supposed to be invisible... care to tell me what browser you're using?

If you don't want that check box to be there, you really should either completely delete it from the published version, or make a change on the server-side to not render it in the first place. Just using attributes to hide it allows users to easily switch it back to visible. I just tried it out and the checkbox is still fully functional.
Logged
Check out my F.DS extension for Chrome! Card links; Dominion icons, and maybe more! http://forum.dominionstrategy.com/index.php?topic=13363.0

Thread for Firefox version:
http://forum.dominionstrategy.com/index.php?topic=16305.0

LastFootnote

  • Adventurer
  • ******
  • Offline Offline
  • Posts: 7495
  • Shuffle iT Username: LastFootnote
  • Respect: +10721
    • View Profile
Re: Announcing Dominion Set Generator
« Reply #158 on: January 14, 2013, 03:07:44 pm »
0

Your homemade randomizer is pretty close to how I randomized games when using real randomizer cards.
I certainly fully support having this as an option for Goko, assuming that you don't feel that you have some sort of ownership over the basic concept.

No, I certainly have no such feeling of ownership.

Quote
As an independent question, would you have any objections if I replicated your generator's behavior in my set generator?  Not as the "default" settings, but effectively a separate generation option.

I was actually considering doing that myself. It may be faster since I'm already familiar with coding the algorithm and the various necessary quirks to take care of edge cases. I'm much more familiar with HTML and JavaScript than I was two years ago when I coded my Java app, or chances are I would have made it HTML in the first place.
Logged

onigame

  • Thief
  • ****
  • Offline Offline
  • Posts: 96
  • Respect: +73
    • View Profile
    • Dominion Set Generator
Re: Announcing Dominion Set Generator
« Reply #159 on: January 14, 2013, 03:38:36 pm »
+2

Onigame, please answer me this question: why are you convinced this convoluted algorithm is necessary, especially for casual players? Have you played Dominion with new players a bunch of times using full random and had them complain that too many kingdoms weren't fun? Because I've never had that happen.

Your actual question is loaded -- it's confusing "new players" with "casual players".

Currently, I refuse to play full random with new players.  I pretty much hated the game after my first five plays or so, and that was just full random with base set.  After I grew to like the game, I initially would play full random with new players, and I noticed a similar reaction.  I have played probably around 3 games with players for whom they had played the game once with experienced players on full random, and felt that they just "didn't get the game" and weren't interested in playing it again.  I convinced them to give it another chance, but with the First Game Starting Kingdom, and generally their feelings towards the game improved.

Casual players I define differently.  Even though I've probably played the game a few hundred
times by now, I consider myself a casual player (although I'm definitely on the edge).  The line that I draw is this: A serious player is looking to better themselves at the game and is willing to spend time improving their skills.  A casual player wants a few dozen minutes of entertainment.  The fun in Dominion, for casual players, is not about learning cute-but-rare interactions, or about making strategic decisions that increase your chances of winning by 1% at a time, but about playing the common interactions that most players who are no longer "new" know about.

I remember a discussion I had, not soon after Intrigue came out, in one of my regular gaming groups.  Some players wanted to veto Saboteur and Torturer, on the grounds that games with them aren't very fun.  David desJardins said something along the lines of, "Why aren't they fun?  The purpose of the game is to win, why does it matter whether the winning score is 30 or 2?"  That was probably the starting point of me noticing this divide.

I think it is likely that casual players are the "silent majority" of Dominion players.  I would say that every single employee here at Goko is a casual player, which is why sometimes have to fight pretty hard about getting card interactions to actually work correctly.  I bet if you put a poll within Goko, asking the question, "Is it important that when-buy and when-gain work correctly?", the votes for "Not really, but we should do it because the fan base wants it" would outnumber the votes for "Yes, because it's the right thing to do."

There was a period in my life when I played on isotropic regularly, maybe about 3 games a day.  At some point I realized that I wasn't enjoying the overall experience.  I hated getting crushed by high-rated players because I didn't want to invest the time into bettering myself.  I kind-of enjoyed crushing lower-rated players but I couldn't get the full enjoyment because I knew there was another human on the other side that was losing.  I hated going through "crazy" kingdoms. 

Now on Goko, I get a much happier play experience.  I play lots of bots and I don't have to feel bad about crushing them.  I still suffer through the "crazy" kingdoms, but now that I actually got off my ass and wrote a generator that decreases them I play them much less.

I really don't feel that my play desires are that uncommon.  I play in a local pinball league; few of the other players there spend time doing card games or board games.  I know of at least three people there who play Androminion on their phone, against a bot.  I've asked them if they would rather play on isotropic, and they say "no, I tried isotropic and all the players there are too good."

Quote
If you want it so that new players can get the hang of the game with some good kingdoms, Adventure mode already takes care of that. Hell, the fact that they can only play with the Base Set takes care of that. I've said it before: Dominion has become very popular without your fancy algorithm that reduces variety.

I'm not denying that it's become very popular without my fancy algorithm.  But if you're trying to make an argument that giving people the option of my fancy algorithm is going to decrease its popularity, it's going to require more than that assertion to convince me.  Dominion is a popular game, but I want to see it reach the popularity of Farmville.

Quote
Why are you so convinced that there is a problem and that an algorithm of your design is the solution? You yourself have said that there are a bunch of other things you'd like to be working on. Why is this such a high priority? Every time you post, it seems like you're operating under the assumption that this randomizer is necessary for Goko. Why? Why can't you just tell them, "Hey, turns out this was a bad idea. Nevermind"?

I can tell them that, but I won't tell them that because I don't believe it.  I do want to get user feedback into the next re-design, so that if it is a bad idea, I can find out.  Then I will tell them that.

Quote
Again, the reason I'm so vehemently against your algorithm is that it only serves to reduce variety. In my eyes, Dominion's main draw for new players is that variety. I'm guessing your algorithm is more likely to make the experience of playing several games less fun for new players, rather than more.

Right, this is a fundamental disagreement between our philosophies.  I believe that Dominion's main draw for players is a certain level of manageable variety, but then any additional variety added on top of that is lost on most players.  I bet the most popular game on Windows right now is Solitaire.  People don't play that because they want variety, they play that game because they know the rules already and it's a fun way to waste some time.  (Did you know Windows Solitaire has six different ways you can play it?  I believe most people don't because they're not looking for variety.)  I want to try to get Dominion into that niche.

(My favorite game, Race for the Galaxy, by the way, totally has more variety than Dominion.  I think one big reason Dominion is more popular than RftG because Dominion has less variety.)
Logged

onigame

  • Thief
  • ****
  • Offline Offline
  • Posts: 96
  • Respect: +73
    • View Profile
    • Dominion Set Generator
Re: Announcing Dominion Set Generator
« Reply #160 on: January 14, 2013, 03:44:21 pm »
+1

confusing "new players" with "casual players".

One thing I forgot to mention -- I had this similar confusion as of, oh, 24 hours ago.  It was one of WanderingWinder's posts on this thread that made me realize that I was mixing the two groups.  The "experience" axis is independent from the "casual-serious" axis -- you can have new players who are serious about learning the game and doing well, just you like can have new players who want a laid-back casual experience.  The two axes are correlated but not absolutely so.  I happen to be the somewhat rarer example of someone who is high on the "experience" axis but on the low end of the "casual-serious" axis.

So thank you, WanderingWinder, for that piece of enlightenment.
Logged

onigame

  • Thief
  • ****
  • Offline Offline
  • Posts: 96
  • Respect: +73
    • View Profile
    • Dominion Set Generator
Re: Announcing Dominion Set Generator
« Reply #161 on: January 14, 2013, 03:47:19 pm »
0

Mm, that checkbox is supposed to be invisible... care to tell me what browser you're using?

If you don't want that check box to be there, you really should either completely delete it from the published version, or make a change on the server-side to not render it in the first place. Just using attributes to hide it allows users to easily switch it back to visible. I just tried it out and the checkbox is still fully functional.

I don't want it there, but it's not important enough to me to prioritize it over other more interesting features I want to add.  I don't mind if people manage to generate some sets with Guilds in them by hacking my generator -- the important data, such as what the names of the cards are and what they do, aren't on the server.
Logged

GendoIkari

  • Adventurer
  • ******
  • Offline Offline
  • Posts: 9701
  • Respect: +10741
    • View Profile
Re: Announcing Dominion Set Generator
« Reply #162 on: January 14, 2013, 03:50:06 pm »
+2

Mm, that checkbox is supposed to be invisible... care to tell me what browser you're using?

If you don't want that check box to be there, you really should either completely delete it from the published version, or make a change on the server-side to not render it in the first place. Just using attributes to hide it allows users to easily switch it back to visible. I just tried it out and the checkbox is still fully functional.

I don't want it there, but it's not important enough to me to prioritize it over other more interesting features I want to add.  I don't mind if people manage to generate some sets with Guilds in them by hacking my generator -- the important data, such as what the names of the cards are and what they do, aren't on the server.

I'll assume you're telling the truth, and stop wasting my time trying to get the program to show me the real names of the cards then!
Logged
Check out my F.DS extension for Chrome! Card links; Dominion icons, and maybe more! http://forum.dominionstrategy.com/index.php?topic=13363.0

Thread for Firefox version:
http://forum.dominionstrategy.com/index.php?topic=16305.0

onigame

  • Thief
  • ****
  • Offline Offline
  • Posts: 96
  • Respect: +73
    • View Profile
    • Dominion Set Generator
Re: Announcing Dominion Set Generator
« Reply #163 on: January 14, 2013, 03:51:38 pm »
0

Quote
As an independent question, would you have any objections if I replicated your generator's behavior in my set generator?  Not as the "default" settings, but effectively a separate generation option.

I was actually considering doing that myself. It may be faster since I'm already familiar with coding the algorithm and the various necessary quirks to take care of edge cases. I'm much more familiar with HTML and JavaScript than I was two years ago when I coded my Java app, or chances are I would have made it HTML in the first place.

That was not an answer to my actual question.  :)  May I have your approval, please?

Independently, if you would like to write your code in Javascript, provide an API with examples, and allow them to use it for free, I would be very enthusiastic to convince Goko to put your generator in as an option.

A longer-range plan would be for Goko to provide an API that would allow any programmer to use their own generator when making kingdoms.  I will put that in as a stretch goal.
Logged

hsiale

  • Duke
  • *****
  • Offline Offline
  • Posts: 383
  • Respect: +244
    • View Profile
Re: Announcing Dominion Set Generator
« Reply #164 on: January 14, 2013, 04:00:07 pm »
+1

Dominion is a popular game, but I want to see it reach the popularity of Farmville.
If you achieve this, great for everyone (Goko has good income and big crowd of casual players means big pool to recruit serious players making the experience of high level playing more enjoyable). But before you go this way, you really, really need to improve your rulebook. You don't make manuals for online games the way Goko has it done now. This is way more important than set generators and I think you should do this soon (as I've heard facebook launch is under preparation and if it's done with current manual it will fail even if all other problems in the game are fixed).
Logged

LastFootnote

  • Adventurer
  • ******
  • Offline Offline
  • Posts: 7495
  • Shuffle iT Username: LastFootnote
  • Respect: +10721
    • View Profile
Re: Announcing Dominion Set Generator
« Reply #165 on: January 14, 2013, 04:11:32 pm »
0

That was not an answer to my actual question.  :)  May I have your approval, please?

Independently, if you would like to write your code in Javascript, provide an API with examples, and allow them to use it for free, I would be very enthusiastic to convince Goko to put your generator in as an option.

Let me rephrase. I don't have a webserver. I would be happy to have you host my algorithm and to provide it to Goko for free. Ideally, I'd like to take a crack at writing it myself, using your generator and your API as a starting point and hosting it on your server, both for my convenience and Goko's. I assume our generators can have an identical output format, for instance, and I believe your card data in dominionSetGeneratorData.js is sufficient for my needs. If this is acceptable, perhaps we can discuss the specifics through email.

Quote
A longer-range plan would be for Goko to provide an API that would allow any programmer to use their own generator when making kingdoms.  I will put that in as a stretch goal.

I had this idea as well. It would be neat, but I wouldn't blame Goko for putting it at the bottom of the priority list.
Logged

ftl

  • Mountebank
  • *****
  • Offline Offline
  • Posts: 2056
  • Shuffle iT Username: ftl
  • Respect: +1345
    • View Profile
Re: Announcing Dominion Set Generator
« Reply #166 on: January 14, 2013, 04:27:41 pm »
+6

"no, I tried isotropic and all the players there are too good."

That's not just casual players - playing mismatched games is no fun no matter what your skill level is.

Which is why I think good matchmaking is important! For both casual and serious and experienced and inexperienced players, you want to match up with people roughly the same level as you are. I don't think that's unique to casual players.
Logged

LastFootnote

  • Adventurer
  • ******
  • Offline Offline
  • Posts: 7495
  • Shuffle iT Username: LastFootnote
  • Respect: +10721
    • View Profile
Re: Announcing Dominion Set Generator
« Reply #167 on: January 14, 2013, 04:36:29 pm »
0

Currently, I refuse to play full random with new players.  I pretty much hated the game after my first five plays or so, and that was just full random with base set.  After I grew to like the game, I initially would play full random with new players, and I noticed a similar reaction.  I have played probably around 3 games with players for whom they had played the game once with experienced players on full random, and felt that they just "didn't get the game" and weren't interested in playing it again.  I convinced them to give it another chance, but with the First Game Starting Kingdom, and generally their feelings towards the game improved.

This may be why I never had this experience. I always start new players with the First Game setup. I'm not surprised that new players thrown into the deep end would have a negative reaction.

Quote
Casual players I define differently.  Even though I've probably played the game a few hundred
times by now, I consider myself a casual player (although I'm definitely on the edge).  The line that I draw is this: A serious player is looking to better themselves at the game and is willing to spend time improving their skills.  A casual player wants a few dozen minutes of entertainment.  The fun in Dominion, for casual players, is not about learning cute-but-rare interactions, or about making strategic decisions that increase your chances of winning by 1% at a time, but about playing the common interactions that most players who are no longer "new" know about.

I remember a discussion I had, not soon after Intrigue came out, in one of my regular gaming groups.  Some players wanted to veto Saboteur and Torturer, on the grounds that games with them aren't very fun.  David desJardins said something along the lines of, "Why aren't they fun?  The purpose of the game is to win, why does it matter whether the winning score is 30 or 2?"  That was probably the starting point of me noticing this divide.

I think it is likely that casual players are the "silent majority" of Dominion players.  I would say that every single employee here at Goko is a casual player, which is why sometimes have to fight pretty hard about getting card interactions to actually work correctly.  I bet if you put a poll within Goko, asking the question, "Is it important that when-buy and when-gain work correctly?", the votes for "Not really, but we should do it because the fan base wants it" would outnumber the votes for "Yes, because it's the right thing to do."

This is all very reasonable, but I'm not convinced your algorithm addresses these issues. If people think games with Attacks are not fun (and I know a few such people), how does your little-bit-of-everything randomizer help that? A card veto list is certainly desirable for casual games. I won't argue that point.

Quote
There was a period in my life when I played on isotropic regularly, maybe about 3 games a day.  At some point I realized that I wasn't enjoying the overall experience.  I hated getting crushed by high-rated players because I didn't want to invest the time into bettering myself.  I kind-of enjoyed crushing lower-rated players but I couldn't get the full enjoyment because I knew there was another human on the other side that was losing.  I hated going through "crazy" kingdoms. 

Now on Goko, I get a much happier play experience.  I play lots of bots and I don't have to feel bad about crushing them.  I still suffer through the "crazy" kingdoms, but now that I actually got off my ass and wrote a generator that decreases them I play them much less.

I really don't feel that my play desires are that uncommon.  I play in a local pinball league; few of the other players there spend time doing card games or board games.  I know of at least three people there who play Androminion on their phone, against a bot.  I've asked them if they would rather play on isotropic, and they say "no, I tried isotropic and all the players there are too good."

Sounds like your bad experiences and those of your friends have a lot more to do with your opponents than with the Kingdoms you played. I agree that Goko needs a good ranking system and a good matching system based on it, but that's a separate issue (and one that deserves a higher priority). This is a problem for any game. I wouldn't want to play chess against a grandmaster other than to say I've done it. Here you only mention crazy kingdoms as a brief aside.

Quote
Quote
If you want it so that new players can get the hang of the game with some good kingdoms, Adventure mode already takes care of that. Hell, the fact that they can only play with the Base Set takes care of that. I've said it before: Dominion has become very popular without your fancy algorithm that reduces variety.

I'm not denying that it's become very popular without my fancy algorithm.  But if you're trying to make an argument that giving people the option of my fancy algorithm is going to decrease its popularity, it's going to require more than that assertion to convince me.  Dominion is a popular game, but I want to see it reach the popularity of Farmville.

Giving them the option is fine as long as:

A) Your algorithm actually provides something that people want.
B) It's not the default.

Yeah, you say you want there not to be a default, but in a UI there is ALWAYS a default. You say you want to make it so a player has to choose which randomizer to use before each game. I ask you, "Which option is the top one?" Because that's the default. That's what new players are going to try first. As long as your algorithm isn't the default (the top option) and has a descriptive name, I'm fine with it being available.

I admit that I'm wrestling to come up with a good name for kingdoms produced by your algorithm. I guess "Balanced Game" might be OK once you've evened out the card biases a little more.

Quote
I believe that Dominion's main draw for players is a certain level of manageable variety, but then any additional variety added on top of that is lost on most players.  I bet the most popular game on Windows right now is Solitaire.  People don't play that because they want variety, they play that game because they know the rules already and it's a fun way to waste some time.  (Did you know Windows Solitaire has six different ways you can play it?  I believe most people don't because they're not looking for variety.)  I want to try to get Dominion into that niche.

The people who don't want variety won't buy more sets, which only serve to increase variety. Those people provide revenue through ads rather than Gokoins. If they don't want variety, they can play the recommended sets of 10 and the Adventures over and over. Nothing wrong with that.
Logged

onigame

  • Thief
  • ****
  • Offline Offline
  • Posts: 96
  • Respect: +73
    • View Profile
    • Dominion Set Generator
Re: Announcing Dominion Set Generator
« Reply #168 on: January 14, 2013, 04:47:19 pm »
0

That's not just casual players - playing mismatched games is no fun no matter what your skill level is.

Which is why I think good matchmaking is important! For both casual and serious and experienced and inexperienced players, you want to match up with people roughly the same level as you are. I don't think that's unique to casual players.

Actually, I enjoy games when playing against people who are ranked much lower than I am.  I am a bit ashamed to admit that it's probably even true that I enjoy them more than playing against people roughly the same level as I am.  When I turn on a game of Civilization (the computer game) I set all the AIs on easy and stomp all over the map.  It's because I don't derive enjoyment from being a better player or from beating a close challenge.  I derive enjoyment from the comfort of exploring a system I know.

I do suspect that serious players are very unlikely to feel this way.  If you're a serious player you want a real challenge and you want to get better at the game, and playing against a significantly weaker player doesn't help.  Also, inexperienced players don't get this option, because there aren't any weaker players around.  So it tends to require the combo of experience and casualness.
Logged

onigame

  • Thief
  • ****
  • Offline Offline
  • Posts: 96
  • Respect: +73
    • View Profile
    • Dominion Set Generator
Re: Announcing Dominion Set Generator
« Reply #169 on: January 14, 2013, 05:05:35 pm »
+3

Yeah, you say you want there not to be a default, but in a UI there is ALWAYS a default. You say you want to make it so a player has to choose which randomizer to use before each game. I ask you, "Which option is the top one?" Because that's the default. That's what new players are going to try first. As long as your algorithm isn't the default (the top option) and has a descriptive name, I'm fine with it being available.

Well, how about this as a proposal: Initially the pure unbiased randomizer will be listed as the top one.  We implement a rating/approval system for kingdoms that players can give after they play a game.  The order of randomizers will be re-sorted based on whichever randomizer has a higher (Bayesian) rating.

That way, in the long run, whichever randomizer, yours, mine, unbiased, generates the better kingdoms, as determined by the players, becomes the UI default for new players who have decided to use a randomizer for the first time.

Would that work for you?
Logged

LastFootnote

  • Adventurer
  • ******
  • Offline Offline
  • Posts: 7495
  • Shuffle iT Username: LastFootnote
  • Respect: +10721
    • View Profile
Re: Announcing Dominion Set Generator
« Reply #170 on: January 14, 2013, 06:04:05 pm »
0

Well, how about this as a proposal: Initially the pure unbiased randomizer will be listed as the top one.  We implement a rating/approval system for kingdoms that players can give after they play a game.  The order of randomizers will be re-sorted based on whichever randomizer has a higher (Bayesian) rating.

That way, in the long run, whichever randomizer, yours, mine, unbiased, generates the better kingdoms, as determined by the players, becomes the UI default for new players who have decided to use a randomizer for the first time.

Would that work for you?

I think that sounds very reasonable. The biggest issue I have is that the buttons shouldn't jump around too much. People don't like it when the options are suddenly in a different order for no explained reason. If one of the randomizers is the runaway winner, this would be a non-issue.

It would be nice to somehow give players Kingdoms generated by a randomly selected randomizer. That would remove some biases and give players an opportunity to rank sets generated by all the randomizers. However, I wouldn't want to force that on people by default and having a 'Random Randomizer' button seems strange. Maybe it'd be OK.

One other minor note is that, if and when my suggested randomizer is available on Goko, it should only be available to players who have purchased at least one expansion. The reason is that if you only own the Base Set, it's identical to unbiased random. The extra option that doesn't change anything would only serve to confuse players.

I think we need some nomenclature for these randomizers and the games they generate. For yours, "Balanced Game" seems most appropriate, assuming that you will/have already gotten it to the point where it successfully creates Kingdoms with no one clear strategy. For the algorithm I use, I suggest the name "Flavorful Game", since it produces games that have concentrations of flavor from just two or three expansions, rather than mixing in everything.

But what's the best name for unbiased random? "Default Game" won't work if it doesn't end up as the default. "Unbiased Game" seems a little technical. "Random Game" isn't specific enough; they're all random. "Pro Game" implies the game is being ranked on the Pro leaderboard, which won't necessarily be the case. Maybe "Basic Game" is the best bet.

Of course, it's also important that players get a short description of each randomizer when they choose it, so they know what to expect.
Logged

Donald X.

  • Dominion Designer
  • *****
  • Offline Offline
  • Posts: 6357
  • Respect: +25671
    • View Profile
Re: Announcing Dominion Set Generator
« Reply #171 on: January 14, 2013, 06:09:22 pm »
+5

Well, how about this as a proposal: Initially the pure unbiased randomizer will be listed as the top one.  We implement a rating/approval system for kingdoms that players can give after they play a game.  The order of randomizers will be re-sorted based on whichever randomizer has a higher (Bayesian) rating.

That way, in the long run, whichever randomizer, yours, mine, unbiased, generates the better kingdoms, as determined by the players, becomes the UI default for new players who have decided to use a randomizer for the first time.
For sure I would not have the order of a list of options change over time, that would obv. suck. You could change which radio button is the default, but only for players who haven't played yet - for people who've ever seen this screen, it should retain what they previously picked. Aside from those things, if for a completely-new-to-Goko player the default radio button is the method most people are upvoting sets for, that's fine. Of course a lot of the time they will just join a game someone else started, without even seeing that list of options.

I got better at making sets after Intrigue. I can believe that people playing the base set by itself might feel like it needs something. There are some duds, and those duds reduce strategic options. The base set has done great anyway, but you know, there's room for improvement. But I mean take out the duds and you're there, that's what I think. If someone plays a random ten from Dark Ages they are not going to see all these boards with nothing to do.

I know there are people out there who don't like Dominion. Like I always say, it didn't win game of the year in Austria. Maybe some of those people just needed better first experiences; I don't have that data. But I have, over the course of thousands of games, taught lots of people the game, and my experience does not match yours. I was not teaching them with the base set though, not ever; I had whatever I was testing and whatever set I was pairing it with and I taught them with that; no time to lose on a game with the first game setup, and I might not even have those cards with me. I think the first game setup is a good first game setup, I recommend it; I am just giving my data here. Pure random from two sets, no lost customers.

If the concern really is, that the game needs to be at its very best for new-to-Dominion players, which seems to be your pitch in that one post, then you maximize that by giving those people human-generated carefully chosen sets of 10, like the first game setup; a spiffy randomizing algorithm is not as good. So no argument about disappointed new players really defends the spiffy randomizer. And a lot of people go straight to the campaigns, so hey, we are already doing well there, except that the main set campaign sucks. The spiffy randomizer would be to make games more fun for people who have been hooked already and need endless sets-of-10. And for those people, those non-isotropic casual players, I have a lot of confidence in random, but if you want to provide your thing as another option, that's fine.

There's nothing sacred about pure random; it's important to have because it's the suggested real-life algorithm, and lets you play with all of your cards, and the expansions (the later ones at least) have been balanced with it in mind. I like the premise of "let's generate sets that are more fun." If the game had started as a computer game, it might have had a non-trivial card-selection algorithm from the start. The way the card frequencies fall out in your version though make me think that's not the approach for me. I don't want to see Margrave more often than Noble Brigand, etc., that's just making the sets worse for me.
Logged

Donald X.

  • Dominion Designer
  • *****
  • Offline Offline
  • Posts: 6357
  • Respect: +25671
    • View Profile
Re: Announcing Dominion Set Generator
« Reply #172 on: January 14, 2013, 06:28:30 pm »
+2

I don't want to see Margrave more often than Noble Brigand, etc., that's just making the sets worse for me.
I feel the need to hammer on this point. As it stands, Margrave shows up more often than average and Noble Brigand less often (for example), because Margrave has +3 Cards and +1 Buy and Noble Brigand just looks good in leather. But that means that with this randomizer, I play more games with Margrave and fewer with Noble Brigand. That cannot possibly be making games more fun; attacks in particular are a place where variety is crucial. Discarding down to 3 every turn is much more tolerable when you aren't doing it game after game. And the flexibility of Margrave gets it purchased in more games that it's present in; maybe we didn't even buy Noble Brigand that game it came up. So Margrave extra-dominates.

Again I recommend trying to group cards via your algorithm but with flat card frequencies; just as many Noble Brigands as Margraves.
Logged

onigame

  • Thief
  • ****
  • Offline Offline
  • Posts: 96
  • Respect: +73
    • View Profile
    • Dominion Set Generator
Re: Announcing Dominion Set Generator
« Reply #173 on: January 14, 2013, 06:50:28 pm »
0

For sure I would not have the order of a list of options change over time, that would obv. suck. You could change which radio button is the default, but only for players who haven't played yet - for people who've ever seen this screen, it should retain what they previously picked.

Yes.  Absolutely I approve of "sticky".  If you actually read my proposal (which you can't yet, because I'm not done with the first draft) you can see "sticky" all over the place.

Quote
If the concern really is, that the game needs to be at its very best for new-to-Dominion players, which seems to be your pitch in that one post,

That used to be my pitch but I've now changed my mind.  It has to do with the epiphany that "new" players and "casual" players are not the same thing, that I mentioned above.

Quote
If the game had started as a computer game, it might have had a non-trivial card-selection algorithm from the start. The way the card frequencies fall out in your version though make me think that's not the approach for me. I don't want to see Margrave more often than Noble Brigand, etc., that's just making the sets worse for me.

Yeah, if the game had started as a computer game, you'd be doing all this work and not me.  :)

I totally see the desire for having a uniform distribution.  Previously I felt unable to accommodate it, but your idea on BGG was pretty good and I'm working on that now.

(Me personally, I want to see Margrave more often than Noble Brigand.  But options are good.)
Logged

dondon151

  • 2012 US Champion
  • *
  • Offline Offline
  • Posts: 2522
  • Respect: +1856
    • View Profile
Re: Announcing Dominion Set Generator
« Reply #174 on: January 14, 2013, 06:57:44 pm »
0

So I've stated this already, and it seemed to have been ignored, but I think that weighting the cards such that the most frequent cards appear at most 150% as often as they are supposed to vs. a uniform distribution and the least frequent cards appear at least 66% as often would be a good middle ground to lower the frequency of "dud" cards (and even non-dud cards) while still allowing them to be strategically relevant.

Although, if your goal is to create "fun" kingdoms... doesn't veto mode do more or less the same thing? Cards that people dislike or are useless get played less often because people choose to strike them out.
Logged
Pages: 1 ... 5 6 [7] 8 9 10  All
 

Page created in 0.068 seconds with 22 queries.