??????
You are not logged in.
Josh, your out-of-range stake would be slightly more likely to get picked, simply because it would live on the list longer.
I suppose the server could wait for two people to pair up and pull one away into the amulet game instead. Say the amulet holder is waiting for a game. There are a bunch of waiting games on the list, but none of those are chosen. Instead, we wait for X of those games to be joined (where X is a random number). On the Xth one, we pull one of the players away and pair them with the amulet player, and simply leave the other player "waiting for opponent."
So, if you're picking stakes that no one wants, you'll never get an amulet game.
The randomness of being paired with an amulet player is a potential issue.... though it's not a pure lottery, because once you are picked, you have to exercise skill to even get the amulet, let alone rack up wins while holding it.
I don't see this as being all that different from random pairings in a tournament. It sucks to be paired against the very best player first, but that doesn't make it an illegal lottery.
The primary legal question with lotteries is: does an unskilled player have a chance of winning?
Here, the answer would be NO, or at least as much of a NO as for the main CM game itself. Is an unskilled player going to be able to rack up more wins than anyone else while holding the amulet?
The flipside is whether a skilled player could randomly NOT win, and I don't think that matters. As long as the winner clearly won by skill.
But, thinking about radio call-in contests which aren't lotteries, there may be a more skill-based way to pick the next amulet player. The hard part is preventing those colluding with the amulet holder from having an advantage in terms of getting picked through their skill...
My concern with altering player behavior is this:
1000+ brand new players will be experiencing the game for the first time and learning how to play during the contest.
If the game is skewed toward less normal, less interesting player behavior during that time, that will be BAD for those new players. They won't get a taste of the game. They'll experience some twisted, other game.
For example, if people jumped from all non-amulet games looking for amulet games, new players would experience a game where their opponent almost always leaves right away. That's not "normal" CM, and it would be a bad experience for them.
If there were one fixed stake for all amulets (say 25 cents), brand new players who wanted to try 1c or 10c or $1 or $10 games would have trouble finding an opponent. Most players would be playing nothing but 25c games trying for the amulets. That's a change to normal behavior, where normally your opponents are likely to jump at many different stakes that you propose. You'd create a 1c game to practice and then just sit there. CM would become a game of waiting a long time for an opponent at any stake (except the amulet stake), regardless of active player stats. Lots of normal player behaviors would be discouraged (like going up in stakes after winning a bunch at lower stakes, etc).
I even worry about this for multiple fixed stakes. While it's a nice match for the amulet rarity and other things that you point out, it would still stratify most players into those three stake "bins". Stakes outside those three bins would find very few opponents. New players who wanted to experiment with other stakes would find the player base behaving differently than normal, and that would be bad.
I'm trying to find something that can lay on top of the existing, normal player behavior without disturbing it too much, other than encouraging players to play more than normal. Especially for the (N-36) players who don't currently hold an amulet. I want them generally going about their normal business at a more fevered pitch. I don't want them to have a way to behave differently that gives them a higher chance of getting an amulet ("Don't waste time with any game other than 25c"---there should be NO tips like that).
I see the problem of someone getting mad that they didn't win the money that they thought they would win from a match and got this lousy amulet instead. Is that the only problem, though? This might have the effect of discouraging higher stakes... though you also have a chance of being surprised by NOT losing $500 when you thought you just did... Seems to come out in the wash. And most players, at most stakes, would be thrilled to trade their win $ for a chance to rack up some amulet wins. Someone did suggest a "No amulet games please checkbox" for high stakes players who really didn't want to be involved in the contest. That could be implemented!
Two amulet-holders would never be matched together.
People who get "stuck" with a copper one might be tempted to throw it away and search for gold. That seems okay. It's kinda the same dynamic that you point out with the varying stakes idea. Go for the copper, while I have the chance, or dump it for a slimmer chance at finding a gold?
New idea:
When you hold the amulet, you are paired with a totally random opponent (who is also currently looking to join or create a game). The true stakes of the game don't count (no money is won or lost), because you're playing for the amulet itself.
Example: you hold the amulet, and let's say you create a $25 game.
This $25 game never shows up on anyone's game list. Instead, you are paired automatically either with an already-waiting player (maybe someone who has created a 15-cent game, for example), or with a player who clicks on some other game, trying to join it.
The other player THINKS they are playing a normal 15-cent game. When they leave, they find out that the money stakes were actually bogus, and that the game was for the amulet only.
Essentially, we'd pair up a $25 player and a 15-cent player together in what is effectively a 0-cent game, with the amulet going to the last player standing.
But the non-amulet player wouldn't know it until after the end of the game. If they lost, they'd be out 0-cents, but they'd be expecting to be out 15 cents, and see that they also could have won the amulet. If they won, they'd gain the amulet.
So, this wouldn't change player behavior at all (play lots of games at whatever stakes you would normally play at---any of these games could end up being a 0-cent amulet game instead, but you won't know until the end of the game).
As long as the randomized matching was good and took it's time (okay to keep the amulet player waiting while a random opponent is picked, but this won't slow down other players who are joining games), this would also prevent collusion.
When an amulet holder goes to create a game, the game could have a counter on it with a random starting value. First, the counter is decremented for each existing, unpaired game that's waiting. If it ever hits 0, then that other player is matched with the amulet player.
If no, then every time some other player calls join_game (either to join an existing one or create one), the counter goes down, but that other player does not join the amulet game (they instead join or create a game normally). When the counter hits 0, THAT other player is chosen and joins the amulet game in the place of whatever stake they were trying to join/create, unbeknownst to them.
To prevent some player with 50 accounts from having them all join_game right in quick sequence (where one would surely get through, when the hidden random counter was finally decremented to 0), we could also have a randomized-each-time minimum time delta that must pass before the counter can go down another step. So, join_game calls before that random time delta has passed won't decrement the counter. And the time delta could be set to a new random value after each time the counter is decremented.
And to prevent join-cancel-join-cancel spamming (where an accomplice account automatically creates and kills an unmatched game repeatedly until getting through), any player that just canceled an unmatched game could be out of the running to join an amulet game for X minutes.
To prevent some player from creating X accounts (where the max random counter the server ever sets is X-1) and simply creating loads of games with them at the right moment (so that the counter decrements to 0 while working through the list of existing games), the max possible counter value can be 2 * E, where E is the CURRENT number of unpaired games existing. So, when we pick a random counter in that range, we have a 50/50 chance of picking an existing unpaired game to match with or waiting for future game joiners/creators.
This means that when there are lots of waiting, unpaired games, the amulet holder will wait longer to find an opponent. Not sure if this makes sense or if there's a better way. Wouldn't want a bunch of accomplice accounts to swamp the pool late at night when there are only 5 "real" players.
Maybe there's a better way? The ideal would be to pick a player at random from the pool of all active players and simply wait for them to eventually join a game. That would make the amulet holder wait too long, obviously.
So, it looks like this CAN be done!
The key insight here was that there are only 36 people with an amulet. While we don't want to drastically change the behavior of all the other non-amulet holders, it's OKAY to change the behavior of those 36. We can make them wait longer for an opponent, openly ignore the stakes they pick, or otherwise change their experience without harming the main game at all (because after all, when you're holding the amulet, your behavior is going to change anyway).
A few things have become clear:
Amulet games can't be marked as such for non-amulet holders because it changes player behavior too much. They can't be marked on the game list, because then players would be refreshing waiting for nothing but amulet games. They can't be marked once you join a game because then people would dump non-amulet games shortly after joining (to go back to the list and try joining a different game, hoping that it's an amulet game).
Of course, the amulet holder would know that they have the amulet. Without this piece, the contest wouldn't be very exciting or interesting.
We can't let the amulet holder name any stakes they want, because they could use high stakes to guarantee that their accomplice account joins their game.
Having one fixed stake for all amulets would change player behavior too much (no one would play other stake games at all).
We could have 36 different fixed stakes, with each amulet getting a different stake. The problem with this is that a given stake would be "set" by the amulet holder creating a game. This means that non-amulet holders would never want to create their own games. They'd always be trying to join games matching one of the 36 known stakes. (This would change player behavior too much, with everyone joining games and no one creating non-amulet games).
If the amulet holders could not create games and could only join them, then all non-amulet holders would never want to join games---they'd all be creating them---so that would change player behavior too much also.
If the stakes were confined into a range (say, anything under $2), then anyone who wanted an amulet would never play games outside of that range.
Leaving those extra steps in there when people are all-in has just been laziness on my part. It keeps the game structure the same no matter what, so the server code is simpler with fewer special cases.
Gotta do it, though.
In terms of time, I think most online poker gives you 30 seconds per choice. I used to have it set to that, but a lot of players complained that it was too fast.
Remember, in Poker, the only choice you're making is the bet, and there are no choices to make AFTER the bet, either, besides more bets.
Here, you're making a real choice (column picking). Even on the betting choices, many of them depend on future picking choices that you still haven't made yet. So evaluating those possibilities in a deep way requires more time then simply looking at 5 or 6 cards in poker.
Jere has pointed out that after pick 2, during the post-pick-2 bet, you can actually discover that you could force a win on pick3. You can't just glance at the board and know that. It takes some mousing around and studying the graph.
If you are NOT studying the graph post-pick-2, you are going to miss those forced-win certainties and play worse as a result. I'm not saying that everyone should do this, but many advanced players do, and it takes time.
How fast is it to look at the board in poker before the river and know you've won? Like, let's say you have an A-high flush before the river. Well, even then, the draw could give someone a full house. So really, only 4-of-a-kind and straight flush before the river let's you know you've won. Simple to recognize instantly, and it almost never happens.
How did I end up with the smartest players in the world playing my game?
:-)
I felt like if amulet games were at a set stake (kinda like tournaments, with an always-there stake button), you'd often join that stake and be disappointed, or people would all be spamming that stake, and it would change player behavior too much. Maybe you're a new player and you're playing nothing but penny games. Maybe you're Judge Doorman and are bored even by ten dollar games. This is "normal" player behavior. But everyone who cared about the amulets would be playing nothing but 10-cent games for a whole week.
You're right about high stakes making it easy to join with your accomplice, even with a substantial display delay. And people using accomplices late at night would have a big advantage too. Even fixed stake doesn't really help much at a time when everyone else is asleep.
Yeah, with the display delay, if two people started a $23.45 game at the same time, they'd never join. One would start and wait for a joiner. That game would exist with one slot empty. The second player would go to start one, but check for an open one first, and not see the existing one because of the display delay, and thus start a second game with one slot empty at the same stake. They'd both continue waiting.
And I was suggesting that non-amulet games would have no display delay. So of course the amulet holder would just insta-join the game of their non-amulet-holding accomplice.... sheeeesh!
And if an "amulet" game could only be created by the amulet holder, you'd have all the rest of the players never creating games.
And if you said, "Server gathers X people who want to play a game at this stake before finally picking one to let join at random," someone would just use X accomplice accounts.
Rolling back here a bit....
Is there no way to uniquify accounts? Clearly credit card numbers don't work, because of ubiquitous throw-away numbers from gift cards.
What about phone numbers? I do have an auto-dialer API that can call you and deliver a unique code.
Maybe it's easy to create dummy phone numbers now through GoogleVoice...
Truly uniquifying accounts in a non-intrusive way (everyone fax me your driver's license!) seems like a dead end.
Went outside and stared at the sky for a few minutes... Aha!
So, clearly amulets have to be hidden during games to non-amulet holders. That means no join delay or anything, and no special stakes. Also, it means Josh's suggestion of no huge leave penalty for amulet games. The last one in the game, or whoever takes all the chips, gets the amulet. Score for an amulet based on number of games played while holding it (can you hold it more than once? Sure, why not?)
So, how do you prevent collusion to boost your amulet win count?
Instead of a random join delay, there needs to be a random display delay, where each game appears on each player's game list (and is joinable) after a slightly different amount of time. This should be the way it works anyway, for all games, to increase anonymity.... or maybe just for amulet games (don't want to add extra delays, generally).
Let's say the max appearance delay is 30 seconds.
Each player is permanently assigned a random number between 0 and 30.
Each new game is assigned a random number between 0 and 30.
When a player asks for the game list, for each game, their number is added to that game's number, mod 30. If the game is older than that number of seconds, the game is shown to them and is joinable. If not, the game is hidden and cannot be joined.
This effectively assigns each player a random delay for each new game that is created. These delays are always in the same order, in a modular loop, but no player has the advantage of seeing all games before other players.
This can be exploited with many accounts (because they'll all be spread out in the mod space, so one of them will have a low delay for any game). But if you have multiple accounts hammering on JOIN for a given stake level, and a bunch of them fall through because their delays aren't up, they're going to create their own games at that stake level and end up joining each other (especially if the display delay only applies to amulet games---spamming JOIN from multiple accounts is just going to get those accounts instantly matched against each other).
Also, we can make this as hard as we want by increasing the max display delay. For example, it could be 10 minutes. The trade-off is that the amulet holder will likely wait longer to find an opponent, but they've got a hot potato anyway, so they can wait a bit.
This makes collusion to rack amulet wins much harder than simply picking one accomplice account and joining the right stake at the right moment
I don't know that I want to change the WHOLE game too much during the contest.
Remember, this will be the very first taste of the game for 1000s of new players. So, turning on some huge leave penalty for all games doesn't seem right.
Amulet hopping will happen. If you want an amulet, and that's your goal, then every second you spend playing a non-amulet game is a wasted second. Granted, hopping has a cost of 6 coins, but if you're playing penny games looking for an amulet, it would be worth it. Even at a fixed stake of 25 cents, hopping costs less than 2 cents. I can imagine a bunch of people putting in $10. That's enough to cover 500 hops. Okay, they can do this if they want to. But imagine the experience on the other end for a brand new player who is NOT amulet hopping. Even worse for a new player is someone who wins 8 coins and then leaves (just enough to cover the leave penalty).
Even if amulets are hidden to non-amulet holders, any other hints, like an extra delay before joining the game, will be used to sniff out amulet games. And we'd need join delays to stop collusion. And I wouldn't want to add join delays to ALL games (again, that would suck for new players).
The cool thing about Steal Real Money is that it measured and rewarded "normal" behavior in the game. It didn't use an outside metric like "who has the most pitbulls," which would have skewed behavior.
The problem with measuring normal behavior and attaching an outside reward is that it is vulnerable to collusion.
The cool thing about CM is that it isn't vulnerable to collusion at all in its natural state. In this regard, it's unlike TCD, where collusion could give you a large advantage. So, Steal Real Money just amplified the collusion tendency that was already there... it didn't create a new one.
Anyway, clearly EVERY game can't be an amulet game, because that's vulnerable to collusion. It also doesn't solve hopping, because some amulets would be more sought-after than others.
Even flagging games at random as being amulet games doesn't stop collusion (just have two colluding accounts play tons of game---this would ensure that they get lots of amulet games).
Storeroom... I read this a few times and have a slight grasp of it, but I'm having trouble seeing how it would stop someone with 20 accomplice accounts from having an advantage.
Every time you win while holding the amulet, you get the stake points that are attached to it and the total points on it go up (so the next stake you'll win will go up too). When you finally lose, you get all those built-up points.
Amulet starts out with 100. I win. I get 100. Amulet goes up to 200.
I win again, I get 100 again, Amulet goes up to 300.
I win again, I get 150, Amulet goes up to 450.
I win again, I get 225, Amulet goes up to 675.
I lose, I get 675 points, the amulet passes to someone else, and now has 338 points in it.
My total points are 675 + 225 +150 + 100 + 100 = 1250
Or do I have that wrong? Maybe I don't get points each time I win, but only get all the points in the amulet when I finally drop it. But you said:
If you win you'll get x number of points and keep the amulet, if you lose you'll get y points and lose the amulet
Maybe I only get 675. Still, winning a lot before dropping it benefits me.
Anyway, as long as winning while holding the amulet is good, I'm motivated to make a bunch of accomplice accounts that I can win against, right?
Yeah, cash along with each amulet is fine. It can match the rarity of each amulet. More cash for gold, less for silver, less for copper. But the gold ones could certainly be sold for their gold content... still, it would only be a $600 value if they are half an ounce each. I think they'll be worth far more at auction as unique, one-of-a-kind objects. There are 728 registered CM players, but a lot more people who know my work. Limited editions of Diamond Trust go for around $100, and there are 1000 of those in existence.
Given what you guys are observing about winning streaks being generally short, it seems like ties are going to be a huge problem. Need to deal with that up front...
Regarding the 20-account collusion, who has 20 unique credit cards? Oh, crap, visa gift cards. Could easily buy 20 of those...
I don't know that forcing new accounts to play X regular games first would help.... someone would just create their 20 accounts ahead of time and rack up the necessary games.
Some kind of randomized match-making for amulet games could help. However, you could imagine the colluding amulet holder just making a game with stakes that most new players wouldn't want to join ($50 or whatever), which would give their accomplice account plenty of chance to be the one to join.
Amulet games could have a fixed stake (like $1.00), and then randomized match-making would work (just a simple random delay of when a game shows up on the list).
I was hoping to let amulet games happen at any stake, though, but I wonder if there would be a race to the bottom anyway (if you're trying to win the most games in a limited time, proposing $100 games is probably a bad idea). Though a set $1 stake might push it out of reach for brand new players.
One bad side-effect of this contest would be people game-hopping looking for the amulets. If you join a game and it's NOT an amulet game, you'd want to leave that game as quickly as possible and try joining another one.
So, if you have the amulet, you'd want to play as much as possible, and all your games would go to the end (winner-take-all).
If you don't have the amulet, you'd want to JOIN as many games as possible... but not play as many games as possible.
Could the amulet status of your match be hidden until it's over? YOU JUST PICKED UP AN AMULET. Then everyone would be afraid to leave any game before conclusion, which would change the game quite a bit.
I just thought of something else.... maybe the 200-penalty for leaving should only apply to the player who is currently up in coins (or both if they are tied).
So, if the amulet status of a game is hidden from you as the non-amulet holder, and you're down 50 coins, you can safely leave for the normal 6 penalty, whether or not its an amulet game...
This would motivate non-amulet holders who are up in coins to stay in every game, just in case it's an amulet game, but still let them leave a game when they are way down without a shocking penalty for leaving. Because, obviously, the leave penalty would have to be hidden (it can't say LEAVE AND PAY 95, because that would let you know it's an amulet game).
On the other hand, as an amulet holder, you'd always know it was an amulet game...
Hmm... we've kinda come full circle here. Hard nut to crack indeed.
Josh is right about collusion. Imagine someone posting in here... "I've got 13 tokens worth $819.20. If someone is willing to meet me online and lose 1 token to me, which is only worth 10 cents to you, I'll go up to $1638.40. I'll pay you $400 for that one 10-cent token."
AnoHito: part of why I don't want a tournament involved, not even as an end game, is that it seems to cater to the established players too much. I mean, who won every tournament? Jere! Or at least he placed very highly in every one. Most likely, he'd win the contest if there was a tournament involved. Or JA or BG... the usual suspects.
The one-football idea was interesting because someone like Jere might not ever get his hands on it (though he and everyone else would obviously try), and even if he did, he'd only hold it for one day max. Granted, that idea simply doesn't touch enough players. Though... given Chain World and A Game For Someone, I'm not sure that *actually touching* players matters as much as giving them something exciting to hope for and talk about. 3 people in the world actually played Chain World, but it is by far my "most popular game." 2x the Google results of TCD. Well, not more popular than Passage, but still, for a game that 3 people played...
And Steal Real Money also made it clear how important the STORY of the contest is.
I do like the idea of a Golden Ticket... trying to conceive of a skill-based version of that.
So... I'm coming back to the "one amulet" idea (or maybe a small handful of them). I think, if you were trying to rack as many wins as possible while holding it, it would be pretty hard to hold it for long, if one loss makes you drop it to someone else. If you could only hold it once during the week, and could only play each opponent once while holding it (to avoid racking wins against an accomplice), I think it would bounce around through lots of different people. Really, I can't imagine one person holding it for more than an hour if they were actually trying to rack wins and not just hold it to be a jerk. So that one amulet would touch at least 168 people during the week. Make that 6 amulets, and we're talking about 1000+ people touched (who actually held it).
Beyond that, there's all the people who had a chance to get the amulet. How many people can the amulet holder play in an hour? At least 6. So that's 7000 people who would have the excitement of at least one AMULET GAME during the contest.
If those six are half ounce gold amulets, this would cost $3600.
Maybe throw in some "lesser demon" amulets as well in silver and copper... those would be cheaper, and thus there could be even more than 6. I dunno... 36 of them, why not? 6 gold, 12 silver, 18 copper. 6000 people could hold one at some point during the week. If we have that many people playing during launch week, we're doing pretty good.
Maybe you can hold each amulet once, so after holding one, you can keep trying for others. But when they're actually awarded at the end, you can only win one, starting with gold and working down to the smaller ones. After you win one, you're taken out of the pool (like, if you won a silver, you couldn't win a copper too, even if you had more wins while holding that copper amulet than everyone else).
The game client could be modified to show the actual amulet on the screen during an amulet game, and the game could have a 200-coin penalty for leaving.
I've actually got some experience with lost wax casting (I made my wife's wedding band), and the UC Davis craft center has lost wax casting equipment---they have copper and silver there for purchase and say that I can bring my own gold. So I'd probably just make these myself, which means they'd be really unique and rustic and strange and spooky looking.
(Seems like, after launch week, something like this could continue on a longer scale.... like a whole month with an amulet passing around.... maybe over time, they could get bigger and better, add precious stones, etc. I guess the protocol should let the server specify the amulet image that the client will display...)
There's an extensive document in server/protocol.txt
This doesn't necessarily spell out the order in which things happen, but every server call is documented there. It's all straight HTTP, with simple text responses back.
This is an excellent idea!
One down-side: it limits the number of tokens that can be floating around. 16 is doable... 100 is not!
Unless there's some cap on the maximum number held by one person.
Also, if this was combined with "win one or lose one," where only one token is ever at risk even if you have more, I think people would be willing to keep playing and not turtle...
Well, figuring out what the rule should be is itself kinda confusing.
If you have no tokens, how many times can you try playing a given player with tokens? If you win, will they have a chance, ever, to play you again and win their tokens back?
In the tournament, you stopped seeing someone as an opponent once you had won $X or more from them.
You're suggesting a kind of "Play Token Game" that's separate... that you have to pick on purpose.
I was thinking of this as something that is happening more under the surface. Like, rewarding you for playing the regular game. Surprise! You just beat someone who had tokens and you got some.
Maybe a quota of games per day isn't needed to hold your tokens.
Maybe you can just decide to quit when you want to quit. Won't greed encourage people to press their luck to win more tokens, especially if they only stand to lose half of them?
'Cause that's the only reason you need to control repeat match-ups (so someone can't use an accomplice account to fill their quota).
One worry (without quotas) would be that 500 people would just keep their starting token and be happy with that, and then stop playing for fear of losing it.
Yeah, I dunno... it seems like this idea will encourage certain people to stop playing (or even with a quota, play as little as possible).
Seems more like bonuses should be un-lose-able once you get them.... and that there are only more to be gained by playing more. The original one-football idea had this, because just having the football meant nothing... the bonuses came from playing a lot while you had it (and though you could lose the football itself, you couldn't lose the win count that you had built up so far).
Oh, so it is essentially computing the whole partial-information game tree up front. So, your results pretty much agree with mine, then.
Hmm... yeah, this is getting close!
So... one thing with "winner take all" in terms of the golden tokens is this: they will snowball, and the snowballs will keep getting bigger. There's no way for a heap of 20 tokens to ever get split back up into two heaps of 10.
As Jere points out, people holding snowballs will become as conservative as the game lets them be. This was "fixed" in the original single-football idea by counting wins while holding the football, so you want to play as much as possible when you have it. Here, when you have one or many, you want to play as little as possible.... especially against people who have no golden tokens. Getting stuck in a game with a non-token-holder when you have tokens yourself is the worst. Nothing to gain, everything to lose.
One simpler idea is that when you lose, the winner gets half of your tokens. This would break up the snowballs, and be closer to what you're talking about, Jere, without being too complicated. If you have only one, you lose it, of course.
But still, aside from the "to keep my tokens" quota, there's no reason to play once you have a bunch. And the quota can't be THAT frequent, because people do need to sleep. Hmm... if it's just X games every 12 hours, it seems like someone would play dummy games against an accomplice account to fill their quota. Unless we go back to blocking you from seeing players that you've already played (for a single big football, that made sense, but with 1000 flying around, it will be hard to explain and for players to reason about).
What about... when you have tokens, you can only play people who don't have them... Then there are no snowballs at all... Two accomplices would just pass them back and forth.
Well, my wife got this book out of the library for our kids. When she saw these pictures, she showed them to me. "You just HAVE to use these for something."
So, they are "about" CM in a way...
Yeah, .. is right on.
Start in game.cpp (CordialMinuet/gameSource/game.cpp).
This file both implements and calls parts of the minorGems/game/game.h platform independent game framework that all of my games use (a platform independent wrapper for networking, graphics, sound, and user input).
game.cpp splits the game up into multiple GamePages, and keeps track of the current one. It checks that page every frame to see if its status has changed, and if so, executes the page change-over logic, and then sets a new current GamePage. For example, if they click DEPOSIT on the MenuPage, that will take them to the DepositPage, but if they hit CANCEL there, it will take them back to the MenuPage.
The main logic for the actual gameplay all happens on PlayGamePage, which is the page that you go to after joining a game when the game starts. That page is actually pretty hairy, because it processes a bunch of different protocol responses (most of the other pages just handle one protocol request/response). That page also computes and draws the graph. There's a lot of code in there.
Yeah, Jere, the "luck" of being blessed with the football at the start is a problem. I wouldn't be legally allowed to do that randomly. Probably give it to the player with the highest Elo or something. There's also the idea of some wealthy player "buying" a football. Like, it costs $100, but if you win the most while holding it, you get $5000. So there could be a bunch being passed around.
I'm not totally clear on your mote idea. Yeah, I could imagine "The first 1024" people who deposit after time X getting one (that counts as a contest of skill, like the radio call-in contests). But what happens then? Do you pass it to someone else when you lose? Do you gather them by winning against someone? So, by winning against a lot of people, you could build up a bunch, and then get paid for them at the end? Or if you've built up 100 of them, and then you go up against someone who has one 50 of them, and you win, do you then have 150?
Anyway, in addition to each one being worth $2, we could have first, second, third, etc. places for the person who gathered the most of them. And along with the cash they'd get, they'd also win a physical keepsake prize.
Amping up the leave penalty makes sense to enforce binary win/loss for "special" games. But I do worry about changing the game skill too much.
Also, if you're holding a football (or footballs), and we amp up the leave penalty, it means that every game you play is winner take all until you drop it.
This seems pretty safe from a 2- or 3-account exploit. Yeah, you could get a lot of money by registering 100 accomplice accounts, but that would require 100 credit card numbers.
Would also need to figure out how to make this sound really exciting in one catch phrase. STEAL REAL MONEY is pretty tops. "$5000 Football" (or thematic equivalent) is also pretty inspiring sounding, even if in reality, it would suck for most players----it would still get them excited. Explaining this in a concise, exciting way is a bit harder...
One thing that I really like is the idea of handing out physical gold (been thinking about "medals" or "amulets" in gold, silver, copper, etc for a big tournament). Maybe tiny occult beads of some kind that are each 1/100 of an oz. So, instead of money "motes" there would be tokens that are actually worth real motes of gold that would be mailed to you. Maybe if you collect a lot of them during the contest, you can literally combine them into a larger, more interesting amulet.... like, 100 of them will get you a 1 oz amulet, or a ring or something.
By having a "first 500" limit, there could be nice hard limit on how much gold is at stake and how much mailing I would potentially be liable for.
That might be more exiting. WIN REAL GOLD or something like that. Get them blessed by The Order of the Golden Dawn or something.
Especially since so many games have "gold" in them... but never real gold.
Josh, does your method pick all 6 moves up front, before seeing any of the opponent's choices? Or does it make decisions step-by-step, two picks at a time?
Yeah, Jere, I was thinking of physical prizes here as well. There are no shortage of interesting occult items, for example....
Storeroom---this sounds like an interesting tournament structure. I'm really hoping to stay away from "tournament" for this promotion, though. It sounds way too intimidating to brand new players.
One idea that just came to me:
Call it "The $5000 Football" plan.
Maybe just one "football" that moves from player to player in the game.... you have to play frequently to keep it (at least X games every 12 hours), you know when you have it, but no one else knows that you have it (or maybe it shows up on the leaderboard). You have to play DIFFERENT people to keep it. Maybe when you have the football, you stop being able to join games against people that you've already beaten while holding the football (so you can't exploit a few accomplice accounts repeatedly)
Once you lose while holding the football, that's it, it's gone, and the person who beat you gets it.
And you WIN the big prize if, at the very end, you won the most games while holding the football (or won the most chips, so stakes don't matter, but clean-sweeps matter).
So, this encourages everyone to play a lot.
The person with the football wants to play a ton (to up their chips-won count while holding the football).
The others who haven't touched the football yet want to play a ton to increase their shot at playing the football-holder and winning it.
Seems like there needs to be some way to handle partial wins/losses where one player wins more chips. If a few chips change hands and one leaves, who won? Can you win the football by winning a few chips? Maybe you can only win it once you take at least 75 chips total (maybe across multiple games) from the football holder.
Maybe there could be a big prize for the longest streak with the football and ALSO a big prize for the player that ended that long streak. Other prizes for shorter streaks. This would give brand new players reason to try (they may have no chance at a long streak, but a good chance of sniping.)
Maybe you only hold it for 24 hours max. If you lose during that time, you lose it, if you hold it until the end, it's given to another random player. But a top player can't hold it all week that way...
Maybe there's more than one football traveling around at the same time. Maybe they each count for a different prize, with separate stats tracked for each one...
It seems exciting for this football to be traveling around in the game.... you never know when it might reach you.
Thoughts about this?
(Of course, it wouldn't be a football... an amulet or something more thematically appropriate... maybe it could even appear visually on the game screen).
I'm trying to come up with some kind of contest to promote the launch of the game.
Those of you who played The Castle Doctrine may remember the Steal Real Money contest that I held the week before launch:
http://thecastledoctrine.net/stealRealMoneyContest.php
This worked well because it attracted both existing players AND brand new players. There was no entry fee (beside buying the game), so even new players felt like they had a shot at winning.
One problem that I didn't discuss too much at the time is how an outside prize really motivates cross-account collaboration and cheating. In the case of the TCD contest, people were motivated to build up huge bounties on one account and then "dump" the bounty in their main account by dying. If they waited until the last minute to dump the bounty, then the money was safe from being stolen by other players, and their main account could suddenly jump to the top of the list. People were also motivated to camp in their houses right at the end. I have a foggy memory of instituting emergency "house kicking" right at the end to keep things fair. Or maybe I made the robbery time limit more strict to prevent robber camping.
The point is, these issues came up. It sucks to have winners of a contest who may have exploited some loophole to give themselves an unfair advantage.
I've thought about holding some kind of tournament with a big outside prize, but it feels like a tournament would scare away new players "I don't have a chance, so why bother?" It also seems like an entry fee is necessary to prevent a tournament from being exploited by multiple accounts. If it's free to enter, you could pool thousands of accounts together and have them all lose to one account, causing that account to win the tournament. Entry fees make this prohibitively expensive (especially combined with a tournament structure that blocks one account from winning too much from another account). BUT, an entry fee would make it a no-go for brand new players. Also, if the entry fee is way smaller than the prize, then multi-account exploits become more attractive.
So, I'd like to just have an "open" contest over a week or something. Reward biggest profit ratios, biggest Elo jumps, most games played, most games won, I don't know.
The problem is that these are all easily exploitable with multiple accounts.
Been stuck on this for a while now...
Yeah, I tried that too.
Assume a random opponent, and pick the optimal order for the 6 columns that I'm going to pick. That method was only netting a very small advantage... something around 2% (52% win rate when the number of trials got large).
Still, having it refresh "out from under you" when you're about to click on a game would be annoying. I forgot about that.
It almost seems like the refresh button should disappear for 5 seconds after you hit it, just to prevent you from spamming it.
It currently auto-refreshes if there are ZERO games returned. I figured that in that case, there's clearly no strain on the server.
It doesn't auto-refresh if at least one is returned. The idea is that once the server got busier, auto-refresh when someone was away from keyboard would cause a lot of wasted load. Of course, with the refresh button there, someone can simply spam the button as much as they want, and cause all kinds of load.
Maybe it should auto-refresh if less than one page was returned, and just have NO refresh function (auto or manual) if multiple pages are returned, because paging can be used to refresh.
Or maybe auto-refresh should be on always...