I found THE solution to the collusion problem

jere · 2015-04-01 18:11:38

I was taking a walk in the park today and I had a sudden flash of insight. There are several criteria that a solution has to meet for this contest and they cannot be bended whatsoever:

The contest has to measure REAL SKILL (i.e. playing and winning games against a REAL OPPONENT).
The contest CAN'T CHANGE the experience for players.
The contest has to be TOTALLY AND UNEQUIVOCALLY immune to cheaters and colluders.

I have a solution that meets ALL THESE CRITERA. It is 100% collusion proof. The best part is this contest requires NO ADDITIONAL CODE.

The solution is simple: amulets will be awarded based on who wins the most games against Jason Rohrer. Yes, that's right. Jason Rohrer will play Cordial Minuet by hand continuously for 5 days. In the course of those 5 days, all you have to do to increase your chances of winning an amulet is play a game against Jason Rohrer and beat him. That's it.

Jason Rohrer will know with certainty that these players were not colluding with Jason Rohrer and thus the solution is perfect.

One more detail about this solution: (Happy April Fool's!)

CaravanDisturber · 2015-04-01 20:12:32

This could work.

jasonrohrer · 2015-04-01 20:28:31

HOLY MOTHER OF DUNG!

I screamed with joyful laughter when I read this.

This is perfect.

Isn't taking a walk in the park the best thing for lightning bolts like this?

So... details would need to be ironed out.

First of all, how am I going to stop playing Rust for five days straight?

More seriously, is it the most total money won from me? Or maybe most chips won? I don't think the "last person standing" rule should matter here. Would losses against me count as negative points?

The whole "RECEIVING AMULET" and amulet point count display in the client could be salvaged so that the top 36 players would each hold an amulet currently and their "points won against me so far" could be displayed.

Or something... would be cool to preserve that exciting moment in the game of HOLY CRAP, I GOT AN AMULET. Maybe right after you win from me, there's some amulet display...

Anyway, beating the creator of the game, anonymously and without being sure if it, fits the occult themes of the game perfectly. It is also a great PR story.

There is some concern about my exhaustion/sanity/health.

Maybe there could be some way to elect 36 "high council" members who would each be defending one amulet during the week. Of course, that would make you all (Jere, et al.) ineligible to win an amulet, which would suck for you.

Yeah, I guess I would just have to suck it up and do it.

jasonrohrer · 2015-04-01 20:34:12

It's funny... I was logging into the forums today thinking... man, when is one of the smart people in the forums going to find a solution to this mess and save me?

And there it was.

But wait, were you joking? It doesn't matter if you were!

(Also, by "suck it up and do it," I didn't mean play for non-stop for 120 hours... I meant play quite a bit each day at mysterious times.)

CaravanDisturber · 2015-04-01 21:13:36

If Jason does play, I think his username should be "Cordial Minuet". Would be exciting seeing that come up afterwords. We should also always just have a non drop able Amulet Titlebelt. That could be passed around, like a wrestling belt.

jere · 2015-04-01 21:15:37

The "walk in the park" thing was a joke. I figured one day you would just dump the solution on us out of nowhere and that's probably what you would say.

And yea I was joking, but there's a grain of truth here. I saw your post about CM being cheat-proof "end of story" and it made me realize that, yes, CM's metagame is totally cheat-proof in the degenerate case that you are one of the players. Anything else is vulnerable to collusion and a search for a perfect solution will only result in one place: this absurd idea!

It's not too crazy I guess. You really could play hundreds of games over a week if you wanted to. Indeed, it would be an easy PR sell. The headlines write themselves:

Win Jason Rohrer's Money

Jason Rohrer Will Pay You to Lie to Him

Even if it was just the kickoff to another system (e.g. amulets are handed off from you to a victor, the gain points until they lose it, and then it goes back to you), it would still add an interesting layer. If nothing else, I'm glad someone got a kick out of it.

Ghost Amount · 2015-04-01 21:51:52

I for one would play my ass off during that time in the hope of playing the "Creator"

LiteS · 2015-04-01 22:05:32

I'm actually super down with this solution. It solves the problems that arise with multiple accounts, slow play, and even last two hours of the contest game spamming. On a meta-level, it even solves the "No games played for more than $3" effect observed during the trial contests. The only thing left is to flesh out exactly how to score each player. "Chips won at end of game vs Jason" would be a nice metric if we want penny-stakers to stand a chance.

jasonrohrer wrote:

There is some concern about my exhaustion/sanity/health.

It'd be a personal challenge to distribute your play time to make amulets available during all time zones, but I think this is a risk we're willing to take.

EDIT:
Jason's account name should probably be kept as secret or anonymous maybe? With Canto Delirium people would probably flock to the game when there's activity on that account, but then it'd just be up to Jason to play as sporadically as he chooses.

Last edited by LiteS (2015-04-01 22:09:11)

jasonrohrer · 2015-04-02 00:48:17

Well... you all don't know my name on the leaderboards, do you?

I just did a forum search for it, and my alias has never been mentioned here.

If there was some kind of flashy display in-game after playing me, people would try to use Canto to figure it out collectively, which would be a cool meta.

And then whenever I would come online and start playing, there would be a collective HE'S HERE! moment.

Yes, it does solve all of the other problems too (no longer need fixed stakes, or stake caps... is that Jason trying to bait us with a $25 game?... no longer need random delays of any kind.... player behavior isn't altered at all, besides playing more... essentially, the game is played exactly how it's been played.... heck, I've been playing myself from time to time, so even that would be the same.... also, we've already had HE'S HERE moments in the past when people were waiting for judge doorman to come online, so that's business as usual too...).

Also, I just realized that chips lost must count against positive chips won. Otherwise, if losses didn't matter, there would be a motivation to play a risky aggressive style in hopes of racking up some big wins that would add up.

If I play really well, the winners could actually be the ones who lose the least against me!

Also, there's something a little strange about counting just chips and not money. So, you could lose 100 chips in a $100 game against me, and then win 180 chips in two penny games against me, and then you'd be up by 80 points? But you're out over $99.94. So that seems to change the game somewhat, where the money doesn't matter. When people were trying to play against Judge Doorman, they were probably playing very differently on the $1 games vs the $500 games. Also, clearly Doorman himself was at least trying to play differently.

And I'd want to preserve that... maybe you'd be throwing several penny games to lure someone up into 10-cent games, but if you cared about the amulet and chips counted over dollars, you wouldn't want to do that.

Well, I guess even if we were counting dollars, you wouldn't want to do that, because the thrown penny games could be against me, and the 10-cent games might not be. Of course, when counting dollars, thrown penny games don't really hurt you much.

It's interesting that having dollars count doesn't necessarily privilege rich players, because playing high stakes against me just means their swings could be bigger. If you play against me in that one rare $50 game that I play and lose, you would never catch up if I never played at that stake level again.

There would clearly have to be some kind of leaderboard for this...

I'm still waiting for a bit more lightning here about how the amulets fit thematically into this. Are there still 36 of them? Are they just 1st - 36th place, which is kinda boring? Or do they represent something else? Times of day, stake levels that I'm playing, numbers that you stuck me with during the match? It would be interesting if you could still score separately on each amulet somehow, by playing games against me.

Jere, I now really appreciate how you baited me perfectly with the whole "walk in the park" lead-in!

Cobblestone · 2015-04-02 01:03:24

I'm not sure if I love jere's original post more, or the fact that you're off and running with this idea, Jason!

joshwithguitar · 2015-04-02 06:14:41

You would have to make sure there is no great advantage to Americans just because they are awake at the same time as Jason. The idea of different amulets for different times might fix this - though still someone with similar sleeping patterns to Jason will have the opportunity to complete for more amulets and this will still improve their chances.

Jere: It's not 100% collusion proof - you can get a slight advantage with colluders by simply exiting any game that you start between another colluder as you know it isn't a game against Jason. Also, it will require additional server code to keep track of score etc.

jere · 2015-04-02 13:21:24

What I meant with no additional code was that technically the winners could be determined with a few mysql queries. It would never actually be implemented that way, but I was trying to amp up the joke (which I really didn't think anyone would take seriously). I do empathize with you on the time zone issue. If Jason is willing to adopt some odd sleeping patterns and keep an eye on which hours he has played (easy with a query), then he could balance his games throughout the day. But then does that give you the advantage because you're competing against fewer players for games?

If we're going back to games on demand as opposed to the random delay stuff, then your particular collusion concern isn't an issue. Because you normally wouldn't join games with yourself. It makes no sense.

Your comment made me wonder about a few similar issues. If you had 50 visa gift cards, you could open up 50 stakes and dramatically increase the chance that you might play a game with Jason. But then you have a few big problems: what happens when several of your stakes get joined? Do you play all of them simultaneously? And even if you do manage to funnel Jason into a game, you're reducing your chances of wracking up many wins against Jason down to nothing. So I'm still not seeing collusion issues. And even if they do arise, there's certainly will never be a way to get free points.

Jason, you'd have to put a cap on the stakes right? I assume you're not going to be playing $1000 games against Judge Doorman or Nate.

Are they just 1st - 36th place, which is kinda boring? Or do they represent something else? Times of day, stake levels that I'm playing, numbers that you stuck me with during the match? It would be interesting if you could still score separately on each amulet somehow, by playing games against me.

One cool thing about time of day is it seems to match the occult theme very well. Right? I'm kind of speculating here, but I think it would. It reminds me of three characters in American Gods:

The Zorya Sisters - The Zorya Sisters, relatives of Czernobog, are sisters representing the Morning Star (Zorya Utrennyaya), the Evening Star (Zorya Vechernyaya), and the Midnight Star (Zorya Polunochnaya). In Slavic lore, they are servants of Dažbog who guard and watch over the doomsday hound, Simargl, who is chained to the star Polaris in the constellation Ursa Minor, the "little bear". If the chain ever breaks, the hound will devour the world.

You could have the most valuable amulet only start after midnight PST. Aren't demonic rituals most effective at night?

LiteS · 2015-04-02 14:07:12

joshwithguitar wrote:

It's not 100% collusion proof - you can get a slight advantage with colluders by simply exiting any game that you start between another colluder as you know it isn't a game against Jason.

Saturating the game queue with colluding accounts doesn't really give much of an advantage, since you still won't know if you're playing against Jason. I suppose there would be a slight advantage to players who can play simultaneous games on multiple accounts, but each individual game still must be won. I think the mental challenge of having to split focus between opponents during simultaneous play easily balances any advantage gained.

jasonrohrer · 2015-04-02 15:14:48

Well, I was thinking there would be no stake cap.

The game would function as it normally does in every respect.

Hours of the day or something like that could be good for spacing out the amulets.

The question is whether Nate winning $1000 from me would put him in the lead by $1000 or only put him in the lead by 85 chips.

It seems like counting chips only, and not dollars, would change the game and encourage people to play lots of penny games... well, not if people thought I wasn't playing penny games at the moment.

If people felt like it was probably me proposing a $50 game that was on the board at the moment, and if we were only counting chips, I think everyone would avoid the $50 game, because it counts for the same number of points as a penny game, and there's a 5000x opportunity cost. Well, not really an opportunity cost, but a risk of loss. There's also a chance of winning $50 in addition to winning points toward the amulet. But... why not wait for a penny game against me? Or maybe take the chance, because you know that the penny games against me would be swamped with people trying?

It almost seems like, by making "dollars not count," we might give rich players an even bigger advantage, because they're the only ones who can afford to risk $50 to get 85 points. Whereas if you're risking $50 to get 5000x points, it's much more affordable.

Also, keep in mind that I can play however I want. So if I'm in a $1000 game with Doorman, I might bully him for the first few rounds, taking 10 chips, which is $100, and then leave. He'd then be down $40, both in terms of balance and in terms of amulet score. At that point, he'd be wishing he played a penny game against me, if dollars counted.

jere · 2015-04-02 15:48:52

I don't think counting dollars is a good idea. You're coming up with scenarios where it would be bad for high rollers, but they're kind of a stretch. If I'm Judge Doorman, I'll just do martingale betting starting at $100. If I lose, I do $200. Then $400. Then $800. Once I win a single one of those games (and if I can't win any of 4 games against you it's a moot point), I've probably won the amulet AND I'm up $100 in real dollars. Martingale is insane, but provided you can support a long run of it (you are actually rich) it makes perfect sense here because of the incentives.

Also, keep in mind that I can play however I want. So if I'm in a $1000 game with Doorman, I might bully him for the first few rounds, taking 10 chips, which is $100, and then leave. He'd then be down $40, both in terms of balance and in terms of amulet score. At that point, he'd be wishing he played a penny game against me, if dollars counted.

You can do whatever you want, but there's a question of fairness here. If you leave a game after 10 coins and you play other games to conclusion, the person who lost 10 coins is really going to feel cheated. The "last man standing" rule was nice in that sense. It's not like you need code to enforce it, but it'd be nice if you played games out (or at least played for a reasonable number of rounds).

Similarly, it would seem a bit unfair if you would be much less likely to play certain stakes. That's part of the reason I mention a stake cap. Without a stake cap, there's surely some value that is outside your limit. $999,999? It'd be good to know what that limit is for those that want to play really high.

I would think the contest should be attractive to the people who only want to put in a few bucks and having dollars count for the amulet is discouraging. At the end of the day, dollars do count because money is still going to be changing hands as normal.

jasonrohrer · 2015-04-02 15:59:49

Well, likewise, someone could bully me and walk away with a few chips.

Essentially, whatever strategies work to make money in the main game would work here.

The problem with martingale is that Doorman would have to know that I'm willing to ride the elevator up with him. I could get off on any floor, and it might not even be me. Though at some point, he could learn that it's me by watching Canto. Still, there's no guarantee that it's me continuing to ride up with him.

But yes, I can see that it would be discouraging if dollars counted. If not theoretically (because of amplified risks of big negative amulet scores), then emotionally.

As far as a maximum stake.... well, try me.

jere · 2015-04-02 16:12:25

I guess where my confusion stems from is I assumed that if the "play the game how you normally play" attitude is going to be respected, you would basically select games/stakes to play at random. So Doorman could just play games at whatever level he wanted without having to hunt you down and he would be just as likely to get a game as anyone else.

If instead it's our responsibility to figure out which stakes you're playing that changes how the game is played. I will admit however that the metagame of trying to predict not only what columns you're going to pick but what stakes is somewhat interesting.

As far as a maximum stake.... well, try me.

Yup. I figured you'd have to start ramping up the trash talking if we all know you're going to be playing the game.

jasonrohrer · 2015-04-02 16:29:10

CHOKE ON MY BOOMSTICK

Well, I don't think I could chose games randomly. I mean, I could use dice or something, but baring that, I am human and it is my money that's at stake in each game. I would just be playing normally. Also, I can't just pick random games to join, because then players who wanted an amulet would never join games themselves. Of course, I have to BOTH create games AND join games. And how can I create games randomly? No, I'd be naming stakes on my whim.

You know, there have been times in the past where players were trying to play a certain opponent or trying to avoid a certain opponent. Thinking you were playing Judge Doorman when you were really playing each other. Or trying to bait someone with stakes. Or that time that Nate put up the $6666.66 stake for days.

So, that's a normal part of the game. The stake game. The opponent game. Fishing.

What stakes are "Jason bait" that I won't be able to resist clicking on? And what stakes will I be proposing? And then there's the meta of other players trying to mimic me so that they can trick other players into thinking that they're playing me.

Also, last time I checked, I was really bad at the game. Or maybe I'm making that up. Who is Jeopardy Alcohol anyway? All I said was that I knew him... Does Judge Doorman really exist? Doesn't that story sound too good to be true? Andy Kaufman lives.

LiteS · 2015-04-02 19:14:12

jere wrote:

Without a stake cap, there's surely some value that is outside your limit.

jasonrohrer wrote:

As far as a maximum stake.... well, try me.

Careful not to get carried away guys, he is the house after all.

jasonrohrer · 2015-04-02 19:41:38

So, as I implement this server-side, which as Jere predicted involves very little code, I'm noticing something funny.

First, we have to track both coins won from me and lost to me (the net) so that people don't play recklessly against me. That means that people can go negative, and negative scores will show on the leaderboard.

But obviously, some players will never play me. They don't have a 0 score, they just have no score at all. So they don't beat the people with negative score, because they never played me. Though someone who played me once and tied me would have a legitimate 0 score.

The result of this is that someone could win an amulet by LOSING coins to me. This seems strange, because in terms of their game balance, they would be better off not playing me at all, but here they are on the leaderboard.

This is only an issue if less than 36 people have a positive score against me. (I'm not wedded to 36 amulets, by the way, but it's a good placeholder number). That actually seems pretty likely if I play well.

Unless anyone with a negative score against me simply cannot win an amulet. Then maybe I hand out less than 36. Maybe I cycle back through the list handing out the rest, and some people get two? That seems like a good solution.

jasonrohrer · 2015-04-02 19:45:07

Also, though you wouldn't want to play your own dummy accounts, collusion between friends who are both trying to play me is still likely. Like, Jere and Josh could talk in the chat room and determine if they are playing each other. If they are, they are certainly not playing me, so they should quit their game and try joining again.

Still, that's a brand of collusion that I can live with (does not give free points to anyone), and it is completely impossible to prevent.

Cobblestone · 2015-04-02 20:53:07

jasonrohrer wrote:

Who is Jeopardy Alcohol anyway? All I said was that I knew him... Does Judge Doorman really exist? Doesn't that story sound too good to be true? Andy Kaufman lives.

Good lord, reading this made me feel incredibly paranoid. >_>

jasonrohrer wrote:

So, as I implement this server-side

Wait, you're actually doing this? THAT'S AWESOME.

computermouth · 2015-04-02 23:12:04

jasonrohrer wrote:

Maybe there could be some way to elect 36 "high council" members who would each be defending one amulet during the week. Of course, that would make you all (Jere, et al.) ineligible to win an amulet, which would suck for you.

Hey Jason, in all seriousness, I'd gladly play as an amulet guardian. I just like playing the game. I win/lose against/to players across all ELOs. And for once I wouldn't be losing my own money

jasonrohrer · 2015-04-03 01:21:07

Oh, no, as an amulet guardian, I'm going to be playing with my real money.

So if I did go that route of electing 36.... real money play would be part of the deal!

But I'm currently leaning toward it just being me for a bunch of reasons.

computermouth · 2015-04-03 01:55:01

jasonrohrer wrote:

So if I did go that route of electing 36.... real money play would be part of the deal!

Color me intrigued. CM: Factions!? Clan Wars! Clan Wars! Imagine the rage, haha

CORDIAL MINUET ENSEMBLE

#1 2015-04-01 18:11:38

I found THE solution to the collusion problem

#2 2015-04-01 20:12:32

Re: I found THE solution to the collusion problem

#3 2015-04-01 20:28:31

Re: I found THE solution to the collusion problem

#4 2015-04-01 20:34:12

Re: I found THE solution to the collusion problem

#5 2015-04-01 21:13:36

Re: I found THE solution to the collusion problem

#6 2015-04-01 21:15:37

Re: I found THE solution to the collusion problem

#7 2015-04-01 21:51:52

Re: I found THE solution to the collusion problem

#8 2015-04-01 22:05:32

Re: I found THE solution to the collusion problem

#9 2015-04-02 00:48:17

Re: I found THE solution to the collusion problem

#10 2015-04-02 01:03:24

Re: I found THE solution to the collusion problem

#11 2015-04-02 06:14:41

Re: I found THE solution to the collusion problem

#12 2015-04-02 13:21:24

Re: I found THE solution to the collusion problem

#13 2015-04-02 14:07:12

Re: I found THE solution to the collusion problem

#14 2015-04-02 15:14:48

Re: I found THE solution to the collusion problem

#15 2015-04-02 15:48:52

Re: I found THE solution to the collusion problem

#16 2015-04-02 15:59:49

Re: I found THE solution to the collusion problem

#17 2015-04-02 16:12:25

Re: I found THE solution to the collusion problem

#18 2015-04-02 16:29:10

Re: I found THE solution to the collusion problem

#19 2015-04-02 19:14:12

Re: I found THE solution to the collusion problem

#20 2015-04-02 19:41:38

Re: I found THE solution to the collusion problem

#21 2015-04-02 19:45:07

Re: I found THE solution to the collusion problem

#22 2015-04-02 20:53:07

Re: I found THE solution to the collusion problem

#23 2015-04-02 23:12:04

Re: I found THE solution to the collusion problem

#24 2015-04-03 01:21:07

Re: I found THE solution to the collusion problem

#25 2015-04-03 01:55:01

Re: I found THE solution to the collusion problem

Board footer