CORDIAL MINUET ENSEMBLE

Here you go: http://cordialminuet.com/incrementensem … hp?pid=521

As many people have pointed out, the contest encourages non-amulet holders to play very recklessly and amulet holders to play very conservatively. Amulet holders are risking their amulet AND up to $3. Non-amulet holders are risking a single penny. I don't know if this is a problem, but it was very clearly there. And this also meant that the number of games you played seemed far more important than your skill.

There's the obvious strategy of securing a win 2 hours before the comp ends and just holding it. Nothing wrong with that, but it's probably going to happen with many amulets.

One person who shall remain nameless intentionally drew out games with the stated goal of encouraging non-amulet holders to bail out. This is the first time I've experienced this and it was incredibly frustrating. I could see someone drawing out a handful of rounds to 2 hours (and only losing a couple dozen coins).

One interesting thing in this contest was targeting specific amulets. Most of the time, the top amulets weren't held so it wouldn't make sense to play for an amulet if you thought you had little shot of out competing everyone else for the top spot. Josh had 3799 points at the end, so he must have won at least 20 games with that amulet. If it's the last few hours of that contest, that seems pretty hopeless.

So at one point I knew I had a decent chance at the 3rd amulet and what I had to do is wait for the other two to get picked up before going after mine. In the real contest, that probably won't be a viable strategy. The way I see it working is that if you get your hands on an amulet, you will have to treat it like the only time you will ever see that amulet and make the best of it.

Cobblestone, while I can't rule out a bug, I can likely explain what's going on here. Even if it is sort of a cop out.

You gave me some tweets as examples. What happened in those cases is that you were getting part of a "bundle" of tweets. The bot runs every 5 minutes and, during periods of high activity, several people are involved per 5 minutes. Obviously, it has to split up all these announcements if they grow bigger than 140 characters. So the "random" one you got should have arrived at the same time as another one that involved you.

It's probably just lazy programming and I could fix it if necessary, but I didn't think it'd be that big of a deal.

And actually this behavior will tend to be useful. You want to get @ when your "followers" play but you also want to see who they played, which might be the person in the random tweet.

Hope that helps....

Yea these games are taking forever.

http://techcrunch.com/2015/03/11/rose_gold/

I haven't looked too deeply into this, but I thought it might be semi-relevant:

Sounds like you are giving away more/purer gold than is in Apple's $10,000 watch.

Ah, you're bold to go in with a hand like that, but looks like it worked out. Welcome, .txt!

cullman, I was saying the exact same thing on collusion, which is why I want more details. If it's set up the way I think it is, I might try to "collude" just to prove it.

.., it would probably be simpler and in some ways more fair to have a single score for each type. Simpler for many reasons but mainly because you can have 3 leaderboards instead of 36.

The pros of having individual scores?
-It makes each amulet feel more like a real object rather than a variable sitting on a server.
-It might convince new players they have a better shot. If I get a gold amulet, I don't have to beat the best 6 players in this game to win. Instead I just have to win on this amulet. The flip side is I could have the bad luck of gaining an amulet on which someone has already racked up the most points.

This is outdated unfortunately. I wrote this when the ante was 1 coin every time. Storeroom basically summed it up, but to give you some context:

Betting 10-20 coins on every round was Judge Doorman's original strategy. Now imagine you're playing a $100 game and your opponent bets 20 coins. You match it and you're on the line for $20! Probably more than that because if they bullied on the first turn, they'll probably do so later on too. But it's not that hard to counter and that's what I was referencing. I just folded every time I had a bad opening pick. Sometimes, it was clear that not only did I have a high number, but I clearly had a huge advantage. At that point, I would match. I could fold 9/10 hands and still win the entire thing consistently. That's how Nate and I won a good bit of our money. Anyway, Doorman got much better, he stopped betting like that, and the ante was changed.

Fighting bullies is tougher these days, but the basic idea is that a bully betting huge amounts every round indicates they're probably bluffing often. You have to be willing to take some calculated risk and jump in when you have even a slight advantage.

You should conduct a thought experiment: would I ever be willing to match an all-in on the opening bet? If the answer is no, an opponent can bully you with impunity and maybe you should think about lower stakes. In other words, if you're not comfortable risking a $1 on one hand, go down to $.10 stake and an all-in there is roughly equivalent to risking a few coins at the higher level.

^^Clever insight. Would never have thought about making the coins not count.

It works. The random part is good. The problem here is: the winner of the amulet doesn't get the money. What if it's Judge Doorman putting up a $500 game, he wins it, and is promptly told he hasn't yet won anything. He's won a chance at the prize, but the reality is most of the people that get their hands on an amulet are going to be receiving $0. He's going to be pretty bummed: he was risking $500 of his own money, he played well and won, and didn't get the usual reward. You could explain this rule up front, but it changes things for everyone playing the game.

I don't get the "changing player behavior" concern. If you implement what you're describing, EVERY single game played by a non-amulet holder is a chance to win an amulet. Whatever player behavior you were worried about with people joining a fixed stake games (because they would "know" there's a chance at an amulet) now applies equally to ALL games.

Unless your concern is the stake is too low or something.... but this new idea would just incentive most people to trawl for amulets at 1c anyway.

The only difference with these approaches is that with a fixed stake, you can opt out of the chance at amulet. I suppose you could say if tons of people don't want to play amulet-chance games, it will skew the chances of getting one of the players who do. But I think practically everyone playing during the contest is going to want to play the contest. The opt-outs are probably the exception.

I get that you prefer playing fast, but many of us don't. I think your best bet is to find someone like minded (through here or the CM chat room) and play against them under the conditions of a gentleman's agreement to take shorter turns.

Personally, I'd hate to have to make a move 5s within my opponent. Besides the fact that I am slow thinker, every game would have the potential to turn into a coin toss: your opponent just all-ins and immediately selects moves at random and you are forced to make a guess without time to evaluate. No thanks.

I also think you're underestimating the complexity of this game. There may not be the possibility space that Go has, but Cordial Minuet has a layer Go lacks: hidden information and trying to read your opponent. Sure, I only have 12 choices on the second turn. But I also have to consider 60 permutations of my opponent's row at the end of that turn (1st turn red row and 2nd turn red/green rows). I have to consider what my opponent is thinking, what they're thinking I'm thinking, what they're thinking that I'm thinking that they're thinking, and so on. I don't see how you can be considering all this in 5 seconds.

If I run into one those smart players you're talking about, I'll point them here.

Re: uniquifying, Having seen people complain about the existing, tiny barriers to entry (entering your CC, deposit fee), I definitely wouldn't add any additional friction. Anyway, it still doesn't stop a group of people from working together.

Basically, I think the fixed stake thing with delay (for everyone) is good enough and you shouldn't lose sleep over it. Hard to predict what turnout you will see, but Steal Real Money got 800+ people on the last today. With several hundred people playing everyday, with all of them going into the same bucket, colluding will be a losing proposition, not to mention a huge hassle.

I'm not sure on that myself. I keep going back and forth. As long as you encounter the amulet often enough (and keep getting motivated to play your best), I guess afterwards might be better. If I play 30 games and never seen an amulet, I'd probably get pretty bored. I know you don't want to have to cast 1000 of these things, so maybe you consider having a good number of cash-only (like $10) amulets.

Also, you were probably planning this, but it'd be good to see after a game if you had a chance at winning the amulet, but lost it.