??????
You are not logged in.
I can understand the desire to abandon it all at this point but I'm not sure it would be that hard to fix with a few comprises and a different approach.
I think you ran the second tournament much too fast. After the first we outlined that there was lots of potential collusion exploits. You added some hasty patches and quickly ran another test. If instead you'd posted the rules and the server code and asked "how do you exploit this" you'd have had a large list of problems from us within 24 hours. Many of the problems had already been noted before the comp. The only reasons exploits didn't come up faster in the test is because we were being fair.
If we ran through a number of fixes and threw them to the players to figure out an exploit I don't think it would take long to get to something workable - partly because I think we are almost there.
So, currently there seems to be two lines for exploitation:
maximising amulet pickups and therefore points
maximising amulet games - hopefully against your colluders
The first is very easy to deal with. Get rid of the 200 points given when you pick up an amulet. It shouldn't happen too often anyway. This completely removes any chance of this exploit being a thing.
The second is more tricky. The only way I can think of doing the exploit is this:
play lots of games against colluders where one player leaves/loses quickly
lose to any colluding amulet holder
The only way I can see for doing this is:
- chose an unpopular stake
- have two accounts join the stake at the same time
- if both join each other one leaves
- if one is left waiting treat the other game as an amulet game.
- if brought into a game with a colluder, lose
Now, if this is the only exploit available then it seems there are lots of ways of dealing with it, many of them simple:
- have it so that facing the same opponent within X minutes disqualifies that matchup from drawing into an amulet game.
- have it so that after the start of any game a player will not be considered for amulets games for 5 mins (new games during this period will not reset the timer though).
- put a cap on the number of games per day that once passed makes players no longer eligible for further amulet games that day. (this would also restrict the power of bots and people with far too much spare time).
These are just a few but they all seem to do the job on their own and I'm sure there are plenty more.
The other thing I'd recommend is this to basically make it so the contest isn't running - no amulet games occur - if the player base gets too low at any point in the day. Hopefully this won't happen very often. But it is a simple fact: the idea really only works if there are enough active players that any player who wants to can start a game within a couple of minutes.
Anyway, I hope I have managed to make you rethink things, but if not good luck with thinking up something new!
Caravan: So in the end were you holding onto the amulet while the other two played amulet games against you or were you dropping the amulet and picking it up again?
Jason: Regarding the question to Caravan, if it was the answer is first one then I'm pretty sure the leaderboards were lying to me at the end until it was too late to do anything.
What I saw with regular refreshes:
I lost the amulet to JA
JA dropped the amulet Caravan Picked it up and made 200 points
Caravan drops the amulet
From this point until I saw Caravan with the amulet ahead of me in points the amulet was always unheld
I tried joining a few .11c games during this time getting a wait screen each time. Thinking that I possibly sent them into an amulet game I would check to see if any amulets were held. None ever were.
I never noticed Caravans score increasing - (though it is possible that this went under my radar as I was focusing more on if it was held? This seems unlikely...)
At one point the server seemed to go down for a few seconds.
and then suddenly out of nowhere - bang Caravan held it and had more points than me, giving me no way to make a comeback.
One question: Is it possible the leaderboards were not updating properly and gave me a false sense of security at the end? And that CD was holding the amulet the entire time? I did noticed that the server went down for a bit.
Or did Caravan Distributor just get extremely lucky?
It's the last hour of the competition and I end up losing the amulet to JA. After that I watched the leaderboards continuously and never noticed anyone holding the $100 amulet.
Suddenly I noticed Caravan Distributor held the amulet with 1640 points on it - 800 more than when I had dropped the amulet 20 mins ago.
Checking Cando Delirium caravan had played against two opponents in that time Chaos Building and Nursery Malice.
Now, it is possible that CD just got extremely lucky at the end and faced off against 2 opponents who both lost very quickly in the last hour of the contest, but I think it is more likely that they are alts or friends of CD. I don't know how CD was able to drop and pick the same amulet in such fast succession though - I thought that was supposed to be fixed.
Has anyone else noticed that the end time of the contest keeps being pushed back?
I just lost the $100 amulet because I expected there to be a gong and there wasn't so I timed out on the "accept amulet game" screen...
You need to add a gong.
Also, JWG, you played a lot of games, what stakes were you typically playing? Since the stakes aren't revealed to the amulet player, I'd imagine $3 games wouldn't be as efficient as penny games.
I played mostly <= 10c because that was what was on offer. In the real comp though I will be trying to play as many $3 games as possible. This is for two reasons:
- Now that game stakes will be revealed to the amulet holder $3 games will increase my chance of the amulet holder leaving in amulet games giving me a free amulet.
- With a large influx of players I imagine that the majority of games played, even in the $3 range will be against new easy to beat players and so the larger the stakes I play each game the more profit I will make overall.
One thing I noticed, which shouldn't be too big a problem, is that after I had obviously dropped an amulet according to the leaderboard it still showed up in my client, and when I clicked "new game" it still showed "play amulet game". It only corrected myself when I went back to the main lobby.
How I got the top place for the two top amulets:
Played lots of games during low activity dropping an amulet as soon as I picked it up. Sometimes against other players doing the same, sometimes against new players that I didn't lose against and so the amulets tended to stay in circulation ready to be picked up again. During popular times I'd hold onto any amulet I got a hold of (well, at least the top two) and play amulet games.
Of course this strategy won't work at all if there are always lots of active games. But with low activity this kind of behavior is what will win it, if not full blown collusion with players or alt accounts intentionally leaving in order to very quickly boost another players amulet points. This would have been extremely easy in this competition - I'm personally glad no one decided to do it. If you have another test comp I wouldn't be surprised if someone tries this though.
So, overall what will make or break the comp is whether you can get consistently have lots of active players. If at anytime during the day the number of active players drops below or close to the number of amulets then it will be very open to collusion. Basically you want all the amulets to be held most of the time.
I'm personally not too worried about this happening at the very start of the competition, but I do worry about it towards the end - especially if it goes on for more than 2-3 days. Also, I wonder how many people you'll get playing from my part of the world - living in Australia definitely gave me an advantage in this comp.
Perhaps one preemptive solution is to have the number of active amulets restricted to some ratio of active players. So if there are 36 or less active players then maybe only one gold will be able to be picked up which will mean that winning a game won't guarantee a pickup and 200 points.
Another possibility is to have players lose points when they lose the amulet - so you get 200 points when you receive it but lose those 200 when it is dropped or lost. This will mean that picking up the amulet and dropping it has no net gain - you need to actively play amulet games to get points. This combined with restricting multiple amulet games with the same players should severely restrict collusion opportunities even with a low active number of players. Or even simpler you could just revert to there being no points gained when you pick one up from the pool - there should be no need of this when there are lots of players and games.
Congrats to the other amulet winners - Claspa, Jere, Cobblestone, Context Fabric and Ano!
Josh, yeah, I think it could work to pull a player away from a game that's about to begin. A random delay would be needed before picking a such a game to pull players away from, of course.
Good to hear, I was getting a bit worried about the collusion problems that Thomas brought up as well.
You're right that if the amulet holder is shown the stakes at the start, higher stakes will have a greater chance of scaring them into leaving the game and dropping the amulet. But won't higher stakes also be less likely to find a joiner, so be less likely to end up in an amulet game? I.e., it wouldn't be a viable strategy to propose high stakes ($3!!!) in hopes of scaring the amulet holder, because the game would sit there, unjoined, and thereby be ineligible to be pulled into an amulet game (according to your suggestion).
Well, this all depends on how many players there are. Just say there were lots of games and the average wait time for a 1c game was < 1 second while the average wait time for a $3 game was 30 seconds. The majority of players might still find $3 too high, but 30s isn't really that much of a wait given the average length of a game. 30s is enough though to give a much higher chance of being brought into an amulet game under your old scheme though. Of course, as this time increases there will come a point where it is no longer worth the wait, but we can't really know how frequent games will be at this stage.
Unless the wait time is prohibitive, I will personally be tempted to sit on $3 anyway because I imagine that most of my games will still be mostly against new players and my odds of winning will be high. This does mean that if game stakes are transparent then I will have the added bonus of amulet leavers, and again this kind of strategy is only something that someone with a decent bankroll can pull off sustainably. Still, perhaps this (slight?) advantage to me is not as bad as amulet players unknowingly being playing $3 stakes against me.
I'm glad that things are moving forward, but, just to clarify some things:
You're still planning to use a system in which a random staked game can be made? I cannot see any problems with the alternative of pulling a player away from a game that is about to start and, again, I see three advantages:
- games will tend to be around the median game stakes
- there is no advantage given to waiting around for odd valued high stakes games to pop up.
- there is no advantage given to posting a less frequently played stake in order to increase wait time, and therefore the chance of joining an amulet game.
All of the corresponding disadvantages of your current system favour high bankroll players.
Secondly, you've decided to ditch the idea of making the stakes transparent to the amulet holder? If not, you're fine with the advantage given to higher stakes games in terms of forfeited amulets?
Josh, what is your suggestion?
That we pull either a game creator or a joiner out of a game that's about to start, and match it with an amulet-holder who is waiting. But have some limit on it, like $5 max?
So the amulet-holder still needs to have at least $5 in their account to play an amulet game, but they will never create a game themselves at random stakes (with expected value of $2.50, which is high). Instead, their fate will be decided by the current stake-climate in the game in the sub-$5 range.
Yep, that is what I'm suggesting. Combined with letting the amulet player know the stakes this will mean that they most likely will be playing low stakes, but if they are unlucky and get drawn into a high stakes game they can opt out (for a leaving fee, though maybe this could be waved somehow).
But.... this again gives an advantage to playing high stakes games as you are more likely to have an amulet holder leave when they join you. I would be tempted to play lots of games at the highest amulet stake as many players would not be willing to risk it if they were joined to me.
When it comes to rising stakes games, I agree with storeroom: if only the maximum and not the minimum stake increases then this gives a strong incentive to play low stakes as it will ensure that you can join all amulet games, especially if the stakes reset when the amulet changes hands. Also, lower stakes amulet holders are less likely to have held it for long, and are therefore less likely to be good.
If the minimum stake rises then you again have a system prone to collusion, especially as the stakes get high.
Jason, I think instead of having amulet holders join/create a random sub-$5 game you should match them to a player about to join a game as you outlined in an earlier post. This, I think, would fix most of the problems faced by the suggested system.
Firstly, it would mean that instead of having a high chance at a $2.50 stake game you would instead on average join a median stakes game, which I imagine would be much lower. Sure, there would still be a chance of joining a $1+ game, but as long as these are much less common than sub $1 games this would be pretty rare. Still, under this system there is no advantage to playing lower or higher stakes in order to get into an amulet game.
Also, under the current proposed system I would be tempted to just sit around waiting for odd priced $1+ games (with stakes like $3.86) and just join them as I think it would give a much higher chance of joining an amulet game. If I wanted to be lazier I would just set a $5 stake in order to increase my waiting time, and thus increase the chance of an amulet game joining me (and giving them a nasty surprise when they lose $5!). These are both things that most players will not have the luxury of being able to sustain, giving high rollers an advantage.
Under your earlier proposed system of stripping players away from games, none of these things would be problems.
Josh, if your goal during the tournament is to make money.... hmm... I guess any amulet structure that encourages low stakes won't be good for you. But you would be a counter-example for your money pool idea. If a good player played loads of high stakes games against low-stakes players and won most of them, they would suck the money pool dry.
I don't think this would be a real problem in my system as: 1) amulet games won't come up that often, and 2) those holding an amulet will tend to be stronger players.
What if it was just "a random stake of $5 or less." and you had to have at least $5 in your balance to play an amulet game? Maybe $5 is too high in this case. Then you'd have NO control over your stakes, there'd still be a whole range of low-stakes games on offer. Maybe it could be a random stake of $2 or less.
Hmm, this could mean some unlucky player just learning the game and given an amulet suddenly losing $2 in their first amulet game.
And now, if the system was closest to a random sum from 0-$2 this would encourage play in the $1 region, which would lead to a bunch of players soon not having the money to play amulet games. And if very few players played in the $1-$2 stakes area then this would be the best place to hang out to get an amulet. So it would be better just to pull someone from a randomly selected game starting in that region.
I've just thought of one further annoying thing about the current system.
CM is a game in which as the stakes rise, the level of play tends to rise as well. This will mean that the higher your bankroll (up to $50) the more difficult your opponent will be on average when you have the amulet. This gives a strong, and annoying, incentive to reduce your bankroll to as low as is practical for the completion. I would personally be encouraged to either cash out or create a second account for the competition rather than just play on as usual.
I guess I have some selfish reasons for not liking your system. Times of great player influx are generally great times to play if you actually want to make some money playing, but your system will strongly encourage games that are < 50c as these are far more likely to be amulet games. I would have to make the choice between maximising my in game profits or maximising my chance of an amulet. And that is only if there are enough $1+ games for that choice to exist.
Also, if players are encouraged not to play 1c games you still haven't given a solution to the problem of players intentionally reducing their bankroll to 19c once they have an amulet in order to maximise their chance of meeting their 1c playing colluders.
cullman: Even with no stakes amulet games the vast majority of games you will be playing during the tournament will be for real stakes and you won't even know that you are in no-stakes games until you actually have an amulet. And then you will be playing to keep the amulet, so it won't feel like no stakes.
Josh, as cool as the lightning bolt was about "hey, we could actually match two people up who are trying to play different stakes," I've become convinced that anything bordering on funny-money is a bad precedent to set here. We're tracking balances down to the 100th of a penny here. Money is hard and real in this game. The person you're playing is always in for the same amount that you are. I think it's important to keep it that way.
I personally think the system here is excusable in this case as it only affects a certain limited number of games as part of the competition. It is not something you will sustain, and so it is simply an additional prize pool in the competition. You of course would not sustain such a pool after the comp ends. Amulet games will already be strange and non-transparent anyway and I don't imagine players complaining about winning/losing money in exactly the way they expect to given the stakes of their game. Money remains hard and real - you are just betting against the pool rather than the other player.
What exactly do you mean by funny-money?
What you're suggesting is good: that I would create a buffer pool to pad out "missing" winnings of some matches with the "extra" losings of other matches. But it's still kind of funny. Imagine if you matched your friend and talked about it later. "Ha, I won $18 bucks from you!" No, your friend would say, I only was in for $5"
If you met a friend after in this case, it wouldn't seem that strange because it was an amulet game and it is quite obvious from the amulet holders point of view that they were matched randomly and would not know what stakes they were playing.
In order to guarantee that you're playing only penny games when you hold the amulet, you need to have a sub-40-cent balance. But how would you get the amulet in the first place with such a low balance?
As long as you are also able to play non-amulet games it should be pretty easy to lower yourself down to 40c.
Also, you could cash-out down to 40c.
Also, regarding the incentive to play low stakes games in hopes of being paired with an amulet holder, it depends on how we do the pairing.
If we randomly pick from the existing games that are below 5% of the amulet holder's balance, then yeah, proposing penny games would make you most likely to get picked (because EVERY amulet holder could play you).
But if we pick a random stake first for the amulet holder from their whole 5% range and THEN look at the existing games for a match, proposing only penny games won't work as well for non-amulet holders.
The most attractive stake then becomes 2.5% of the average bankroll in the game, whatever that is, because it's the value that's closest to the most values in the 5% range.
So you're proposing that a value is chosen randomly from within the 10% and the amulet holder will join a game of that value if it exists? And if it doesn't it will try a different number until it finds one? Or are you proposing that if the value is not found it will go to the closest one? If it goes to the closest one then you might get a strange meta game going on where players are picking numbers that are inside the other numbers that other players currently have active. There will definitely be tactical ways to open games in order to maximise the chance of meeting another amulet holder. I personally don't like this - players should be rewarded for playing the game as usual, not picking strange and low stakes in order to increase their chances.
If you really don't like the idea of a pool/buffer, I think the next best option is simply to have amulet games play for no stakes and add an opt-in check-box on the main page in order to allow for amulet games. When you first tick the checkbox the game could make it clear that if you are matched with an amulet game you will not know, and no money will change hands. If you want to play a high stakes game without the risk during the comp you could always just uncheck the box for that game and recheck it after.
In this system you could even have it so that the amulet holder names a maximum stake as I suggested above, so that if the chosen game is higher than that stake then the game will simply be for the max-stake.
I'm going to repost some ideas I put up on the "Thoughts" page because they seem to have been buried without any responses and I would be interested in peoples opinions. I think there are big problems with the way you are currently selecting amulet games and think my earlier proposed solution would fix it.
- One problem I see is that a player could intentionally have an account with a low overall bank balance in order to force all amulet games into 1c games. If their alts or fellow colluders always then start 1c games it will give them a decent chance of joining a game.
- Another problem with this system is that those who want to play amulet games will be rewarded by playing low stakes - as the stake goes up the chance of joining a game will become much smaller. So this could strongly affect player behavior (perhaps countering the above point because everyone could be playing 1c games. This is not something you want though).
So what I propose is to go back to assigning the amulet holder to a completely random game, stealing a player away from a game they are joining. This will ensure there is little impact on gameplay for non-amulet holders and allow for players of any stakes joining amulet games. My earlier idea was that by adding a small additional prize pool you could make sure the game stakes are still relevant to the game:
"One further possibility, if you don't mind increasing the prize pool a little, is to have a pool from which you compensate players who join amulet games until the pool runs out. So if you join a $20 game and get pulled into an amulet game and end up winning with 180 chips you will get $18 from the pool as well as the amulet so long as the pool has $18 remaining. You could also have it so that money wagered that would normally have been lost in an amulet game feeds back into the pool up to the starting pool value. This would create a slight intensive to play higher stakes games while the pool is full, but I don't see that as a problem. I would not be surprised if a $100 pool was never exhausted in a system like this, and so the nature of the game would not change very much for those who are brought into amulet games.
Further still, amulet holders could nominate a value whenever they start a game to set the maximum money they are willing to wager. If they end up being paired with a lower priced game, money transferal is treated like a standard game at the lower price with no money going in or out of the pool. If the amulet holder is paired with a higher value game then the payout for them is treated like a standard game based on the higher stake. Money would only then have to come from/enter the pool to make up for the gap in winnings for the other player. So, in this final system if you start a $20 game and you are matched with an amulet holder who is willing to wager up to $5, then if you end up with 180 chips and they have 0 at the end you will take $5 from the amulet holder and be given a further $13 from the pool. If you lose all the chips and your opponent ends up with 180 you will give them $4.50 to them and a further $15.50 to the pool."
I think the random opponent assignment idea for the amulet holder can be saved by simply adding an opt-in check box on the main lobby page that once checked could mean any of your games will be drawn into an amulet game without you knowing. The first time it is checked it can bring up a page explaining what you are allowing. This would mean that if you want to join a high stakes game without any risk of an amulet game you could simply uncheck the box before joining. And yes, the idea that it will pull a player away from a game that has just started to play you instead fixes the problem nicely as it encourages playing as many people as possible.
One further possibility, if you don't mind increasing the prize pool a little, is to have a pool from which you compensate players who join amulet games until the pool runs out. So if you join a $20 game and get pulled into an amulet game and end up winning with 180 chips you will get $18 from the pool as well as the amulet so long as the pool has $18 remaining. You could also have it so that money wagered that would normally have been lost in an amulet game feeds back into the pool up to the starting pool value. This would create a slight intensive to play higher stakes games while the pool is full, but I don't see that as a problem. I would not be surprised if a $100 pool was never exhausted in a system like this, and so the nature of the game would not change very much for those who are brought into amulet games.
Further still, amulet holders could nominate a value whenever they start a game to set the maximum money they are willing to wager. If they end up being paired with a lower priced game, money transferal is treated like a standard game at the lower price with no money going in or out of the pool. If the amulet holder is paired with a higher value game then the payout for them is treated like a standard game based on the higher stake. Money would only then have to come from/enter the pool to make up for the gap in winnings for the other player. So, in this final system if you start a $20 game and you are matched with an amulet holder who is willing to wager up to $5, then if you end up with 180 chips and they have 0 at the end you will take $5 from the amulet holder and be given a further $13 from the pool. If you lose all the chips and your opponent ends up with 180 you will give them $4.50 to them and a further $15.50 to the pool.
The best option to limit collusion while altering player behavior the least is to just have a cap on games that could count as amulet games. Say $5, where any game for $5 or under could potentially involve an amulet, but $5.01 and higher guarantees the game is NOT for an amulet. That cap could even change in the middle of the contest if there was some good reason to change it.
In this alternative system, setting the top price for a amulet game might be tricky. You want to set it in such a way as to not discourage people from determining their own stakes but also at such a level that at the highest stake there are still lots of games being formed/joined. Perhaps it could be a variable rate that is displayed on the main lobby page. The problem here is determining the rate considering the fact that the rate itself will strongly affect player behaviour. The best idea I can think of is to have the rate increase slightly every time a game is joined at or close to that rate and decrease over time when nobody plays at it.
Also having the amulet taken away for timeout after few games against slow opponents would be infuriating and perceived as unfair (even if the rules are clear), aggravated by the fact that you cannot leave an amulet game while in advantage without loosing it.
One way of making it so that slow games are less annoying when you hold an amulet given a time limit on holding it would be to have it extended each game based on the amount of time spent waiting for the other player. So every second between the time make a move/bet and receiving the opponents move/bet would be added on to your time.
--Amulets handed out initially based on how many games you played until the end in the last 24 hours (the same way we count amulet points, except you don't have to win a game to have it count).
This gives a big advantage to those with bots when it comes to starting with an amulet. I would not be surprised at all if Clock Form ended up with one to start. And if timed out amulets are common and are given to those who are playing the most then bots might end up being handed lots of them (although I guess you could restrict each account to only one free amulet).
One issue with random game assignment: To increase the chance of me playing in an amulet game I might put the stakes outside of most peoples range. For instance, if I put a $100 game up, it's unlikely anyone will join the game, though I now have a game up waiting around for the next amulet game that randomly attaches itself to it.
Also, to throw away an amulet would you just join a game and leave straight away? This might be odd behavior, but I guess the other player gets a bronze amulet they won't complain too much. Unless they will just go and throw it away themselves...
If you wanted to avoid this kind of behaviour but also keep it fairly simple you could allow for multiple amulets to be held and just have it so that only the highest value amulet is active at any one time. So if you have a silver and you win a gold, the gold becomes active and your silver streak becomes dormant until you lose the gold amulet. You can still restrict gold vs gold games etc.
Are there ways to check to see if a card being used is a gift card? Perhaps address verification? Do gift cards have addresses? If so you could just not allow them or at require that people using them provide additional identity verification.
I think it will be quite hard to prevent someone with multiple accounts getting an advantage without unwanted side effects like restricting the price range of amulet games. So perhaps it is best just to make find methods of making it prohibitively difficult.
I still think there is no need at all for the 200 coin leaving penalty. It changes the game unnecessarily for people who happen to join an Amulet game, which would be especially bad if you didn't know you were playing one, as has been suggested.
Why not just work it as I said before: if you leave a game whilst holding the amulet you lose the amulet to your opponent. To win an amulet you must either take all of your opponents coins or have them leave. This has the same basic effect re:amulets while meaning that you can still treat an amulet game the same as any other if you want to.
Edit: To put it simply: if you leave an amulet game early, your opponent ends up with the amulet. Otherwise the games are no different to normal.
I personally like the idea of having hidden amulets. Even if you had a 200 coin leaving penalty this would not stop people joining many 1c games and leaving until they find an amulet game. The thing about a competition like this is that it will affect the way people play, but there is nothing really wrong with that. Encouraging people to play out full games during the period would make things interesting. Sure, a game that you know is for an amulet would be more intense, but if any game could be for the amulet then EVERY game would be more intense.
One big problem with my idea above is that it is very vulnerable to collusion as it is about as far from zero-sum as you can get: player one wins 1 token so now has 13 = $409.6 gain while player 2 loses their one token = a 10c loss.