How does the system work?

We introduced a small glitch yesterday which resulted in the markets being unusually unresponsive to money being placed into them. Actually in the longer term we are considering making them less responsive than they are, though possibly not to quite the extent we did.

In the course of fixing that error we somewhat overcompensated though, making the Yes/No questions ridiculously responsive for half of today.

Everything is now back to normal.

cheers
chris

Is that what happened to the leaderboard?

I noticed that the markets were going wild this morning. I am guessing someone gamed the system this morning and was able to go from nothing to $300k in one morning.

It's made the leaderboards change somewhat because it moved prices on the markets. Frustratingly not to my advantage I notice

It definitely makes the case for resetting the game a whole lot stronger. It'll take weeks to catch the top two for anyone from the rest of the top twenty. And how discouraging is it now to someone with $1500? They'd have to increase their net worth by 12,680% to reach where first is now!

Some people have gained from gaming the system, or in this case just blind luck in being able to take advantage of a wide-ranging system problem. At least 2 of the top 3 seem quite happy to have things reset, and I can see the reasons for feeling like that would be a good idea, although at the same time it's difficult for us to say that we are certain we have ironed out all the glitches in the system at this point, so it would be premature to do some kind of reset this early in the game.

I'm also aware though that we have thousands of users who have been plugging away trying to make an honest profit from legitimate trading, and by resetting them as well I think we are kind of kicking them in the face. It also doesn't address a potential issue going forwards of new users starting off with their thousand dollars and playing against long term users who have large (and eventually entirely honestly gained) fortunes.

We are certainly open to suggestion that allow us to address both these issues and keep everybody happy.

I think that the issue at play is that this is a public beta. Weird things will happen. Cheating loopholes will exist, be found, be closed. Bugs and server failures will happen, and hopefully solutions will be found. Experiments in liquidity and other market factors will occur, will have dramatic results, will call into question the fairness of the net worths. This is all understandable -- in fact, expected. You can't know what might need fixing until you see it happen.

But as a public beta, I don't think there is any expectation that this period of time is official (ie the real game). So I think that at some point where you guys feel the system is streamlined and fair, then it goes out of beta. That's also the appropriate time to reset the user balances. Game companies do this. They often run the beta until game release or a week or two later, and start up a new game as the official game. That's the best practice here, I believe. Everyone knows long in advance that this version will close at some point, and we all start equal again in the new one.

A second issue is within this version. People get upset, maybe for good reason, maybe not, when arbitrary events change the nature of the game. I was for obvious reasons shocked and upset for an hour or two, and really frustrated when I found out why. I'd asked the day before for warnings to be posted about changes to the inner dynamics of the game - it's the least we might expect. Also, I think that HubDub should come up with a policy about what to do about the results of such changes. This is a beta issue, as I assume there will come a time when changes like yesterday aren't being made - you've worked out a formula that works.

But pics4d should not feel he/she has to apologize to us for good fortune. I certainly had no suspicions of them. Nor should they have to return the money (because there was no policy in force at the time that said so).

But tinkering with the game in-play raises the issue of the (virtual?) contract that is made at the time of virtual-wager by each player. So there will be questions. And letting people keep their winnings into perpetuity (rather than just for now) is unfair, and discouraging to many current players and certainly to all new players.

I know there are new ideas for leaderboards that measure growth. But with the exception of anomalies like yesterday, I can easily predict (from experience) the amount of their net worth ($15,000), because that's around the ideal amount given the total number of players at this time. So I'll still be more interested in net worth, because this displays (hopefully) the long-time players who are good at the game. The quarterly gain (after the first quarter) will show the up-and-comers, and the weekly percent will show the lucky and the up-and-coming up-and-comers. Obviously these have their appeal to people in these various situations, but ultimately as it stands the net worth is the gold standard, the premiere league. So it matters that the amounts there reflect fair gains.

I think right now it comes down to how much HD seeds the market with at question creation. All the bets on Yes (for example) are pooled, and that amount is divided by the total of Yes and No bets to get the percent for Yes, and similarly for No. But think about betting on a virgin question. The odds don't go to 100%, which means there is "hidden" money already in the market. And on a 50/50 Yes/No question, where would the money come from to pay you back 200 on a 100 bet (pretend you are the only bettor on that question).

So if HD puts more cash into the question, split into pools of cash according to the initial likelihoods, then the market is less responsive, and if less cash is put in, the market is more responsive. I don't know whether or not money is paid back to HD from the market as it comes in from users. (Caveat: I'm supposing this is what's really happening. Official response?)

There are problems with this mechanism. HD is essentially underwriting low participation questions, and pumping far more money into the economy this way than they are from issuing new user accounts.

Think about this. The money that the two of you made yesterday, that was like 400K. But did a large number of other players lose (collectively) 400K? I don't think so. In the top ten (previous), I saw no change in the net worths of anyone else, which was surprising.