Topic: Just-Dice.com : Invest in 1% House Edge Dice Game - page 87. (Read 435457 times)

I wish there was a clearly best way of implementing variable risk for investors.

Can you explain why 'your way' is better than the way I've described?

I think the main difference between the two ways is that I'm saying that every bet should be shared between investors according to how much they are willing to contribute to the maximum profit on the site, and you're saying that investors should share the action only based on their amount invested, disregarding how much each is willing to risk.  Is that a fair summary of our positions?

Here are his stats on the new account:

wagered: 366.42000000
bets: 30
profit: -120.00000000
losses: 20
wins: 10

Dooglus what are you working on these days?

Yay, bitcointalk is back up.  Dooglus, have you thought about implementing my method for risk management of investors?

Well, in fact he divested and made a few gambles from a new account, but stopped suddenly (after he lost a few dozens BTC) and wrote on the chat:


And invested again.

Funny stuff if you ask me. And I have to say I'm glad to see Nakowa talking some sense, despite of the chance of recovering the losses being smaller without him gambling.

While I was away from the cold wallet, nakowa ended up with a balance of over 10k BTC which I was unable to cash out for him.  Instead he invested it, and asked me to block his account from betting, which I did.

So yesterday after mech's win, he was unable to use those 10k BTC to try to win his lost investment back.  He asked for a withdrawal while I was sleeping last night, but had reinvested it by the time I woke up.

I imagine sooner or later we'll be online at the same time, he'll withdraw, make a new account, redeposit, and try (and probably manage) to regain the coins mechs took from him.  The chance of turning 10k into 10.4k is pretty good.

additional variable = your vote

that is an additional variable needed in both cases

Dooglus,

Can you please refill the hot wallet?  Maybe you should leave more then a few hundred in there at a time?  It always seems to be empty.

thanks.

SL went down in flames. BTC price is down in flames... Is BTC gambling next?

Keeping track of what exactly?

I suggested: vote weight = sum over: deposits * duration deposited.

This information is already there (go to your investment history. nothing more is needed).

Not saying it isn't some work to be implemented (some interface for voting, code that calculates the vote weight), but: no, there are no "additional variables" that need to be tracked.

In the vote system, no one gets what they want (but everyone gets some average). In the market system, everyone gets exactly what they want.

Both introduce a similar amount of complexity (keeping track of additional variables for each user and the site as a whole).

You can't rely on a single source for generation of the roll outcome.

A better mechanism would be to have 2 servers controlled by different people with each its own seed (this concept can be expanded beyond 2, but my example uses 2 for simplicity).

For any user, both servers have their own server seed and publish its hash as usual. The player can provide his client seed and the nonce is generated as usual too. When a player makes a roll, server 1 computes hash(seed1, clientseed+nonce) and broadcasts the hash to server 2. Server 2 then computes hash(seed2, received_hash_from_1) and returns to server 1.

Server 1 is the regular JD server operated and controlled by dooglus. Server 2 in this example is a server controlled by a third party, be it some form of trusted escrow or a large investor.

The last server to compute a hash, knows whether a bet will win or lose and can in principle feign a network interruption to not transmit bets that have an undesirable outcome. To combat this, other server(s) keep a log of all bets for which the hash was transmitted, but no reply was received. Servers will retry to transmit their hash to the next server in case of a timeout and prevent users from placing new bets until the previous one has resolved.

In this setup, there is no single party that has sufficient information to predict bet outcomes. With servers keeping logs of bets they receive and transmit, the system can be audited to ensure that no single party is witholding bets with undesirable outcome. Obviously, this is a relatively complex system and it introduces additional latency and potential failure if a secondary server is disconnected. But this is just a rough sketch, one could extend the system with numerous servers and some form of redundancy.

This setup would make JD (or any similar game) provably fair for its investors under the assumption that the operators of the servers are all independent and not conspiring together.

I'm really liking the vote idea. Between 0.2% and 1%, vote weight is based on amount invested for what period of time.

EDIT: this vote is for max bet profit.

This would require dooglus to build a system to track vote weight per account in real time.

However an additional thing to add would be a time half life. So older time is less relevant past a certain point, perhaps after 3 months each month on has halfing effect.

An example: days 1 to 365 (site open for a year)

Example involves person A and person B.

Day 1: A has 100 bitcoin invested, B has nothing invested
Day 275: A divests, B now invests 100 bitcoin.

Day 365 vote weight:

A should have half the vote weight that B has, and every month more that B stays divested the time should half off instead of degrading like normal, because its old irrelevant time.

If no half life is implemented, then at day 365 A will have 3x the vote power of B even though B has been far more current.

I think the half life should not be too close to the current day because whales can just invest for a short period of time, but if after 3 months the time invested past that point becomes worth less and less, it will give current investors a better vote.

The site will need to count vote weight all the time, if B divests 10 bitcoin for 1 day this needs to be continuously monitored and weighed.

EDIT 2: The downside is that this is very complicated for something that is currently 1 variable that dooglus has set and most people are happy with.

THIS is stupid.

- you add a hell lot of complexity & points of failure
- you have to trust more people

what does "little more provably fair" even mean?!    It's either A (Rock solid, cause of math) or B (math do not proof the outcome of your 20 Satoshi bets)

So far it's A.


JD is doog's thing. either trust it (the math behind it AND doog for not running with the coins) or GTFO. End of story.

Yes, but he would have access to only one server seed, which would be worthless without the main server seed, and the other 3, 4, 5 you'll have distributed around....

and the whale as a big investor......?

The only way that I can see to make it a little more provably fair to investors is making a system of distributed server seeds, and giving access to each new server seed to a different big investor. That way you can even distribute the risk of holding everything on a single hosting service. So on each roll the central server asks to each seed server what's the hash for that roll of that user, and the seed servers answer with a corresponding hash which is then used to roll the final number.

Main drawbacks would be a lot of added complexity, maybe increased unreliability if you need all the servers up all the time, probably a cost/benefit ratio too high...

This would be a much more practical and fair way to set the max profit percentage.  Let every user choose between .25% an 1%. Then at any given time the max profit percent = the average weighted by the time-value of the votes.
Time vote value= Principal Investment Amount x days invested (gets reset to 0 with divestment).
Fixes to issues at once, fixes how max profit decided in a fair and simple way & gives a strong disincentive to rapid investing/divesting to day trade whales.

Or maybe not. It's not as if I never thought of that possibility. My suggestion was:

vote weight, per account := sum: amount invested * time invested

Which means nakowa or anyone else still can show up, invest for 5 minutes, vote and then divest, but his vote weight would be the same as someone who had a much smaller amount invested, but since the beginning of j-d. The biggest vote weight would go to those investors who had large sums invested, since the beginning, the way it should be, IMO.

What? Someone beat my max bet? (Which I haven't made.) I missed some action again huh.