Bitcoin Institute

Bitpredict Update Thread

40 messages BitcoinTalk kiba September 16, 2010 — December 29, 2010
kiba September 16, 2010 00:55 UTC Link to this message

I will post daily developer log of what I done for bitpredict, a bitcoin prediction market site. However, today I am too tired to work after returning home from college. I might or might not got anything done tonight. We’ll see.

kiba September 17, 2010 05:01 UTC Link to this message

After realizing how late it is, I updated to rails 3 and restart from scratch the bitpredict project directory. A disappointing start, but the key is progress.

If you can move, you will eventually get to your destination no matter how long it take. So, my goal is to alway be somewhere further along at the end of the day rather than be stuck at the beginning of the day.

kiba September 18, 2010 05:54 UTC Link to this message

I got a bit further along today. Finally got the test server from the rails directory running. My goal is to make much more progress than that tommorow. I also attempted to set up my netbook for development. It was an utter fail.

kiba September 19, 2010 04:58 UTC Link to this message

More progress: figured out how to generate static pages in rails.

kiba September 21, 2010 04:43 UTC Link to this message

Yesterday, I didn’t do anything. This is bad.

Today, I only modified the home template. I think I’ll do something with a bit more substance tomorrow. Perhaps, a generic registration system will be in order.

kiba September 23, 2010 03:40 UTC Link to this message
Quote from: Babylon on September 23, 2010, 3:39:07 AM UTC

Yay!  Bitcoin prediction market.  I am glad to see this coming together.  I wish you luck, let me know if there is anything I can do to help.

Ugh. I missed a day. No excuses there.

kiba September 23, 2010 04:13 UTC Link to this message

Very tired person from college. All I did today was beginning the use of an authentication framework and generating a user model.

kiba September 23, 2010 15:52 UTC Link to this message

I created a register link, fix spelling, and figure out how to do routing. Today is the most substantial progress we see.

Let see what happen tonight.

kiba September 24, 2010 06:23 UTC Link to this message

I did another session of coding tonight. I created an empty template page for new user, wrote unit testing code, and generally making sure that you can’t actually have empty username.

kiba September 24, 2010 20:08 UTC Link to this message
Quote from: S3052 on September 24, 2010, 8:02:17 PM UTC

sorry if this question was asked already: what will the    bitcoin prediction market actually predict?

People put up a bet on things that they will think will happen. For example, who will get elected in 2012 US presidental election. Will the republican take over congress? The marketshare of the search engines. It will be mostly be something specific and easy to verify.

kiba September 25, 2010 06:11 UTC Link to this message

Well, I was busy the whole day with another web application, a MVP if you will. I was close to completion so I decided to pour all my effort into getting it online. With any luck, I will integrate bitcoin as a payment option.

As for this project, it will resume as soon as I fixes things that went really wrong and reveal the MVP for feedbacks.

kiba September 25, 2010 06:26 UTC Link to this message
Quote from: S3052 on September 25, 2010, 6:18:41 AM UTC

sorry if this question was asked already: what will the     bitcoin prediction market actually predict?

People put up a bet on things that they will think will happen. For example, who will get elected in 2012 US presidental election. Will the republican take over congress? The marketshare of the search engines. It will be mostly be something specific and easy to verify.

that would be fantastic. I am ready to join!

Well, some bets will be excluded on what the site operator and me thinks it will encourage dangerous and unethical events to happen. Examples are not up for discussion.

kiba September 25, 2010 13:00 UTC Link to this message
Quote from: Babylon on September 25, 2010, 8:03:13 AM UTC

sorry if this question was asked already: what will the     bitcoin prediction market actually predict?

People put up a bet on things that they will think will happen. For example, who will get elected in 2012 US presidental election. Will the republican take over congress? The marketshare of the search engines. It will be mostly be something specific and easy to verify.

that would be fantastic. I am ready to join!

Well, some bets will be excluded on what the site operator and me thinks it will encourage dangerous and unethical events to happen. Examples are not up for discussion.

I am curious if your sites design will be open source.  Certain dangerous or possibly unethical predictions were one of the advantages I saw to bitcoin, although, obviously, you would not want to be the site admin responsible for allowing such predictions to be made.

It is open source. I just have not decided a license yet.

kiba September 25, 2010 13:01 UTC Link to this message
Quote from: chaord on September 25, 2010, 7:58:35 AM UTC

Well, some bets will be excluded on what the site operator and me thinks it will encourage dangerous and unethical events to happen. Examples are not up for discussion.

Allowing anonymity while enforcing exogenous bets is the biggest challenge IMO.

For now, I am going with the normal user account. It might change in the future.

kiba September 25, 2010 18:00 UTC Link to this message
Quote from: harding on September 25, 2010, 5:49:02 PM UTC

who [confirms the outcome of a scenario] and distributes the coins?

The simplest way is to let the participants confirm the outcome and have the site distribute the BTC.  Here’s how it works:

Every scenario has an end date.  After that date, all participants are asked to log in to their accounts and record whether they lost or won.  As each loser confirms losing, the site divides among the winners that loser’s contribution to the pool.  A proportionate amount of the winners’ contribution is also paid out to the winners, ensuring the pool maintains the same ratio of bets for and against the outcome that it had when betting stopped.

But, you ask, what keeps the losers honest?

On a pseudonymous site, each user account can have an associated feedback score which helps determine honest players.

On an anonymous site, the winners can offer a partial refund to losers who pay their bets.   For example: losers who confirm their loss will be refunded 5% of their bet.  As long as the money is stored in the site account (so the losers can’t use it), the losers’ best interest will be to confirm their loss and retrieve the refund.

This system minimizes demands on the site operator (hopefully keeping costs low) but also protects both parties from unforeseen events.  For example, imagine a bet 10 years ago on “who will win the year 2000 U.S. presidential election?”   Those losers who don’t believe the reported outcome was fairly determined can refuse to pay their bets, or better yet, they can request a high refund rate.  If the winners agree about the unfairness, they will offer the high refund rate.

But, you complain, my money and my winnings will be subject to the whim of some sore loser!  This is true, but a good site will let you sell you future share of the winnings on an open market.  You’ll receive money now and some entrepreneur will try to collect from the sore loser.

-Dave

I am not sure if this is fair, but I’ll implement your suggestion when the time come. The best way to find out something is to actually test and see how it goes.

kiba September 25, 2010 20:15 UTC Link to this message
Quote from: ribuck on September 25, 2010, 8:12:05 PM UTC

The simplest way to determine the outcome is to automate it completely.

The funds are frozen until one outcome is chosen by people who represent 50% or more of the betting stake. At that point, everything is paid out to those who favored the chosen outcome.

It’s in no-one’s interest for the funds to be tied up forever, so the participants will find a way to get 50% support for one outcome or the other.

All the organisers need to do is to run the website, not to get involved in controversies.

Automation would be basically the site gathering data and than computing the result and then paying the bettors accordingly. No humans involved other than the site operator/me coding it.

kiba September 26, 2010 03:40 UTC Link to this message
Quote from: nelisky on September 25, 2010, 11:58:53 PM UTC

My only line of reasoning is: I don’t think we should be imposing moral or creating rules of what should and should not be done with this. If we don’t like it, we don’t use it. Stay clear of all the things you oppose to, but let others do what they like with their own lifes… so as long as they don’t affect negatively the lifes of others, but that’s a completely different issue.

Hey, it’s the site operator and me’s discretion to slap down any bet we like.

But, I would like to give the site users to moderate themselves by killing a bet within say, 5 hours.

Let just say I don’t want bitpredict to be a source of calamity in the world.

kiba September 28, 2010 14:06 UTC Link to this message

Bitpredict is not dead! I just had another web application to attend to, one that will accept bitcoins!

I’ll report on the payment integration with my new web app since it is relevant to know how I will implement such a system for bitpredict.

kiba October 2, 2010 05:24 UTC Link to this message

Today, I worked on the development environment on my netbook at the coffeeshop/bookstore so I can develop bitpredict from there. The only problem is that the netbook take too long to do anything. By the time I leave, I have the development environment barely set up by the time I leave the shop to pick up mother.

kiba October 26, 2010 16:07 UTC Link to this message

Bitpredict will be resuming development with the project to learn how mybitcoin merchant API work for soulplaying.com

Once that is done, development of bitpredict’s market system will begin in earnest.

kiba December 3, 2010 08:43 UTC Link to this message

I started working on bitpredict’s user registration and user login/user sessions and the whole nine yard that come with setting up an account system.

I will come up with an accountless system for those who are too lazy to register.

kiba December 4, 2010 07:13 UTC Link to this message

THIS IS WORKING DRAFT:

How it will work:

Each claim have a stock of 1000. People will seed price estimation by deciding how much each of the stock is worth. Each stock can be sold and brought on the market. Each have an end date.

Traders who don’t invest in a claim will judge. There can only be 3 of them, they must be elected by majority rule on each side. If possible, bots will judge. Bot don’t have feeling, emotions and their code as well data they collect is open to the public for inspection.

The losers’ pool get divided up into the winners.

1% of winner pool get divided into the losers.

Meanwhile, the site operator and me get 0.002 for each trade. 0.001 of which will go to the judges.

UPDATE

Still working on old boring registration and session system.

kiba December 6, 2010 20:41 UTC Link to this message

I am please to announce that I am able to access bitpredict’s VPS again thanks to aceat64.

kiba December 12, 2010 04:58 UTC Link to this message

I created a rudimentary registration system. Not sure sessions work.

So I am working on prediction stuff.

kiba December 13, 2010 08:34 UTC Link to this message

Update:

I created two models; prediction and bid.

Prediction data: title, description, and by date.

Bid: PredictId, UserId, txn.

I also added a funds attribute to users.

Once the rudimentary detail is done, I’ll work on a script to upload site to aceat64’s server so that I can test it with a few people.

kiba December 14, 2010 04:12 UTC Link to this message

A little update:

Worked on making sure you can’t get a negative balance in your funds in your user account.

Barely got started on bid spec test.

kiba December 14, 2010 21:25 UTC Link to this message

Quickie update for today:

I got some of the bid data model set up. Mainly, it hold data relating to the prediction data, and the user data.

kiba December 17, 2010 05:57 UTC Link to this message

I am having trouble with bid and user.funds checking. However, there’s still a bunch of work elsewhere such as creating a CSS file for the site, so I did that in absence of progress.

kiba December 18, 2010 03:35 UTC Link to this message

Doh. It’s my bid spec that’s causing problem. Still, the easiest route to progress as well necessary anyway is to work on the site layout. Now it have a blue background.

kiba December 20, 2010 23:36 UTC Link to this message

Latest update…I am getting close to a proof of concept. User funds are getting subtracted; bids are linked to prediction and user…

Next step is to work on all the forms so people can submit predictions, and people can buy bids…

kiba December 24, 2010 00:09 UTC Link to this message

Today’s Update: I find myself creating a lot of missing pages after working the model for quite a bit.

I also realize at this moment that I forgot the crucial details on how to manage and rate judges.

kiba December 26, 2010 07:08 UTC Link to this message

I was working on the ability to make predictions.

kiba December 28, 2010 06:21 UTC Link to this message

Today, I worked on something non-essential called…the footer. Should be focusing my effort on getting a proof of concept working rather than..ya know…wasting time on footer.

kiba December 28, 2010 20:00 UTC Link to this message

I made it my goal to work on and fix up prediction creations. Then I got distracted by the fact that I didn’t have a logout mechanism. So I fix that.

kiba December 29, 2010 15:31 UTC Link to this message
Quote from: bitanarchy on December 29, 2010, 3:29:19 PM UTC

Would it be enough to allow bitcoin to include conditional transaction to create a prediction market? The conditional transactions then would consist of a normal transaction with a condition attached to it. The condition could be a prediction, like “silver will hit XX/dollar on AA/BB/CCCC”. The nodes would have to verify the outcome of the condition, to know whether the coins are transfered or not. So if you re-spend the conditionally spend coins, you may have to add a large fee to motivate nodes to verify the outcome of the condition.

Such is a complicated thing for the proof of concept.

kiba December 29, 2010 19:24 UTC Link to this message
Quote from: bitanarchy on December 29, 2010, 5:58:11 PM UTC

Yes, but you know that governments are fiercely opposed to unregulated prediction markets, even more than just private currencies I suspect. So, don’t you think, a prediction market must be P2P to avoid being shut down.

The site itself is open source. Shutting down my site doesn’t mean much of anything if people can set up their own exchange.

kiba December 29, 2010 20:04 UTC Link to this message
Quote from: bitanarchy on December 29, 2010, 7:52:02 PM UTC

I have not seen your site. Is it ready jet?

No. If it was…then I would already announce it.

bitanarchy

Anyway, if large sums are bet on a prediction, then your site will store those sums until the outcome is clear. There are a few problems:

  1. The government can confiscate the sums of bitcoins that you temporarily store.

My site is essentially a bank. I don’t store balances for outcomes. This is to facilitate a liquid market. Store at your own risk.

bitanarchy
  1. The outcome of the prediction is determined by your website. After that the bitcoins are payed-out, and this cannot be reversed, even if the outcome turns out to be the other way around. For example, there may be a bet called “Bin Laden brought down the twin towers”. If bets are put in a block chain, and the truth on the prediction turns out the be different then the transaction will be reversed by consensus of the nodes.

My site is just an intermediary. Outcome are decided by judges elected by people on each side.

kiba December 29, 2010 21:26 UTC Link to this message

Assassination politics is rather a dangerous game to play…Death by majority.

It would be one of those things excluded as a prediction in bitpredict.

kiba December 29, 2010 23:24 UTC Link to this message

Death is a dangerous weapon that should only really be used as self-defense.(How would you like it to be AIMED at you because of a person’s grudge or a bloodthirsty mob?)

Otherwise, I do not trust human beings’ judgement on the use of the death.

kiba December 29, 2010 23:43 UTC Link to this message
Quote from: Anonymous on December 29, 2010, 11:38:57 PM UTC

I merely state that the childs parents would be justified in placing a prediction  😊

Just because it fit libertarian ethical theory doesn’t mean it’s a good idea to put up a prediction.

Assassination politics open up a whole can of worm. It’s not going to be used against just evildoers, but people who are rich, people who made several individuals mad, or just being successful.