Outcome-Resolution Demo

[psztorc]

My friend Alex has created a prototype demo, where one can more easily understand the way that outcomes are calculated. On Tuesday, I will be giving a presentation directly on that topic, as well. I hope to have the slides done soon.

Cool Consensus Demo Site: https://lyoshenka.ocpu.io/truthcoindemo/www/


Note that the demo would, obviously, differ from the actual experience in several ways. For example:

1. In the demo, one can set all the Votes one wishes, ie, changes everyone's Ballot as much as they want. In reality, one would only set their Ballot (ie, only one row), and one would do it before any other Ballots were visible.

2. In the demo, (for simplicity) each person's vote counts equally. In reality, each person's vote would count in proportion to their ownership of VoteCoins.

3. The demo is only of one period. In reality, such a vote would take place each Voting Period (with different questions, of course).

4. In the demo, the questions sometimes refer to events which took place several years ago. In reality, the questions would likely all be on events taking place within the previous Voting Period (ie, the preceding 6 weeks).

5. The demo has short 'question titles', to fit them into cells of a simple matrix. In reality, the questions would be longer and clearer. For example, the original question texts were:

    QID1: In the United States, following the 2012 November elections, was Barack Obama elected US President?
    QID2: In the United States, following the 2012 November elections, was Mitt Romney elected US President?
    QID3: In the United States, following the 2012 November elections, did the Democratic Party control 51 or more Senate seats?
    QID4: In the United States, following the 2012 November elections, did the Republican Party control 218 or more seats in the House of Representatives?
    QID5: In the United States, following the 2012 November elections, how many seats in the House of Representatives were controlled by the Republican Party? Range: 0 to 435
    QID6: During the 2011-2012 United States football season, did the New England Patriots (AFC) win the 2012 Super Bowl (XLVI)?
    QID7: During the 2013-2014 United States football season, did the Denver Broncos (AFC) win the 2014 Super Bowl (XLVIII)?
    QID8: On June 27th, 2014, was the closing price of the Dow Jones Industrial Average (INDEXDJX:.DJI) above 16000?
    QID9: On June 27th, 2014, was the closing price of the SPDR Gold Trust (ETF) (NYSEARCA:GLD) above 120?
    QID10: On July 9th, 2014, what was the closing price of the Dow Jones Industrial Average (INDEXDJX:.DJI, USD per Share, 2 decimal points)? Range: 8000 to 20000

6. The demo has only 10 Decisions (columns), I anticipate realistic VoteMatrices to have 30-100 Decisions (columns).

7. The demo does not have any Decisions which I would consider to be vague or un-answerable (ie, the ".5" selection), although this may happen in reality.

8. In reality, there would be an instruction on what to vote for a S decision which was vauge (either its max or its min). In the demo this instruction isn't present.


The graphic is exactly of the form of the two examples in the 1.3 Whitepaper, Figure 4 (page 31). In summary, the rectangles represent votes (this is simple), but the votes gain/lose transparency base on how coordinated "the Ballot they came from" was with other Ballots (this is more complex).

[koeppelmann]

Well done!
I am looking forward to the presentation on Tuesday.

[psztorc]

Currently, a bug is displayed when users attempt to run the Demo. I fixed the bug but it is taking very long to sync over to OpenCPU for some reason.

[martinBrown]

Slick, nicely done.

Might as well share this old ipython notebook, a very incomplete work-in-progress.