Global Weather Forecast and Archive

[Markus]

There is no need for a central feed to be trusted. Nor would anybody have to report the value. After the event longs and shorts will both have their own figure in mind for what happened, but they will have to find a compromise.

The settings must be such that neither party gains from holding out too long after the event. If longs win and shorts lose (in terms of yield as in BitShares) then the settling price would be above the "real" temperature. I want the optimum settling price to reflect the exact value of the event.

[bytemaster]

This DAC is really just a specialization of event-based prediction market DAC. 

Change the form of the question from 'true/false' to a prediction of the temperature at a given time and place.  This prediction market would never have to end, just like the BitUSD prediction market never ends.  What would happen is that after the event and the information is 'known' the prediction would be very obvious.  Shorts would be facing a constant loss by not settling and longs a constant gain... in effect there exists a Nash Equilibrium where the Shorts either cover or eventually they will run out of margin due to the interest they owe and the network will force them to cover.

[phoenix]

This is a great idea, but figuring out what actually happened in a decentralized manner would be very difficult. If you have trusted feeds, then they become subject to manipulation. If you have people report what happened, then there will be incentive to skew the data.

[Markus]

Based on the concept of BitShares. Instead of creating BitAssets whose value is supposed to mirror a real life asset (USD, Gold, …) the "assets" in this DAC would be a certain weather information at a certain point in time and space.

For example: Maximum temperature in Adelaide, South Australia on 15th Dec 2013 in °C.

When the point in time is still far ahead the orderbook spread would be wide, after a while the first trade would happen and the market price should fluctuate according to the best forecast, then approach the real temperature. After the point in time has passed the fair value is known*) and shorts and longs should slowly pull out of the market at that price (same as with BitAssets).

*) This could be defined in advance as the data of a certain weather station. But this could incentivise manipulation. Instead the location doesn't have to be defined that exactly as the market will have to find a consensus anyway.

For those who are just interested in the weather data, there is no need for trading. And even if no trade ever happens, the bid/ask spread still contains valuable information.

Issues I can think of:
- Many rather small markets
- Who decides which markets to create

Of course this must not be limited to "max daily temperature", any weather data would be possible: Min temp, temp at x o'clock, mm of rainfall, hours of sunshine per day, average windspeed of a day, etc.