Finally, the “The Bet Outcome Settlement Process” will be done via the usage of open source Large Language Models (LLMs) in order to ensure that the decided upon outcome is transparent, deterministic and consistently correct.

LLMs are containerized software technology optimized for answering Natural Language prompts accurately without ulterior or hidden motive.

Therefore they are perfect for dictating the outcome of a sport event in a factual and neutral way when given with the necessary information with respect to any specific concluded sport.

In essence, the bet that users are actually placing is something like “Once the agreed upon hypertext of returned by this url is provided to the LLM model at this specific point in time along with this standard text prompt, the LLM will respond with this team's name”.

Technically, the bettors are betting on the response of the LLM model itself, which with the current performance of LLMs retrained for this specific purpose, should always match the real world outcome of the game. The outcome stated by the LLMs once the prompt has been executed will be publicly announced and attached to to the sport event in question, so that the Routing Lightning Node previously described acting as the arbitrator of the bet can programmatically settle the bets for a specific sporting event by releasing the payments to the users who’s bet selection matches the output of the LLM.

I have attached some screenshot examples using ChatGPT-4 for illustration purposes. You can try examples yourself with similar prompts to validate that LLMs currently have this capability. In practice Vigless will only use an open source LLM model such as Mosaic’s MPT-7B model or Meta’s LLama 2 model and only after said LLM has been backtested on thousands of historic sporting events to ensure that it has a zero level error rate. The open source nature of these models would ensure that they can be run hosted free of licensing costs and in a distributed fashion.

The url whose html will be passed to the LLM as well as the accompanying prompt that will be used to settle the response will transparently be displayed within the metadata of the betting event and will be available to any user prior to the placement of any bet so that any user placing a bet on said sport event is acknowledging that they have agreed to that url as being the source of truth and consider the bet to be fair.