In Reply to: And Prediction markets are increasingly gamified posted by Dr.J on May 11, 2026 at 13:46:04
> They are designed to allow some people to make huge profits more than making a scientific calculation of probabilities.
That's a pretty ignorant blanket statement.
There are characteristics that make some questions good or bad for prediction market prognostications.
Here is a Claude summary
"What prediction markets do well
Binary, verifiable outcomes —
Questions with clear yes/no resolutions (Will X win the election? Will Y company merge with Z?) are ideal. The market needs a crisp settlement condition.
Short-to-medium time horizons — Markets are most accurate for events resolving within days to a few years. Longer horizons suffer from discounting uncertainty and participant turnover.
High-information events — Elections, sports outcomes, economic indicators, and corporate events attract knowledgeable participants who move prices toward truth. The more experts care, the better the aggregation.
Aggregating dispersed knowledge — This is their core strength. No single analyst has all the information, but markets pull together signals from thousands of people with different data points, incentivizing honest revelation.
Outperforming polls and pundits — On political forecasting especially, prediction markets have consistently beaten traditional polling averages and expert panels, partly because they incorporate polls plus other signals.
Detecting early signals — Because they update continuously, markets often shift before mainstream media catches on — e.g., leadership changes, legal verdicts, geopolitical developments.
Where they underperform
Tail risks / black swans — Rare, unprecedented events are systematically underpriced
Thin markets — Low liquidity means a few traders can move prices without adding real information
Long-run structural questions — "Will AI surpass human
intelligence by 2040?" is hard to trade rationally
Manipulable outcomes — If a participant can influence the event itself (e.g., a CEO trading on their own resignation), accuracy breaks down
Questions that are hard to resolve — Ambiguous resolution criteria distort prices"
Since 2022 prediction markets have been perfect on interest rate changes and more accurate in the Trump vs Harris prognostication.
They are good for the right questions and bad for the wrong ones.