X-Tweet Market: AI Tweet Count Predictions

AI-powered tweet count predictions for 9 public figures. Live tracking, hourly heatmaps, Polymarket bracket analysis, and historical posting patterns. Every forecast is graded after the market closes: 309 graded, 83% inside the bracket or within 20% of the count, paper ROI +0.4% at $1,000 per call. Misses are on the page.

Tracked Accounts

Select a person to see AI predictions, live count, heatmap, and bracket analysis.

Graded record by persona

Every resolved Polymarket post-count market is graded against the forecast we had 24 h before the close. Hit = within 20% of the final count or inside the winning bracket. Money = $1,000 on the bracket our forecast pointed to, at that moment's YES price, held to resolution; markets without a price at the checkpoint are not staked. Buying the favourite bracket instead would have returned −$4,193 on the same markets. As of 2026-09-10.

PersonaGradedHitStakedWonAvg priceP&LROI
Elon Musk9280%653050¢−$8,136-12.5%
Donald Trump4178%211037¢+$3,141+15.0%
Andrew Tate000
Ted Cruz3997%271548¢+$293+1.1%
CZ Binancenot graded: counter mismatch: our tracker counts original posts, the market counts all posts — grading paused 2026-09-04
White House4098%261866¢−$1,160-4.5%
Khamenei2979%211778¢+$656+3.1%
NYC Mayor3181%161473¢+$3,623+22.6%
Zelenskyy3773%211668¢+$2,349+11.2%
All personas30983%19712057¢+$765+0.4%

How AI Tweet Predictions Work

OddsShift's AI model analyzes thousands of historical posts per person, identifying time-of-day patterns, day-of-week trends, and news-cycle correlations. The model generates hourly expected counts and compares them against live Polymarket bracket odds to identify over/underpriced contracts. Forecast accuracy is shown per market on each persona page, misses included.

Frequently Asked Questions

How accurate are OddsShift tweet predictions?

Graded, not claimed: 83% of 309 resolved markets landed inside the winning bracket or within 20% of the final count. Staked at $1,000 per call 24 hours before close, the same forecasts return +0.4% (+$765) — a hit rate is not a price. The per-persona table above shows who pays and who does not.

Which accounts does OddsShift track for tweet predictions?

OddsShift tracks 9 public figures: Elon Musk, Donald Trump (Truth Social), Andrew Tate, Ted Cruz, CZ Binance, White House, Ayatollah Khamenei, NYC Mayor, and Volodymyr Zelenskyy.

How do tweet count prediction markets work?

Polymarket offers contracts on how many times a public figure will post in a given week. For example, "Will Elon Musk post more than 150 times this week?" You buy YES or NO contracts. OddsShift's AI forecasts help traders identify which bracket is most likely.

Is there an edge in tweet-count markets?

Only where the forecast beats the price, and the record says that is persona-specific. Across all tracked accounts the $1,000-per-call book is +0.4%; the crowd favourite on the same markets is −$4,193. Read the per-persona rows before treating any bracket as an edge. Nothing here is investment advice.