DefiLlama changed 93% of the days it reported. The chain changed none
DefiLlama changed 95% of the prediction market days and 91% of the launchpad days Tickerz recorded, one by 162%, one 130 days after the day it counts. The Labor Department changed 2 of 106 weeks. The Solana chain changed none of 80 days.
A data vendor rewrites. A count of the chain does not.
Every day Tickerz saves each number it records and timestamps the file in Bitcoin. Read those 51 files in order and you see every number a source changed after it first gave it, and when.
| Index | Source | Recorded | Changed | Typical | Largest | Latest |
|---|---|---|---|---|---|---|
| $YESNO | DefiLlama | 171 | 162 (95%) | 0.6% | +22% | 130 days |
| $TRENCHES | DefiLlama | 180 | 163 (91%) | 1.8% | +162% | 130 days |
| $CRIME | City of Chicago | 103 | 83 (81%) | 0.6% | +4.1% | 97 days |
| $LAYOFFS | US Department of Labor | 106 | 2 (2%) | under 0.1% | under 0.1% | 14 days |
| $GIGS | Base chain | 58 | 0 | none | none | none |
| $MINTS | Solana chain | 80 | 0 | none | none | none |
| $WAGE | Base chain | 58 | 0 | none | none | none |
Typical is the median size of a change against the number first recorded. Latest is the longest gap between the day a period counts and its last change. $TRENCHES for Aug 1 went from $1,639,688 of fees to $4,301,531. The Labor Department changed 2 weeks, each by under 0.1%.
A market pays out on one of the two numbers.
When a source changes a number after a market settled on it, each side was paid on a number the source no longer shows. Tickerz keeps the first and shows the second beside it, on every index page and in each settlement file. The bet study counts the settled Kalshi and Polymarket markets the Bureau of Labor Statistics revisions would flip.
Method and caveats
- For every period an index recorded, the value in the first proof that holds it against the value in the last proof that holds it. A period counts as revised when the two differ. Rows: period, first recorded value, latest value, day first recorded, day of the latest change.
- Data: https://tickerz.com/api/indices/proof?day=YYYY-MM-DD, every daily proof from 2026-09-23 to 2026-11-12; each proof holds every period written or revised that day, read 2026-11-13 07:00 UTC.
- The first proof, September 23, 2026, also holds the history Tickerz loaded that day. A period older than that is measured from the day Tickerz first read it, so any change before then is not seen.
- $CRIME, $GIGS and $WAGE are retired. Their records stay in the proofs and are counted here.
- The Bureau of Labor Statistics indexes have too few months in this window to count. The bet study covers them.
- It describes the record and predicts no price.
Corrections
No corrections yet. First published November 13, 2026.