November 20, 2026
Some days the big favorites do the work. Today wasn't one of them. Looking across the board on November 20, the headline numbers barely budged β but the low-probability legs, the ones most people scroll past, carried almost all of the week's repricing.
That's a pattern worth logging. When a 95% book sits still and its 4% counterpart moves five points, the market isn't changing its mind about the outcome. It's changing its mind about the shape of the risk. Here's the polymarket analysis I kept notes on today.
The December FOMC board is split into separate single-outcome books, and the volume is lopsided. The no-change ("hold") market traded $1.29M in 24 hours against $11.4M lifetime β easily the busiest book in today's data β yet it sits at just 2.9% Yes, up 0.8 on the day and 0.5 on the week. Heavy churn, almost no price movement.
The action was one lane over. The 50+ bps hike market prices Yes at 4.9%, down 1.7 points in a session and down 4.8 points over seven days. That's roughly half the leg's value gone in a week. Meanwhile both cut books β 25 bps at 0.2% and 50+ bps at 0.1% β are flat to the decimal, with a combined $14M in lifetime volume essentially parked.
Add the four visible legs β 2.9 + 0.2 + 0.1 + 4.9 β and you get 8.1%. Whatever sits outside those legs (a standard 25 bps hike being the obvious candidate) is implicitly carrying about 91.9%. Run the same math on last week's prints and the unlisted residual was closer to 87β88%. So the week's story isn't "the Fed got more hawkish." It's "the market stopped paying for the extreme version of hawkish." The direction was never in doubt; the magnitude tail deflated.
One caveat I always flag in this kind of polymarket analysis: these are independent books, not a single normalized bracket. The legs don't have to sum to 100, and the residual is an inference, not a quoted price.
The IsraelβIran ceasefire continuation market reads 95.5% Yes as of the November 20 snapshot β unchanged on the day, but up 10 full points over the week. Translated into the leg that actually moved: the break scenario went from roughly 14.5% to 4.5%, a two-thirds reduction in a week of calendar decay plus, presumably, an absence of headlines.
The structural detail I find more interesting than the price: resting liquidity is $50,807 against $341,668 of 24-hour volume. That's turnover running close to seven times the visible book depth, on a market with a deadline two days out. Short-horizon geopolitical books tend to look like this β plenty of flow, very little standing depth β which means quoted mid-prices can be less informative about size than they look. If you track these, note the depth number alongside the probability, not instead of it.
Two dated books on the same question. Rejoin by November 30 prices Yes at 1.3%, down 1.2 on the day and down 0.4 on the week. Rejoin by December 31 prices Yes at 5.7%, down 1.4 on the day but up 3.0 over the week.
Because the later date strictly contains the earlier one, the gap between them is the implied probability of the event happening specifically in December: 5.7 β 1.3 = 4.4 points. Back out last week's prints and that same slice was closer to a single point. So over seven days the near leg drifted down while the far leg climbed β probability didn't leave the board, it migrated down the calendar.
That's a different signal than a uniform fade. A uniform fade says "less likely, full stop." A steepening ladder says "not in the next ten days, but don't write off the quarter." Both books also moved down together in the last session, so the day and the week are telling slightly different stories β which is usually a prompt to wait for a second session of confirmation rather than treat the weekly move as settled.
None of this is a trade recommendation, and trader execution on this project is currently off. These are watchlist entries with three checks attached: does the residual arithmetic on the Fed board hold next session, does the ceasefire book's depth improve or thin further into expiry, and does the NVIDIA December slice widen past 5 points or collapse back toward parity with the November leg.
I post these observations as they develop β odds snapshots, depth notes, and catalyst dates β in our free Telegram channel. If you want the running log alongside the prediction market odds I'm tracking day to day, join the watchlist on Telegram and follow along with fellow traders. Research first, always.
Each outcome is listed as its own independent Yes/No book rather than as one normalized multi-choice bracket. Fees, spreads, and the simple fact that at least one outcome may not have a listed market all mean the visible legs can sum to well under 100. On November 20 the four listed December legs totaled 8.1%, leaving roughly 92 points for the unlisted residual.
It means a lot of contracts are changing hands relative to the amount of resting size on the order book. The ceasefire market showed $341K of 24-hour volume against about $51K of liquidity. Prices can still be informative, but a single sizeable order can move the quote more than the headline probability suggests, so depth deserves to be read alongside the odds.
Subtract them. If the later deadline fully contains the earlier one, the difference is the implied probability of the event landing in the window between the two dates. On the NVIDIA pair, 5.7% minus 1.3% leaves 4.4 points assigned to December specifically β up from roughly one point a week earlier.
They answer different questions. The daily change captures the latest session's flow; the weekly change captures the trend. When they disagree β as they did on both NVIDIA books today, down on the day but up on the week for the December leg β that's usually a reason to wait for another session before treating either as the signal.
No. Everything here is observational analysis of publicly quoted odds, volume, and liquidity at a specific timestamp. Trader execution for this project is off, and nothing above should be read as advice to enter or exit any position.