Updated September 16, 2026 · 18 min read · by Eric Lindquist
Kalshi has a board up that asks the question crypto holders have been dodging since February: which of these coins actually finishes 2026 in the green? There are 10 contracts, one each for Bitcoin, Ethereum, Solana, XRP, Dogecoin, Chainlink, Litecoin, Stellar Lumens, Polkadot and Shiba Inu. Each one pays $1 if that coin's calendar-year return comes out positive, and nothing at all if it does not.
Twelve days ago this page ended on a question about one of them. Chainlink sat 1.1% below the price it opened the year at, closer to its own line than anything on this board had been since January, and it had poked above that line intraday twice in August and closed back under it both times. The question here was whether the third time would be on December 31.
The answer arrived on September 6, and it arrived as a close rather than a poke. Chainlink closed that Sunday at $13.24, above its $12.184 January open, then closed above the line again on Monday and again on Tuesday. Three straight days, and the first closes above that level since January 24. Its contract went with the coin, printing 53¢ in the 9 PM ET hour on Monday, September 7, against the 44¢ this page last recorded. For three days the answer to the question in the headline was yes, for one coin.
It did not hold, and the reason it did not hold has nothing to do with Chainlink. Between September 4 and this morning the futures market's odds of a September rate hike went from 59% to 92.5% by Tuesday morning, and nine of the 10 climbs on this board got longer. Chainlink is back to $10.81, under every close it has printed since August 21. Its contract is 30¢. As of 1:00 AM ET on Wednesday, September 16 — 13 hours before the Fed announces — all 10 coins are back under their January lines, and this board has quietly stopped being 10 questions about 10 coins. The spread between its most expensive contract and its cheapest has gone from 37 cents to 20½ in 12 days, which is what happens when one decision starts pricing everything at once. Which contract survived that compression best, and which one got cheaper on a book with nobody in it, is the whole read below.
The Quick Answer
The market's answer is that probably none of them does it: no contract on this board is priced above 34¢, which is Bitcoin's, and all 10 coins are trading below the price they opened 2026 at. Chainlink is the closest, 12.7% from its own January line and 30¢ on Kalshi, and it is the one coin that actually got above that line this month before giving it back. The full board with every climb and every book, the tape behind the two biggest moves, and where our price-blind AI panel still disagrees with the price are all below.
Why Every Contract On This Board Starts From Behind
The settlement language is short enough to quote. Kalshi's rules line on the Bitcoin contract reads: "If Bitcoin has a positive return in 2026, then the market resolves to Yes." The other nine say the same thing with a different name in the sentence.
If none of that sounds like a bet you have placed before, our explainer on how a prediction market turns a question into a price is the 5-minute version. The word doing all the work here is return. Kalshi's contract terms define it as the last price of the period minus the first price of the period, divided by the first price. The first price here is the opening print of 2026, back on January 1, and nothing between then and the close of the year counts. A coin can spend the whole autumn above its January line and still resolve No if it slips back before December 31, and one that spends the autumn below can still resolve Yes by getting there on the final day.
Every contract here asks the same question: does this coin finish December 31 above the price it opened January 1 at? The terms close the obvious loophole in one line: "A Return of zero is not a positive return."
Chainlink's last two weeks are the cleanest demonstration of that rule this board has produced. Three closes above the line in September pay nothing. The only close that settles this contract is the one on December 31.
There is one oddity in the paperwork. The terms document names nine eligible cryptocurrencies, Bitcoin, Ethereum, Dogecoin, Shiba Inu, Chainlink, Polkadot, Ripple, Stellar Lumens and Litecoin, and Solana is not among them, yet Kalshi lists and quotes a Solana contract on this event alongside the other nine. The Solana contract carries its own rules line on the exchange, "If Solana has a positive return in 2026, then the market resolves to Yes," so the live listing has a settlement sentence even though the underlying terms document never names the coin. We re-pulled the terms document this morning and Solana is still not in it. It is a gap worth knowing about before you size that contract: a documentation question, not a rule anyone can trade against.
Prices come from CF Benchmarks, named as the source agency on the series. Each minute is a trimmed average of that minute's index prints, throwing out the top and bottom 20% of values, which is a small detail that stops one bad tick from settling a market. The board settles on that CF Benchmarks index rather than on any single exchange's screen, so the Coinbase figures throughout this piece are a close proxy and not the settlement number itself. Trading closes at 10:00 AM ET on January 1, 2027, reading the full 2026 year.
Costs matter here more than usual, because some of the books below are very wide. Kalshi's trading fee scales with the price of the contract, and our breakdown of what a Kalshi trade actually costs has the arithmetic.
The starting-line rule is why a coin can be having a good month and still be a long shot here. Litecoin is up 14.7% over the last 30 days and still needs 50.2% to get back to even. Polkadot has had the fourth-best month on the board at 24.5% and still needs to nearly double.
Stay ahead of the markets.
Daily insights and expert picks on Kalshi, Polymarket, and what's moving markets.
Free forever. Unsubscribe anytime.
New to this kind of market? Our plain-English guide to Kalshi's Bitcoin price markets walks through how a cent price becomes a probability and how those contracts settle.
These are market prices and model estimates, not predictions of fact and not financial advice. Prices on this page carry an "as of" stamp and get re-read when the story moves.
What 2026 Did To These Coins
The short version of the first eight months is that rates took the money out. Bitcoin's all-time high was $126,210 on October 6, 2025, and the decline that followed reached 52% peak to trough by the February lows after a 15% global tariff announcement "pushed inflation expectations high enough to take rate cuts off the table." Spring was worse — roughly $2.8 billion out of the spot Bitcoin ETFs across nine sessions in late May — and then Bitcoin ground sideways between about $60,000 and $67,000 from June into early August. Then one week in August put a real question on this board. On the 19th the SEC proposed rules laying out "a framework for crypto companies to raise capital," with a path for mature assets to exit securities classification entirely; Bitcoin was $64,877 that morning and Coinbase's candles have the week running to a $78,326 close on August 21, a 24.6% gain. Spot Bitcoin ETFs took in $1.92 billion that week and ether funds about $697 million.
That week is what the whole board has been arguing about ever since, and as of this morning most of it is gone. Bitcoin closed at $75,584 on Tuesday; the last time it closed lower was August 20, the day before the rally's biggest session. Chainlink, at $10.81, is under every close it has printed since August 21.
| Date | What happened |
|---|---|
| Oct 6, 2025 | Bitcoin's all-time high, $126,210 |
| Late Feb | A 15% global tariff announcement; rate-cut expectations come off the table |
| Feb Lows (after it) | The drawdown reaches 52% from the high |
| Late May | Nine straight sessions of spot Bitcoin ETF outflows, about $2.8 billion |
| June To Early Aug | Bitcoin grinds between roughly $60,000 and $67,000 |
| Aug 19 | The SEC proposes its crypto capital-raising framework |
| Aug 17 To 21 | Bitcoin runs from a $62,837 open to a $78,326 close, up 24.6% on the week |
| Week To Aug 21 | $1.92 billion into spot Bitcoin ETFs, about $697 million into ether funds |
| Aug 28 | Warsh at Jackson Hole: inflation trends have not "meaningfully improved"; Bitcoin retreats from above $81,000; spot Bitcoin ETFs post a $201.9 million outflow, ending a nine-day inflow run |
| Sep 3 | Fed Governor Waller says he can support holding rates in September; September hike odds fall to 54.6% by late morning and 50.5% by the evening close; Bitcoin prints $82,283, its highest in 90 days |
| Sep 4 | August payrolls come in at 162,000 against a consensus near 56,000; Bitcoin falls below $80,000; hike odds jump to 59% |
| Sep 6 To 8 | Chainlink closes above its January line three days running, its first closes above it since January 24 |
| Sep 8 | Polkadot ships a Product Devnet update; DOT trades as high as $1.2838, up 44% from its September 3 close |
| Sep 8 To 11 | Bitcoin gives back most of the week; spot Bitcoin ETFs lose $462.7 million while ether funds take in $196.9 million |
| Sep 11 | August CPI: 0.4% monthly and 3.4% annual headline, 0.3% monthly and 2.4% annual core; hike odds go from 72.4% to about 87% by lunchtime |
| Sep 14 | An attack on a Saudi pipeline pushes oil up more than 11% over five days; hike odds 86.5% |
| Sep 15 | Hike odds 92.5%; Bitcoin trades as low as $74,888 and closes at $75,584 |
The two weeks since this page last ran have a simple shape and a single driver, and the table above is the whole of it. What the table cannot carry is how one-directional the repricing was. Every number in that last column is a rate number: 162,000 August payrolls against a 56,000 consensus, which CoinDesk read as "giving the growing group of hawks at the Fed more ammunition to hike interest rates" and which TechTimes had lifting FedWatch to 59% from roughly 35% post-Jackson-Hole; a core CPI print of 0.3% against a 0.2% expectation on September 11, which took the probability from 72.4% to about 87% by 12:26 PM ET; an attack on a Saudi pipeline that pushed oil up more than 11% in five days and put Monday's odds at 86.5%, "up from 69.4% on Friday morning"; and then Tuesday's 92.5% with the meeting already under way. Not one of those is about a blockchain. The target range they are all arguing about is 3.50% to 3.75%.
There is exactly one flow story in the fortnight that is not a rate story, and it is worth the detour because it did not work. In the week to September 11, U.S. spot Bitcoin ETFs posted $462.7 million of net outflows — $282.7 million of it in one session, their largest single day since July, against a September that had been running positive to the tune of $307.3 million — with ARK 21Shares down $234.2 million and Grayscale's trust $129.1 million. The ether funds took in $196.9 million over the same five sessions, $216.4 million of it on the Friday. Money left one and arrived at the other in the same week. Ethereum's contract is 30.5¢ this morning against 34¢ on September 4, so whatever that reallocation bought, it did not buy this.
Hottest Prediction Markets Right Now
- 2028 Democratic presidential nominee · $209M traded
- 2027 Pro Football Champion · $77M traded
- Pro Baseball Champion · $65M traded
- 2028 U.S. Presidential Election winner? · $64M traded
- 2028 Republican presidential nominee · $62M traded
- Bitcoin price at the end of 2026 · $34M traded
Every market above links to our full AI model verdict; browse them all on the OddsShopper prediction markets hub, and see how every settled call actually scored on the full graded scoreboard.
The Climb Each Coin Has Left
Here is the distance problem in three numbers, all measured against each coin's own opening print of 2026 and its spot price at 1:00 AM ET on September 16.
- Chainlink Needs 12.7%. Down 11.3% on the year, still the smallest hole on the board, but more than 11 times the 1.1% this page recorded on September 4.
- Bitcoin Needs 15.7%. Down 13.6% on the year, on the board's busiest contract at 150,362 traded.
- Polkadot Needs 88.8%. Down 47.0% on the year, and the only climb on the board that got shorter in the last 12 days.
Distance alone undersells the spread, because these coins do not move at the same speed. Measured in each coin's own recent volatility, Chainlink's climb is now about 0.37 of one standard deviation of ordinary movement between now and New Year's, and Stellar Lumens is right behind it at 0.38; Bitcoin's is 0.74 and Ethereum's 0.75; Polkadot's is 1.69 and Litecoin's 1.75. There are 106 days left, and the 10:00 AM ET stamp on January 1 is when trading stops and the determination lands, not when the price is taken.
Put that against what actually happened. Running every 106-day window through Coinbase's daily closes for 2026, Polkadot's best stretch of the year gained 1.0% and Shiba Inu's gained 0.7%, against contracts that need 88.8% and 40.5%. Neither coin has come within an order of magnitude of what it is being asked to do. Chainlink's best 106-day window gained 38.5%, three times the 12.7% it needs — and it ended on September 6, the day it closed above its line. Bitcoin's best gained 20.2%, against a contract that needs 15.7%, though that stretch ran from February 5 to May 22 and has nothing to do with the last two months. Stellar Lumens is the quiet outlier: its best 106-day window of 2026 gained 66.2%, the largest on the board, against a 14.0% requirement.
The second structural fact is that these 10 independent contracts are not 10 independent bets. Over the last 90 days, Ethereum's daily moves correlate 0.89 with Bitcoin's, XRP 0.86, Solana 0.80, Dogecoin 0.79, Chainlink 0.77. Six of the nine non-Bitcoin contracts sit at 0.70 or higher. The two that break away are Shiba Inu at 0.65 and, newly, Polkadot at 0.47 — the lowest reading on the board and well below the 0.57 this page recorded for Stellar Lumens, the previous outlier, on September 4. Stellar Lumens has since drifted up to 0.69 and stopped being one. The last two weeks are what a high-correlation board looks like when one macro number is doing all the work: everything fell together, and the coin that broke ranks is the one whose correlation says it was never really on this ride.
And the calendar-year record says these things turn together. In 2024, nine of the 10 finished green. In 2025, all 10 finished red. Across every complete calendar year Coinbase has priced for these coins, 55 coin-years on this morning's pull, 31 came out positive, a little better than one in two, but they are stacked, not scattered.
The Board
Panel numbers below come from seven AI models that never saw a market price. Each one received the same fetched data card: the binding contract terms, every coin's opening print and spot, the required climb, realized volatility on two windows, momentum and chart levels, the correlation matrix, the full calendar-year ledger, and a dated news record. Each scored all 10 coins in a single pass, then read the other six boards anonymously and revised. The blend is the seat median.
Kalshi quotes are order-book midpoints as of September 16, 2026, 1:00 AM ET, and the rows are in market order, most expensive contract first. The market column is live. The panel column is not. It is the seat median from the August 28 read, frozen since, and every climb it was pricing has changed underneath it — Chainlink's by a factor of more than 11. Read the last column as a dated opinion about a different set of distances, not as this morning's answer. Because these contracts are not mutually exclusive, the column does not sum to 100 and is not supposed to.
| Coin | Still needs | Kalshi | Panel (Aug 28) |
|---|---|---|---|
| Bitcoin (BTC) | 15.7% | 34¢ | 43% |
| Ethereum (ETH) | 23.9% | 30.5¢ | 41% |
| Chainlink (LINK) | 12.7% | 30¢ | 55% |
| Stellar Lumens (XLM) | 14.0% | 24.5¢ | 34% |
| Solana (SOL) | 28.4% | 20.5¢ | 41% |
| Dogecoin (DOGE) | 46.5% | 18¢ | 24% |
| Ripple (XRP) | 41.9% | 16¢ | 33% |
| Polkadot (DOT) | 88.8% | 14¢ | 4% |
| Shiba Inu (SHIB) | 40.5% | 13.5¢ | 28% |
| Litecoin (LTC) | 50.2% | 13.5¢ | 8% |
More live boards from the same panel: the Bitcoin new-high ladder, the Ethereum ladder and the Solana board, each with its own live prices and as-of stamp.
The row I keep coming back to is the fourth one. Stellar Lumens needs 14.0%, 1.3 points more than Chainlink and 1.7 points less than Bitcoin, and it is 5½ cents cheaper than Chainlink and 9½ cheaper than Bitcoin on a 24 bid against 25 ask book, tied with Solana for the tightest here. Nothing in the distance column explains that discount, and the explanation this page has used twice before has just stopped working. In August the argument was the correlation column: Stellar Lumens printed 0.55 against Bitcoin, did not reliably get carried when Bitcoin rallied, and so the same short climb bought it less. That was the panel's reason for ranking it nine points below Bitcoin. Its correlation is 0.69 now, above Shiba Inu's 0.65 and a whisker under Litecoin's 0.72, which is no longer an outlier reading at all. So the discount survives and its stated cause does not, which makes this the one row on the board where I would want to see the panel run again before I read the gap as meaning anything.
Read the last two columns as a disagreement, never as a recommendation. And read the top of the market column carefully: Bitcoin leads this board for the first time since September 1, the last morning this page recorded it on top, and not because its contract went up — it is down a cent — but because everything that was above it came down.
What Moved Since September 4
Twelve days, and the board did not move as a block. The five most expensive contracts on September 4 are all cheaper; all five of the cheapest are dearer. The climb column tells you why the first half happened. The book tells you why the second half did.
| Coin | Kalshi Sep 4 | Kalshi Sep 16 | Contract move | Climb needed, then → now |
|---|---|---|---|---|
| Chainlink | 44¢ | 30¢ | -14¢ | 1.1% → 12.7% |
| Solana | 33.5¢ | 20.5¢ | -13¢ | 19.2% → 28.4% |
| XRP | 27.5¢ | 16¢ | -11.5¢ | 26.7% → 41.9% |
| Ethereum | 34¢ | 30.5¢ | -3.5¢ | 17.5% → 23.9% |
| Bitcoin | 35¢ | 34¢ | -1¢ | 7.7% → 15.7% |
| Litecoin | 13¢ | 13.5¢ | +0.5¢ | 49.5% → 50.2% |
| Stellar Lumens | 22.5¢ | 24.5¢ | +2¢ | 8.5% → 14.0% |
| Polkadot | 10.5¢ | 14¢ | +3.5¢ | 102.8% → 88.8% |
| Dogecoin | 12.5¢ | 18¢ | +5.5¢ | 33.6% → 46.5% |
| Shiba Inu | 7¢ | 13.5¢ | +6.5¢ | 28.7% → 40.5% |
Two of those risers moved on quotes rather than on flow. Polkadot is 10 bid against 18 ask, and the bid is for a tenth of one contract. Litecoin is 5 bid, 22 ask, a 17-cent spread carrying a 13.5¢ midpoint. Both have traded, but barely and erratically. Polkadot has printed 143 contracts in 22 prints since September 4, and its last four went off at 18¢, 11¢, 12¢ and 19¢ — straddling today's 14¢ midpoint by four cents in each direction, which is what a market with no price looks like. Litecoin has printed 138, of which 109 came in one run at 21¢ to 25¢ on September 11 and three have come since. Neither row has traded at all since Monday. Their midpoints went up because a maker moved one side of the quote, not because anybody paid more.
The other three are small but real, and Shiba Inu is the one that surprised me. On September 4 that row was 2 bid, 12 ask and the 7¢ in the old table was arithmetic. Since then 558 contracts have gone through it in 42 prints — more than Dogecoin's 261 and more than Stellar Lumens' 406 — and every print since September 9 has been a buyer lifting an offer: 31 lots at 10¢ that day, 38 at 14¢ and 15¢ on the 12th, 55 at 16¢ on the 13th, 122 at 14¢ and 15¢ on the 14th. The ask moved from 12¢ to 16¢, which no amount of bid-side quoting can do. Somebody has spent 12 days buying the board's worst coin on a climb that got 11.8 points longer. Dogecoin is the same story in miniature — 16 bid, 20 ask, 261 contracts since September 4, the busiest a 102-lot run of buyers at 18¢ to 20¢ on Monday, September 14, against a climb that got 12.9 points longer.
That is the single most interesting thing on this table, and it is the opposite of what the top half did: while money was leaving the contracts with the shortest climbs, a steady trickle was buying the ones with the longest. What happened at the top happened in a specific order, and it starts with the coin this page has been following.
Chainlink: The Line It Crossed, And The Line It Is Back Behind
The coin did the thing. Chainlink opened September 5 at $11.63, ran to a $13.36 high on Sunday the 6th and closed at $13.24, then printed $13.68 on Monday, its highest of 2026 since January and 12.3% clear of the $12.184 it needs. It closed above the line on all three of September 6, 7 and 8, and those are the only closes above that level Chainlink has recorded since January 24.
Its contract chased the coin rather than leading it. The 9 AM ET hour on Saturday, September 5 went the other way — 1,015 contracts changed hands and 1,014 of them were sellers, 802 of those in eight fills inside one second at 9:22:14 that walked 44¢ down to 34¢ before the book refilled — and then Sunday evening arrived and the buyers came in three waves: 667 contracts in the 5 PM hour, 566 of them lifting the ask from 41¢ to 48¢; 295 more in the 7 PM hour, every one a buyer, 44¢ to 49¢; 190 in the 9 PM hour, 48¢ to 52¢. Of the 1,152 contracts that traded across those three hours, 1,051 were buyers. Monday added prints at 53¢ in the 3 PM and 9 PM hours, the contract's high-water mark for the year on the visible tape.
From there it is a nine-day slide with no single bad hour in it, which is the tell that this was the macro doing the work rather than anything about Chainlink. The contract closed September 11 at 43¢ after touching 47¢, then went 44¢ to 35¢ on the 12th, 34¢ on the 13th, 33¢ on the 14th. On Tuesday, September 15 it traded 134 contracts and every single one was a seller hitting the bid: 58 at 31¢ at 1:39 PM, 29 more at 31¢ and 8½ at 29¢ in the 4 PM hour, and 38 at 29¢ at 11:31 PM, the last print on the row. The book behind it this morning is 28 bid, 32 ask, and the 30¢ midpoint is a real price on the way down: there are 129 contracts bid at 28 and 2,000 more at 27. The way up is thinner, three contracts offered at 32 and 15 at 33.
What that leaves is a contract that is 14 cents cheaper than it was 12 days ago on a climb that is 11.6 points longer, which sounds like the market simply repriced the distance. The beta table says it is not that simple. Chainlink moves about 1.27 times as far as Bitcoin on an average day, so its 12.7% hole converts to a Bitcoin move of 10.0%, against Bitcoin's own 15.7%. Stellar Lumens converts to 11.7%. Every other coin on this board still needs more Bitcoin than Bitcoin does. That was the panel's entire argument for putting Chainlink first in August, and the distances have moved a long way without breaking it — Chainlink is still the cheapest climb on the board in the only unit that accounts for how these coins actually move. It is simply no longer the cheapest by a mile.
Bitcoin: The Contract That Barely Moved
Bitcoin needs 15.7% in 106 days, and the book on it is 32 bid, 36 ask, a 34¢ midpoint, one cent below the 35¢ this page recorded on September 4. Its coin fell 5.1% over the same 12 days and its climb went from 7.7% to 15.7%, which makes a one-cent decline the strangest number on this page.
To read what 36¢ means: a buyer paying the ask puts up $100 for about 278 contracts, which return $278 in total if 2026 finishes green, before fees, and nothing if it does not.
Part of the answer is that Bitcoin's contract had already taken its beating before the others did. It went to 41¢ on September 6, the same weekend Chainlink was running, and it has spent the 10 days since in a 28¢-to-40¢ band while Chainlink fell 23 cents from its own high, Solana 15½ and XRP 15. The rest of the answer is that Bitcoin is the only row here where the last sustained flow was buying. Tuesday's session had the shape you would expect: 38¢ at 8:43 AM, then a one-second sell sweep at 11:56:17 that took about 154 contracts from 34¢ down to 29¢ as the coin printed its $74,888 low, then 15 more sold at 30¢ in the 2 PM hour. And then, in the small hours of this morning, 78 contracts, every one of them a buyer, walking back up from 33¢ to a 36¢ print at 12:53 AM. Somebody is buying Bitcoin's 2026 into a 92.5% hike probability. That is not a recommendation, and 78 contracts is not a mandate; it is the only sustained bid on the board this morning and it belongs in the record.
It is worth saying plainly what our own record says about the gap that is left between this column and the panel's. Across 3,523 graded markets in our ledger, there have been 521 settled calls where a panel blend landed at least 10 points away from the market price. It was right on 168 of them and wrong on 353 — 32%. Push the threshold out to 30 points and it is 37 right out of 112, 33%. Both recomputed from the ledger this morning. The panel's 43% against a 34¢ Bitcoin midpoint is a nine-point gap, which does not even clear the 10-point bar; the 25-point gap on Chainlink does, and the record says gaps that size are where this panel has historically lost. Divergence is a question worth investigating. It has never been an edge.
The Bottom Of The Board Went Up, And Mostly It Did Not
Polkadot is the one coin on this board whose climb got shorter. It closed at $0.8876 on September 3, ran to a $1.2838 high on September 8, a 44.6% move in five days, and has faded back to $0.947, which still leaves it 5.3% above where it closed on September 4, a session-to-session basis on which eight of the other nine lost ground. The rally, not the contract, is what shortened the climb, and The Coin Republic tied it to a Product Devnet update published on September 8 bringing new mobile and desktop apps and encryption changes, alongside a sharp rise in daily transaction volume. That is the only coin-specific catalyst on this board in the last two weeks that the record supports, and the correlation column agrees with it: Polkadot's daily moves now track Bitcoin's at 0.47, the lowest reading here and the only one that looks like an asset having its own week.
None of which makes its contract a price. Polkadot's climb is still 88.8%, which is 1.69 standard deviations of its own ordinary movement in the 106 days left, and its best 106-day window of the entire year gained 1.0%. Its book is 10 bid against 18 ask, and the 10¢ bid is for a tenth of one contract. The whole tape since September 6 is 17 prints and about 120 contracts, the largest a 25-lot buy at 18¢ on September 9 and the most recent four contracts at 19¢ on Monday morning. A 14¢ midpoint built from an 8-cent spread on 120 contracts in 10 days is a placeholder, not a verdict. Read the 18¢ ask as the executable number and the math gets worse, not better.
Litecoin is the same shape without the story: 5 bid, 22 ask, a 17-cent spread and the widest book here, three contracts traded since September 11's 109-lot run, and a climb of 50.2% against a best 106-day stretch this year of 9.0%. The panel's 8% sits inside that 17-cent spread, over the 5¢ bid and under the 22¢ ask, which is another way of saying the quote carries no information.
Stellar Lumens is the other row down here that is actually trading. Its contract went from 22.5¢ to 24.5¢ while its climb went from 8.5% to 14.0%, on 56 prints and 406 contracts of genuine two-way flow — 108 lots on September 13 that ran from 25¢ up to a 31¢ print and back down to 22¢, 68 bought at 25¢ to 28¢ on the 14th, then sellers at 25¢ and 24¢ on Tuesday afternoon. The book is 24 bid, 25 ask, one cent wide, tied with Solana for the tightest on the board. That is the cleanest quote on the page, and it is attached to the second-shortest climb here and the largest best-106-day window of 2026 at 66.2%. It is also the row whose old explanation just expired, as the board section above says: the 0.55 correlation that justified its discount in August reads 0.69 today. The panel's 34% sits 9½ points over the midpoint, and I would not spend that gap until the panel has seen this month.
XRP And Solana: The Two Biggest Give-Backs Behind Chainlink
XRP is down 11½ cents in 12 days, from 27.5¢ to 16¢, on a coin that fell 7.4% and a climb that stretched from 26.7% to 41.9% — the largest increase in required distance on the board. Its contract reached 31¢ on September 6 and has not been near it since. The last 36 hours are the whole story in two prints: 115 contracts bought in one second at 11:23 AM Tuesday at 23¢ and 24¢, and then, at 12:48 AM this morning, 148 contracts sold in about 43 seconds from 19¢ down to 17¢, including a single 108-lot at 17¢. Whoever bought Tuesday morning is looking at a 17¢ last print against a 23¢ and 24¢ entry. The book behind it is 15 bid, 17 ask with 1,352 contracts offered at 17¢ against 164 at the next level up, which is the reason the 16¢ midpoint has a ceiling over it.
Solana's decline was one hour. It traded 227 contracts in the 10 PM hour on September 10, 178 of them sellers, and went from 26¢ to 20¢ in that hour; everything since has been chop between 17¢ and 24¢ on a handful of lots a day, ending with 22 contracts bought at 21¢ on Tuesday evening. The midpoint is 20½¢ on a 20 bid, 21 ask book, which is a real quote on a thin market. Solana needs 28.4% and its best 106-day window this year gained 30.8%, so this is one of only five contracts on the board asking for something the coin has actually done inside 2026. It is also still the contract with the paperwork gap described at the top of this page: the terms document Kalshi links from the series names nine coins and not this one, so a reader sizing it is relying on the exchange's own rules line rather than the signed terms.
Ethereum deserves the last word here because it is the row where the flow story and the price story point in opposite directions. Its funds took in $196.9 million in the same week bitcoin funds lost $462.7 million, and its contract still went from 48¢ on September 6 to 30.5¢ this morning. Tuesday alone accounted for 12 cents of that: 976 contracts traded and 963 were sellers, in three clean waves — about 528 lots from 41¢ down to 33¢ starting at 1:38 PM, taking the day's 43¢ handle with them, with a 163-lot at 38¢ and a 200-lot back at 36¢ to close the run; 222 more from 34¢ to 32¢ in the 4 PM hour; and 213 at 31¢ and 30¢ in one second at 9:00:59 PM, 200 of them in a single clip. The book is 29 bid, 32 ask this morning, three cents wide, with 2,200 contracts bid at 28. Ether ETF demand was real and it did not matter, because a 92.5% hike probability is a bigger number than a $197 million week.
Every Seat's Number
This table has not changed since August 28 and will not change until the panel runs again. It is here as the dated record of what seven models thought when Chainlink needed 6.6% and Bitcoin needed 12.6%, which is a materially easier board than the one priced this morning.
| Seat | LINK | BTC | ETH | SOL | XLM | XRP | SHIB | DOGE | LTC | DOT |
|---|---|---|---|---|---|---|---|---|---|---|
| Claude Fable | 48% | 36% | 30% | 33% | 31% | 24% | 22% | 17% | 6% | 3% |
| Claude Opus | 55% | 40% | 33% | 37% | 34% | 27% | 23% | 18% | 8% | 3% |
| Claude Sonnet | 52% | 41% | 38% | 39% | 34% | 33% | 28% | 22% | 8% | 4% |
| ChatGPT | 54% | 43% | 42% | 41% | 32% | 33% | 28% | 24% | 10% | 5% |
| Gemini | 62% | 50% | 45% | 48% | 38% | 42% | 38% | 28% | 10% | 5% |
| Grok | 57% | 49% | 43% | 44% | 38% | 34% | 30% | 26% | 15% | 10% |
| Kimi | 55% | 45% | 41% | 41% | 33% | 30% | 25% | 24% | 7% | 2% |
| Seat Median | 55% | 43% | 41% | 41% | 34% | 33% | 28% | 24% | 8% | 4% |
Every seat on this panel is graded against real market settlements — records to date: Claude Fable 86% on 4,261 graded calls · Claude Opus 86% on 4,362 graded calls · Claude Sonnet 84% on 4,338 graded calls · ChatGPT 85% on 11,122 graded calls · Gemini 86% on 10,420 graded calls · Grok 84% on 8,728 graded calls · Kimi 85% on 5,481 graded calls. Recomputed daily as of September 16, 2026; the full scoreboard is public.
The row worth reading against this morning's board is Chainlink's. Every seat gave it their highest number, from 48% to 62%, and no other coin was ranked first by anybody. They were pricing a 6.6% climb. The climb is 12.7% now, and the market has moved from 44¢ to 30¢ while that column sat still, so the 25-point gap the table implies is partly a gap and partly a stopped clock. The same caveat applies in the other direction to Polkadot, where the panel's 4% now sits under a contract whose distance actually improved.
These are model estimates, not predictions of fact and not financial advice. Seven independent AI models returned a board on this market. Two further seats were unavailable when it ran and are excluded rather than filled in. Six of the seven then completed the revision round; Grok's first-round board stands as its final answer. Every number in this piece gets graded in public once the market settles, and you can read the running results on the full graded scoreboard.
The board has traded 698,660 contracts over its lifetime with 191,280 still open. That is about 19,800 contracts in the 12 days since this page last ran, a quiet fortnight for a market whose underlying moved this much, and worth remembering before reading any single midpoint too hard. All 10 contracts have both a bid and an ask, so every quote above is a two-sided market rather than a displayed one. The spreads are not equal. Stellar Lumens and Solana are one cent wide, Ethereum three, Bitcoin, Chainlink and Dogecoin four, XRP two, Shiba Inu five, Polkadot eight and Litecoin 17.
Where The Panel Changed Its Mind
The revision round is where this panel earns its keep, because the seats have to defend a number against six anonymous critics or move it. Everything in this section is from the August 28 run and is reproduced unchanged.
Gemini moved furthest, revising 92 points across the board, with a 24-point cut to Stellar Lumens and an eight-point trim to its own top pick. The correlation it cites below is the August 28 card's 0.55; today's is 0.69.
I revised my entire board down, accepting B and F's base rate argument that an asset class deep in a red year by late August rarely reverses the sign by December, despite the ETF inflows. Specifically, I drastically cut XLM from 62% to 38% after B and C correctly noted its 0.55 correlation prevents it from reliably riding the macro beta wave, invalidating my round-1 reliance on its raw distance. — Gemini
That is the one revision the last two weeks have argued with. Stellar Lumens' correlation to Bitcoin has risen from 0.55 to 0.69 since Gemini cut it, which weakens the specific reason it gave — and the contract went up anyway while nine of 10 climbs got longer. The base-rate half of its argument, meanwhile, has only gotten stronger.
ChatGPT went the other direction on the tail, cutting six contracts and raising only Chainlink, after a peer converted every required climb into the equivalent Bitcoin move. The sigma and beta it cites below are the August 28 card's; Dogecoin reads 1.29 sigma today and XRP a 1.50 beta, and both of those cut the same way its argument did.
I cut DOGE 29 to 24 because 0.82 sigma plus negative 90-day momentum needs a true meme-led alt surge, not just BTC recovering. I cut XRP 37 to 33 because its 1.30 beta still leaves a much taller BTC-equivalent hill than BTC or LINK. — ChatGPT
The most useful revision was the one that did not happen. Claude Opus held Bitcoin at 40% while the first-round boards around it spread from 34 to 60, and said exactly why.
I held BTC at 40 against a 34-60 spread because nobody surfaced a fact, only differing faith in August's durability, and conditional on eight months of red the ledger says class turns rarely flip sign inside four months. — Claude Opus
Three weeks on, that is the read the tape has been kindest to. Bitcoin's contract is down a single cent in 12 days, the smallest move on the board outside Litecoin's untradeable 5/22 quote.
What Would Change The Panel's Mind
Five things between now and settlement, each one checkable, each with the direction it pushes. One of them lands hours after this read.
- The Fed's decision at 2:00 PM ET on Wednesday, September 16. Futures had a quarter-point hike at 50.5% on the evening of September 3 and 92.5% on Tuesday morning, so the hike itself is not the event — the statement and the path are. A hike delivered with softer language about what comes next is the outcome this board is least positioned for, and it would lift every contract on it. A hike plus a hawkish path does the opposite, and the current target range is 3.50% to 3.75%.
- Spot ETF flows. The shape is one heavy outflow week, $462.7 million out of the bitcoin funds between September 8 and 11 with $282.7 million of it on the Thursday, against $196.9 million into the ether funds over the same five sessions. The late-May shape was nine straight sessions and about $2.8 billion out. A return to that puts the majors back at their August lows; a flip back to the $730.9 million single day The Block logged on September 3 is the fastest route to a repricing, and the seats that leaned on the institutional reading, ChatGPT and Gemini, have the most riding on which it is.
- Charles Schwab actually launching Solana, Avalanche and Chainlink. Schwab's own press release still says these assets "will be available in the coming months," with no date attached three weeks after the announcement. That is unchanged, and it is the only coin-specific catalyst on this board with two names on it. Chainlink and Solana have between them given back 27 cents of contract value since September 4 with that item still sitting unexecuted.
- Bitcoin reclaiming $87,498 at any point this autumn. The figure is Bitcoin's own January opening print, the exact line its contract has to clear. The first time it trades above that level, the read on this whole board changes, because six of the nine non-Bitcoin contracts move with Bitcoin at 0.70 or higher.
- The SEC's proposed framework moving from proposal to final rule. The August 19 announcement was a proposal with a comment period ahead of it. Adoption pushes every number on this board up; a stalled or narrowed rule pushes them down. The reporting on the proposal read the exit-from-securities path as most likely to benefit mature networks such as Bitcoin and Ethereum, so those two contracts have the most riding on it.
Common Questions About Kalshi's 2026 Crypto Return Board
What counts as a "positive return" for settlement? Kalshi's terms define it as the last price of the period minus the first price of the period, divided by the first price, measured on the CF Benchmarks index. The first price is the opening print of January 1, 2026 and the last is the close of December 31, and nothing in between counts. The terms also spell out that "A Return of zero is not a positive return."
Chainlink traded above its January line in September. Why didn't that settle anything? Because only December 31 matters. Chainlink closed above its $12.184 opening print on September 6, 7 and 8, its first closes above that level since January 24, and then fell back to $10.81. Three good closes in September pay nothing; the contract reads one close, at the end of the year.
Do these 10 contracts settle independently? Legally yes; practically no. Over the last 90 days, six of the nine non-Bitcoin coins move with Bitcoin at a correlation of 0.70 or higher, so a basket of them is closer to one macro call held several times over than to several separate wagers. The two that break away are Shiba Inu at 0.65 and Polkadot at 0.47.
When This Settles
| Measurement Period | January 1 to December 31, 2026 |
| Settlement Source | CF Benchmarks index, trimmed minute average |
| Final Determination | 10:00 AM ET, January 1, 2027 |
| Next Known Catalysts | The Fed's decision at 2:00 PM ET on September 16; SEC comment period on the August 19 proposal; Schwab's Solana, Avalanche and Chainlink launch, dated only as "coming months" |
| As Of | Kalshi order books 1:00 AM ET, September 16, 2026; Coinbase spot 1:00 AM ET, September 16, 2026, 13 hours before the Fed announces; panel read August 28, 2026 |
This page gets re-scored when the story moves.
More on this: Where Does Bitcoin End 2026? Kalshi's Price-Band Odds · How To Play (And Not Get Torched) On Prediction Markets · How Bitcoin Price Contracts Work On Kalshi · Is Kalshi Gambling? The Legal Answer And The Honest One · How Kalshi's Rain Markets Decide It Rained
The Bottom Line
Ten coins, 106 days, and all 10 of them back below their own January lines after three days in September when one of them was not. That is the honest summary of the fortnight: Chainlink answered the question this page asked, three closes above $12.184 on the 6th, 7th and 8th, and then a jobs report, a core CPI print a tenth above expectations and a Saudi pipeline took the answer away — not from Chainlink specifically, but from all 10 at once, which is what a board with six correlations above 0.70 does when the rate story turns.
What it left behind is a flatter board than the one this page has carried since August. The most expensive contract and the cheapest were 37 cents apart on September 4; they are 20½ apart now, and the top of the table is Bitcoin at 34¢ purely because everything above it fell. Two of the five contracts that got dearer did it without a trade, on books nobody could transact in, so the real compression is narrower than the table makes it look. But it is real, and it says the market has stopped pricing 10 coin questions and started pricing one rate question ten times.
The panel's August column is now a museum piece, and I would rather say so than dress it up: it scored a board where Chainlink needed 6.6% and Bitcoin 12.6%, and both of those numbers have roughly doubled. What survives from it is not a number but an argument, and the argument is still standing. Chainlink's 12.7% converts to a 10.0% Bitcoin move against Bitcoin's own 15.7%, which means the cheapest climb on this board in the only unit that matters is still priced four cents under Bitcoin's. At 2:00 PM ET on Wednesday somebody at the Fed reads a statement, and the next read of this page will know whether that gap was information or just a slow-moving quote.
