Starknet (STRK) did not put in a 107% rise because the chain locked more value. The token's market did that. The deposits did not. On Binance, the STRKUSDT candle for 5 September opened at $0.02599. At 07:31 UTC on 4 October the last price on that book was $0.05381, which is 107% above the 5 September open and 78.7% above the 5 September close of $0.03011. Only the first of those two measures is a doubling. DefiLlama's historical daily series shows Starknet total value locked at $160.44 million on 5 September and $164.46 million on 4 October, a 2.5% change. A second print from the same firm, the live chains endpoint, read $189.41 million on the morning of 4 October. The two figures do not agree. Neither is a doubling.
What a 30-day chart hides is where that last sits inside 2026. The 31 January close was $0.0549, and the 07:31 last sat about 2% under it. 31 January was the break under six cents: low $0.0498, close $0.0549. The day before, 30 January, was the last settlement above $0.061, at $0.0616. The autumn move repaired later damage. It did not clear the close that stood just before the break, and the 6 January close of $0.0926 is still 41.9% above the 07:31 print.
Price, on that anchor, is back at a late-January settlement. It is not back in January.
A second ticker at 07:52 UTC printed $0.05622, about 2.4% above the 31 January close, with the session high still $0.05657. Every percentage below uses the 07:31 last of $0.05381, so the note has one anchor. Both prints sit under $0.09 and over $0.03.
- Binance STRKUSDT last price $0.05381 at 07:31 UTC on 4 October 2026, up 24.6% over the trailing 24 hours, on $21.56 million of quote volume and 285,833 trades — Binance 24-hour ticker, window ending 4 October 2026
- From the 5 September open of $0.02599 to that last, STRK rose 107%; from the 5 September close of $0.03011 the rise is 78.7% — Binance daily candles, retrieved 4 October 2026
- On 18 September STRK closed up 52.6% at $0.04425, with $33.67 million of quote turnover, 16.8 times the $1,998,750 median daily quote volume from 1 January through 17 September — Binance daily candles, 4 October 2026
- Ether rose 6.7% and bitcoin 5.8% on 18 September, while DefiLlama's historical Starknet series rose 1.2%, from $155.04 million to $156.97 million — Binance ETHUSDT and BTCUSDT daily closes, and DefiLlama historical chain TVL, retrieved 4 October 2026
- The 6 January 2026 candle opened at $0.0907 and printed a high of $0.0963; the 4 October last is 41.9% below that day's close of $0.0926 — Binance daily candles, retrieved 4 October 2026
- DefiLlama's historical point for 4 October is $164.46 million, against $189.41 million on the live chains endpoint the same morning. The historical high for 2026 was $323.90 million on 20 January, a day STRK closed at $0.0772 — DefiLlama, retrieved 4 October 2026
18 September was not an ether session
STRK closed at $0.02899 on 17 September and at $0.04425 on 18 September. The high was $0.04527. The low, $0.02876, was a fraction of a cent under the prior close. Quote turnover was $33.67 million, 16.8 times the $1,998,750 median across the 260 bars from 1 January through 17 September. The next three settlements stayed heavy, $27.48 million, $20.21 million and $11.83 million, and the 23 September close was $0.03832.
Ether did rally. ETHUSDT closed up 6.7%, from $2,447.28 to $2,612.14, and BTCUSDT rose 5.8%, from $76,417 to $80,884. A beta day explains a few percent. It does not explain fifty. From the 5 September closes to 4 October, ether rose 8.6% and bitcoin 6.4%, against STRK's 78.7%.
Follow-through stalled, then a second leg arrived. The 20 September close was $0.04846. On 23 September the close fell 9.8% to $0.03832, and the market sat between roughly $0.038 and $0.044 into month-end. On 3 October it ran again: open $0.04296, high $0.05571, close $0.05428, quote turnover $16.09 million, 26.3% above the prior close. The 4 October bar was still open at retrieval, having traded $0.05657 and last at $0.05381.
The rolling ticker should not be added to those daily bars.
The +24.6% is the move off a $0.04318 open in a window from 07:31 UTC on 3 October to 07:31 UTC on 4 October. Its $21.56 million of quote volume overlaps the settled 3 October candle and the unfinished 4 October bar. The 285,833 trades are a count of prints in that window, not a second pile of dollars.
A calendar of pages, and the days the tape ignored
No Starknet or StarkWare page opened for this note states a reason for the 18 September session. That is a finding about those pages, not about every headline on the internet. The starknet.io post sitemap lists 504 posts, and two carry a last-modified time after 1 August: the strkBTC user guide, visibly dated 12 May 2026, and a quantum-resistance note dated 7 August. The 18 September stamp, 09:59 UTC, sits on the bridges directory, which lists StarkGate, a strkBTC bridge, Stargate and NEAR Intents. A directory timestamp is not a launch.
The closest protocol document is older than the spike. On 9 September the forum carried prerelease notes for v0.14.4, from the account OdedSW. The note calls the version small. The new user-facing item is a SNIP-36 proof of a full-size block, one transaction up to 1.1 billion L2 gas. Testnet is dated 15 September 2026. Mainnet is dated 5 October 2026, pending governance approval, and the note calls the fee changes cost alignment rather than an increase. A search for "v0.14.4" on 4 October returned that thread and no later approval. The mainnet date had not arrived.
STRK closed at $0.03126 on 9 September and at $0.02796 on 10 September, down 10.6%. On 15 September, the testnet date in the note, the close was $0.02665. The 52.6% session came on the 18th. The second leg, on 3 October, came two days before the written mainnet date. A schedule that was public for nine days before the first spike can be context for the 3 October extension. It is not, on the closes, the explanation of the 18th.
Other StarkWare posts in the fortnight missed the heavy sessions. On 10 September the company wrote that seven StarkWare and Starknet Foundation contributors are co-authors of ECDSA.fail, a paper the post says cut Google's March quantum-attack cost estimate by more than half. That was the 10.6% down day. On 16 September StarkWare, Yukon Research and Eigen Labs opened a Quantum-Safe Bitcoin Optimization Challenge with $20,000 from StarkWare plus a further Yukon prize. The post puts the first such Bitcoin transaction at roughly 3,100 GPU-hours and about $320. STRK closed at $0.02791. On 23 September a follow-up put the estimate near $67, and the headline says 79% lower. STRK closed down 9.8%. None of the three mentions the STRK market.
The operational note is an oracle incident, and it also refuses a tidy cause. On 14 September a forum account displayed as Henri, Starknet foundation, wrote that the Security Council had temporarily restricted six addresses after a Vesu / Pragma oracle incident on 4 September, and that the restrictions were later lifted. The note gives no loss figure and does not mention the token. The incident date was a down close, open $0.02766 and close $0.02599. The 14 September post landed on a close of $0.02884.
Where this desk could attach a name, the project had written one down. Optimism's rise from $0.096 to $0.126 followed Upgrade 20. Worldcoin's late-September rise was tied to World Money. The STRK file opened here does not have an equivalent sentence.
The named remarks StarkWare did put under a title are from February, and they are about research staffing. On 5 February 2026 the company said Prof. Scott Aaronson, described there as the Schlumberger Chair of Computer Science at the University of Texas at Austin and founding director of its Quantum Information Center, had joined the scientific advisory board.
"We are honored to welcome Scott to StarkWare's Scientific Advisory Board," said Eli Ben-Sasson, co-founder and CEO at StarkWare. "His deep scientific insight and decades of pioneering research in theoretical computer science and quantum computing align perfectly with our commitment to researching the effects of post-quantum computing on blockchains and our commitment to helping Bitcoin become post-quantum secure."
"StarkWare is tackling some of the most profound challenges involving cryptography, blockchains, and quantum computing," said Aaronson. "I look forward to contributing to the company's scientific vision and supporting its efforts to understand the risk post-quantum computing poses to blockchains and cryptography."
Both passages are on StarkWare's 5 February post. Neither mentions a price. Nothing opened from September carries a comparable named attribution. A later sitemap touch on that URL is an edit time, not a second announcement.
The levels are old prices
The bull case of $0.09 is the 6 January open of $0.0907, shaved to the cent, about 6.5% under that day's $0.0963 high and 67% above the 4 October last. It is a prior print, not a new high and not a model output.
The base case of $0.062 sits a hair above the 30 January close of $0.0616, the last settlement before the break under six cents. It is 15% above the 4 October last. It says the repair finishes through that close and stops short of the January highs. It is the path with the most weight.
The bear case of $0.03 sits inside the 4 to 17 September closes, from $0.02599 to $0.03192. Giving back the 18 September gap lands there, about 44% under the current last. It is not the 12 August low of $0.02217, which is also the low of the whole Binance STRKUSDT history back to the February 2024 listing. On 10 October 2025 a wick touched $0.0297 inside a candle that closed at $0.1114. The cent was visited. The market did not live there.
Daily closes on Binance STRKUSDT from 1 January 2026 through the 4 October session. Bull at $0.09, base at $0.062 and bear at $0.03 run to 31 December 2026. The 4 October point is the last trade at retrieval, not a settled daily close. The January peak on the chart is a close, so the $0.0963 intraday high is not drawn.
| Date | STRK close | ETH close | BTC close | Starknet TVL |
|---|---|---|---|---|
| 17 Sep 2026 | $0.02899 | $2,447.28 | $76,417 | $155.04 million |
| 18 Sep 2026 | $0.04425 | $2,612.14 | $80,884 | $156.97 million |
| 3 Oct 2026 | $0.05428 | $2,687.23 | $84,754 | $162.72 million |
| 4 Oct 2026 | $0.05381 | $2,693.13 | $84,945 | $164.46 million |
Binance daily candles, retrieved 4 October 2026. The 4 October row is unfinished, and the STRK figure is the 07:31 UTC last. TVL is the historical series. From 17 September to that last, STRK rose 85.6% and historical TVL rose 6.1%.
One longer gap, said once: the first Binance daily close, 20 February 2024, was $1.965, and that candle's $7.71 high was a listing wick followed by a $1.892 close. The 4 October last is 97.3% below the first close. This forecast is about the 2026 range, not the listing.
More turnover, not more deposits
After 18 September, STRKUSDT turnover stopped looking like the earlier year. The median through 17 September was just under $2.00 million, and 3 October still cleared $16.09 million. DefiLlama's historical series did not match that. Its 2026 high is $323.90 million on 20 January, when STRK closed at $0.0772. The 4 October point, $164.46 million, is 49.2% lower. The year's low, $155.04 million, was 17 September, the session before the spike.
That is an awkward fact if the marginal trade was a wave of fresh deposits.
Who took the other side of the $33.67 million is not visible here. The comparable deposit series barely moved, so the session looks like a change of hands in the token rather than a larger pool bridged in. The live chains endpoint, at $189.41 million, sits above the historical point. If that gap is a real inflow, deposits are less flat than the daily file says. The gap is unresolved, and neither print matches a token that rose by half in a session.
Volume without a matching page is not new on this book. The largest 2026 quote-volume day in the series was 8 May, $54.67 million, closing at $0.0561. The strkBTC guide is dated 12 May, and that later session turned over $5.07 million. The 7 August quantum-resistance note, about an experimental transfer costing about six cents, landed on a $0.02475 close and $555,207 of turnover. September rhymes with both.
RelatedPolkadot (DOT) Price Prediction: $1.95 Bull vs $0.85 Bear
Injective's September session had the reverse awkwardness, a vote attached to code the project had stopped publishing. Here the pages exist. They are dated, and they are about quantum-safe Bitcoin research, a small protocol version, and a closed oracle restriction. They are about something other than the candle.
Weights to 31 December
The horizon is 31 December 2026. These weights are a reading of the paths, not a market-implied probability and not an instruction. The 4 October candle had not settled at the anchor.
Bull case, $0.09, weight 25%. That is a return to the 6 January open, about 67% above the last. No page opened here documents a use or supply change that forces it. The quarter-weight is there because the book traded the price in January and because turnover since 18 September is a different regime. It is not the centre.
Base case, $0.062, weight 45%. A hair above the 30 January close of $0.0616, the last settlement before the break under six cents. Spot is already about 2% under the 31 January close, so the remaining distance is the smaller one. Flat deposits are why the weight stops here rather than higher.
Bear case, $0.03, weight 30%. A giveback into the early-September band. The invalidation is a daily close under $0.03766, the 19 September low. That low was traded intraday and the close was still $0.04471, so a poke under $0.038 is not the line. A close beneath $0.03766 is. The breakout structure this reading respects would be gone.
The weighted level is 0.25 times $0.09, plus 0.45 times $0.062, plus 0.30 times $0.03. That is $0.059, about 10% above $0.05381. Bias is bull because the weighted level and the base sit above spot. The target field records $0.09, the upside path, not the path with the most weight. Conviction is 2 of 5. The tape is real. The confirming use data is not.
A passed record for the v0.14.4 mainnet, or a primary post with a supply or usage figure, would raise the weight on $0.09. The 9 September note does not qualify while its mainnet line says pending. A daily close under $0.03766, or historical TVL back through the $155.04 million low of 17 September while the token stays elevated, would raise the weight on $0.03.
Questions the tape actually raises
Did Starknet (STRK) double because value locked on the chain doubled?
No. From the 5 September open of $0.02599 to the 4 October last of $0.05381, Binance shows a 107% rise. DefiLlama's historical series moved from $160.44 million to $164.46 million, a 2.5% change. The live chains endpoint read $189.41 million and disagrees with that historical point. Neither print is a doubling of deposits. From the 5 September close, rather than the open, the token's rise is 78.7%, which is not a double at all.
Why is the bull case $0.09 rather than a new high?
Because $0.09 is a price this book already printed in 2026. The 6 January candle opened at $0.0907 and set the year's high at $0.0963 before closing at $0.0926. The bull case asks for a return to that open, about 67% above the 4 October last, not for an extension past the high. It carries a 25% weight in this note. The centre of the reading is the $0.062 base, not this upside path.
What does the $0.03 bear case correspond to on the tape?
It sits inside the 4 September to 17 September closes, which ran from $0.02599 to $0.03192. A giveback of the 18 September gap lands in that band. It is not the 12 August low of $0.02217, the low of the Binance daily history. The 10 October 2025 low of $0.0297 was a wick in a candle that closed at $0.1114, so the cent has been visited without the market living there.
Did the v0.14.4 mainnet date cause the September move?
The 9 September prerelease note schedules mainnet for 5 October 2026, pending governance approval, and testnet for 15 September. STRK closed down 10.6% the next day and closed at $0.02665 on the testnet date. The 52.6% session was 18 September. A forum search on 4 October found no later approval thread. The written date can sit beside the 3 October leg. It does not, on these closes, explain the 18th.
Where does this reading fail?
A daily close under $0.03766, the 19 September low, removes the breakout the base case uses and shifts weight toward $0.03. The reading fails the other way if a primary post documents a supply, upgrade, or usage change these pages did not contain. The two DefiLlama prints already disagree, $164.46 million against $189.41 million, so the deposit side is only as solid as the series being used.
This is analysis, not a recommendation. Capital is at risk. The levels are a reading of Binance candles and of pages opened on 4 October 2026, not an instruction to take either side. The 4 October candle had not settled at 07:31 UTC, and the two DefiLlama prints disagreed that morning.