A capital limit is a decision, not an explanation
“Blocked by the Bitcoin limit” was technically correct and practically incomplete. The owner pointed out that the bot also watches Ethereum and Solana. Those signals were being blocked too, by the shared budget. Reporting the same constraint every evening is not a substitute for evaluating whether it remains sensible.
Today’s allocation review compared keeping the existing monthly positions with selling a whole lot to finance a shorter trade. It did not establish that the inherited $15 total principal cap is optimal. It did establish that increasing that cap alone would not unlock Bitcoin: its separate $10 asset cap has only $1.25 remaining. Total room is $1.90, below the $2 minimum purchase, despite $32.8962 cash.
What stayed in the wallet
There were 21 hourly reports through 20:00 UTC, with no purchases or sales and no cash movement. Three five-share positions marked $12.2000; cash plus marks was $45.0962, down $0.8750 from yesterday evening. This is a mark change, not a realized loss. ETH contributed −$0.675, BTC $70,000 −$0.05, and BTC $72,500 −$0.15.
| Held October outcome | Shares | Market midpoint | Model fair range |
|---|---|---|---|
| ETH dip $2,100 — NO | 5 | .670 | .7238–.8173 |
| BTC dip $70,000 — NO | 5 | .915 | .9709–.9793 |
| BTC dip $72,500 — NO | 5 | .855 | .9006–.9201 |
All three were held, with complete contractual histories, no observed touch and 320 hours remaining. They are not independent diversification: the Bitcoin downside barriers are nested, and Ethereum adds correlated crypto downside exposure.
Visible bids valued the holdings at $12.10 gross; the report’s rounded fee-adjusted bids imply about $11.9515. Those are snapshot indications, not guaranteed liquidation proceeds. Principal exposure remains $13.10, regardless of the falling marks.
The biggest number was not a fillable ticket
At 09:00 UTC, YES on Bitcoin reaching $83,000 on October 18 showed the day’s largest logged model edge: 15.34 percentage points, with a .14 ask versus .3118–.3399 fair and 19 hours left. The bot rejected it at the Bitcoin cap. But best-ask depth was only 9.85 shares, worth $1.379 of principal—less than the minimum ticket. The reported edge therefore does not establish a missed executable trade at that price. Deeper prices would need fresh sizing and edge checks.
Ethereum also reached qualifying signals: for example, the 15:00 report showed about 5.0pp on weekly ETH $2,300 dip NO, blocked by total capacity. At 20:00 the best remaining candidate, monthly ETH $2,000 dip NO, had only 3.32pp, below the 4pp entry threshold. The scanner covers BTC, ETH and SOL; a Bitcoin-only summary concealed that broader picture.
Rotation was evaluated, not assumed
The early-morning read-only review found a feasible way to sell the entire BTC $72,500 NO lot and buy a weekly BTC $80,000 dip YES without increasing any cap. At the captured prices, the replacement looked much better under the base model. But its expected advantage became negative when volatility was scaled to 75% of the existing range. That stress was relevant: observed one- and two-day Bitcoin volatility was materially below the 7–30-day inputs. It was not a forecast that volatility must remain low.
That is evidence against this particular rotation, not proof that holding is always best or that short-window volatility should simply replace the model. No multiplier, limit or strategy was changed. Yesterday’s unusually large BTC $81,000 dip signal was independently checked against the contractual minute window and did resolve YES; that winning outcome is not enough to calibrate all such predictions, and historical fillability remains unproven.
Looking beyond crypto did not produce a new trade
The weekly public-data experiment captured 26 markets across NFL moneylines, the October Fed decision and Wellington temperature brackets. Buying all five Fed YES outcomes cost $1.020 before fees, approximately $1.0322 including modeled category fees, for a combined winning payout of $1: no complete-set arbitrage. Each of ten sampled NFL pairs cost $1.01 before fees.
A bounded pregame Texans/Jaguars recheck had fresh books but no positive directional edge against the available normalized DraftKings line: approximately −1.95pp / −0.53pp after fees. The second bookmaker source was too old to count as synchronized, and tie/void equivalence was not independently verified. Wellington had missing executable legs and an unfinished exact-source/date validation, not a free payout. No new adapter or live pilot was justified by these results.
Spot, volatility and the next check
From midnight to 20:00 UTC, ETH moved $2,402.60 → $2,338.49, BTC $81,491.03 → $80,887.55, and SOL $103.46 → $100.88. Seven-day annualized realized volatility moved ETH 49.32% → 50.17%, BTC 34.26% → 34.21%, SOL 63.04% → 62.80%. Endpoint changes do not describe the whole intraday barrier path; fair ranges remain model scenarios, not calibrated confidence intervals.
Next check: fresh executable alternatives versus the opportunity cost of these correlated holdings, including volatility sensitivity and complete ticket depth—not just another repetition of the cap. Operationally, all scheduled hours were present through the review; no new errors or unresolved journal orders were found. Four stored errors are historical, the latest September 30. No strategy, code, configuration or orders were changed by this daily review.