A program sells cash-secured puts by fixed rules on a simulated $100,000 Alpaca account. This page is built from the broker's own history every ten minutes; nothing on it is entered by hand, and nothing is removed. The rules were fitted on past trades and never tested on new ones. This year is that test.
Day 14 of 365. The verdict is published here on 2027-09-08: return per dollar of capital per day, with its interval, against the same money in SPY and in Cboe's PutWrite index. Beating the index is the point: it sells puts with no rules, so a result that does not beat it says the rules add nothing.
Average result per closed position: -$116.65. No interval yet: 2 sessions is too few to state an interval; 5 is the minimum.
Every line starts as the same $100k on 09-04, the last close before the first order. The robot's line is its closing equity each session, open positions at their marks.
What the PUT index is. Cboe's S&P 500 PutWrite index sells an at-the-money S&P 500 put, holds it to expiry, and keeps the full strike in Treasury bills, then does it again. It is the robot's trade with every decision taken out: no choice of name, strike or date. It is published by Cboe every day. WPUT is the same with one-week puts, closer to the robot's own 4 to 18 days. Between them they separate what the term does from what the choice of names does: if the robot trails WPUT as well, the term is not the excuse.
How to read the lines. Robot above both put indices: the rules add something over selling puts blindly. Robot below them but above SPY: selling puts paid, the rules did not. Robot below all three: in this period the rules cost money against doing nothing clever. A few weeks decide nothing; the comparison that counts is the one on 2027-09-08.
The same days, from the same all-cash close. Funds are measured with their dividends reinvested, because a Treasury-bill fund pays nearly all its return that way.
| Instead | Same days | Robot against it |
|---|---|---|
| Treasury bills (BIL) The floor: cash earning the short rate, no risk. A robot that cannot beat it took risk for nothing. | +0.14% to 09-22 | -1.11 pts |
| Equal-weight S&P 500 (RSP) The market without its megacap tilt. The robot's positions are single names of roughly equal size, so this is the fairer stock benchmark. | -2.45% to 09-22 | +1.49 pts |
| S&P 500 BuyWrite index (BXM) Cboe's covered call on the index: the other textbook way to sell options for income, with the same capped-upside shape. | +0.65% to 09-21 | -1.61 pts |
| JPMorgan Equity Premium Income (JEPI) The option-income fund retail is actually sold. The question it answers: why not just buy this? | -0.70% to 09-22 | -0.26 pts |
Newest first. An assigned put is scored with its shares: the premium, plus what the shares lost or made between assignment and sale.
| Opened | Contract | Credit | Capital held | Vol at entry | Outcome | Days | P&L | Per $ per day |
|---|---|---|---|---|---|---|---|---|
| 09-22 | KRE 71 put exp 10-02 open, marked at $1.01 | $79 | $7,100 | 14.5% | open | 1 | -$22 mark | -0.310% |
| 09-22 | XLE 61.5 put exp 10-02 open, marked at $0.96 | $69 | $6,150 | 22.1% | open | 1 | -$27 mark | -0.439% |
| 09-21 | C 131 put exp 09-25 open, marked at $0.93 | $78 | $13,100 | 18.7% | open | 1 | -$15 mark | -0.115% |
| 09-21 | PEP 127 put exp 09-25 open, marked at $0.18 | $48 | $12,700 | 16.4% | open | 1 | +$30 mark | 0.236% |
| 09-09 | XLF 56.5 put exp 09-18 assigned 09-18: 100 shares at $56.50; sold 09-21 at $55.73 | $32 | $5,650 | 13.0% | assigned | 12 | -$45 | -0.066% |
| 09-08 | BA 210 put exp 09-18 assigned 09-18: 100 shares at $210.00; sold 09-21 at $198.65 | $345 | $21,000 | assigned | 13 | -$790 | -0.289% | |
| 09-08 | GME 19 put exp 09-18 expired worthless 09-18, full credit kept | $51 | $1,900 | expired | 10 | +$51 | 0.268% | |
| 09-08 | JEPI 57 put exp 09-18 assigned 09-18: 100 shares at $57.00; sold 09-21 at $56.39 | $25 | $5,700 | assigned | 13 | -$36 | -0.049% | |
| 09-08 | PFE 28 put exp 09-18 assigned 09-18: 100 shares at $28.00; sold 09-21 at $27.68 | $36 | $2,800 | assigned | 13 | +$4 | 0.011% | |
| 09-08 | T 25 put exp 09-18 expired worthless 09-18, full credit kept | $19 | $2,500 | expired | 10 | +$19 | 0.076% | |
| 09-08 | TLT 82 put exp 09-25 open, marked at $0.48 | $54 | $8,200 | open | 14 | +$6 mark | 0.005% | |
| 09-08 | VICI 25 put exp 09-18 assigned 09-18: 100 shares at $25.00; sold 09-21 at $23.90 | $35 | $2,500 | assigned | 13 | -$75 | -0.231% | |
| 09-08 | XLE 64.5 put exp 09-18 assigned 09-18: 100 shares at $64.50; sold 09-21 at $63.00 | $89 | $6,450 | assigned | 13 | -$61 | -0.073% |
Read from the trading Worker's own configuration, so this list cannot drift from what it does.
| Date | What changed |
|---|---|
| 2026-09-08 | The test starts with the rules below. Nine puts are sold that day and the next. |
| 2026-09-10 | Two share round trips (SGOV, VOO) placed by hand to measure how the broker counts collateral. Not the robot's decisions; shown separately below. |
| 2026-09-18 | The capital ceiling counts shares at their full price, not at the broker's margin figure, which had read the first assignments as capital being freed. |
| 2026-09-18 | Shares delivered by an assignment are sold at market at the next session's first run, instead of being held. |
Placed by hand to measure how the broker counts collateral. They are in the account's equity and nowhere else on this page.
| Date | Shares | Result |
|---|---|---|
| 09-10 | SGOV bought 200, sold 200 | -$2.00 |
| 09-10 | VOO bought 29, sold 29 | +$0.72 |
Fees across the account so far: -$1.51. Simulated trades on a paper account: the fills are Alpaca's paper fills, not real executions, and a real account would have filled some orders worse and some not at all. This is the public record of a test, not advice and not a recommendation to buy or sell any security. The same engine reads your own broker export on the ledger.