At a glance
| Weekly Iron Butterfly (SPY) | |
|---|---|
| Live paper trading, by the rules | +$948 on 5 trades, 5 winners |
| Off-rule trade (Sep 21), shown separately | +$57 |
| Weeks traded of 8 (Sep 14 – Nov 2) | 5 by the rules, 1 off-rule, 2 skipped on the credit floor |
| Average win / largest win | +$190 / +$283 |
| Maximum loss per trade (wing − credit) | $303–$478, average $384 |
| Worst peak-to-trough, daily account closes | −$107 (Oct 26–28) |
| Worst day, by the rules | −$59 (Oct 27) |
| Backtest, current rules (33 weeks) | +$1,618 on 15 trades, 13 winners |
Five for five, but one exit carries half the result. The profit target turned what would have been a maximum loss of about $303 into a gain of $152, a swing of about $455. Without it, the live record would be +$493 with one losing week.
The target fired because of the path, not the destination. SPY opened Thursday near 765 and fell through the butterfly’s 759 center on its way down. That is the moment the position was worth the most. If SPY had dropped overnight instead, below the 748 wing, the target would never have fired, and with no stop loss there would have been no exit until 3:40.
Market-neutral doesn’t mean market-risk-free. Wins near $190 against losses up to about $380 mean one bad week erases two good ones. Eight weeks and five trades show how the strategy behaves. They don’t show that it has an edge.
How the strategy works
The rules are unchanged from our October 23 note except for the profit target, added on October 30.
| Structure | Sell the at-the-money SPY straddle; buy a put and a call about 1.5% of SPY away (about $11–12 wide) |
| Expiry | That week’s Friday |
| Entry | Monday 9:40 ET (Tuesday after a Monday holiday or a failed Monday run), one 4-leg limit order |
| Credit floor | At least $700 per contract; re-quotes each minute for 20 minutes, then skips the week |
| Profit target (new) | Checked every minute at mid prices; closes once the position is up $150 per contract |
| Scheduled exit | Thursday 15:40 ET, a 4-leg limit order that starts at the mid and steps toward the market |
| Stop loss | None |
| Account | Our own paper account, 1 contract |
What it’s betting on. The butterfly collects about $700–800 up front and earns the most if SPY finishes near the entry price. Its break-evens sit about $7–8 either side of the center, roughly ±1% of SPY. Beyond the wings, about ±1.5%, the loss is capped at the wing width less the credit, about $300–480 per contract.
What happened on November 5
The butterfly opened Monday, November 2, at 9:40 a.m. with SPY at 759.16. Strikes were 748 / 759 / 770, with a $797 credit, so the maximum loss was $303. SPY drifted down to about 752 on Monday and Tuesday and rallied to 768 on Wednesday. Thursday it opened near 765, fell through 759 within minutes and kept falling, about 4% on the day.
SPY, November 2–5, against the butterfly’s strikes
| Thursday, Nov 5 | SPY | Position value | P&L |
|---|---|---|---|
| 9:43 a.m. — profit target fills | 757.78 | $645 | +$152 |
| 3:40 p.m. — scheduled exit (if held) | 734.45 | about $1,100 | about −$303 |
The 9:43 row is the actual fill. The 3:40 row uses the legs’ last trades at 3:40 p.m. (759 put $24.69, 748 put $13.83, 759 call $0.17, 770 call $0.03). That puts the butterfly at its full $11 width, the maximum loss.
Why the target fired. A butterfly is worth the least to buy back when SPY sits on its center strike. SPY passed through 759 at about 9:40 on the way down, with three days of time decay already earned, and the position reached +$150. About 40 minutes later, with SPY near 749, it was losing money.
When it wouldn’t have. The same week with a different path would have gone very differently. Had SPY gapped from Wednesday’s 767 close straight to 745 at Thursday’s open, the position would never have passed through +$150. With no stop loss, it would have taken the full maximum loss at 3:40. The target protects against a sell-off that starts near the center, not one that starts from a gap.
Across our SPY options strategies. All three of our other SPY condor and butterfly strategies, the daily iron condor, the daily iron butterfly and the weekly iron condor, lost money on November 5. We’ll compare them in a follow-up note.
Live paper trading
One contract on our own paper account. Results come from Alpaca fills, so they’re net of the spread actually paid.
P&L per trade, one contract
| Week of | Strikes | Credit | Max loss | SPY in → out | Exit | P&L |
|---|---|---|---|---|---|---|
| Sep 14 | 749/760/771 | $741 | $359 | 760.10 → 762.89 | Thu 3:40 | +$283 |
| Sep 21 (off-rule) | 758/770/782 | $596 | $604 | 769.55 → 772.45 | Wed 9:34 | +$57 |
| Sep 28 | 755/767/779 | $722 | $478 | 767.38 → 764.81 | Thu 3:40 | +$226 |
| Oct 5 | Skipped: best Monday credit $680, under the $700 floor | — | ||||
| Oct 12 | Skipped: best Monday credit $648 | — | ||||
| Oct 19 | 755/767/779 | $733 | $467 | 767.33 → 761.90 | Thu 3:40 | +$136 |
| Oct 26 | 752/763/774 | $788 | $312 | 762.82 → 768.17 | Thu 3:40 | +$151 |
| Nov 2 | 748/759/770 | $797 | $303 | 759.16 → 757.78 | Thu 9:43, target | +$152 |
| By the rules | 5 trades, 5 winners | +$948 | ||||
Strikes are long put / short straddle / long call. The September 21 trade was entered at 11:08 ET, under the credit floor, with a resting exit order, so it is excluded from rule-based totals, as in our October 23 note. Account equity rose $1,001 over the period; the difference from the trade total is fees and about $1 of unrelated stock fills.
Would the target have changed the earlier trades?
The first four trades predate the target. Rebuilt from minute-by-minute option prices, each one first reached +$150 at or just after Thursday’s open. With the target in place they would have closed then, at roughly the values below.
| Week of | First at +$150 | Est. exit at target | Actual (held) | Difference |
|---|---|---|---|---|
| Sep 14 | Thu 9:30 | about +$203 | +$283 | about −$80 |
| Sep 28 | Thu 9:30 | about +$193 | +$226 | about −$33 |
| Oct 19 | Thu 9:30 | about +$192 | +$136 | about +$56 |
| Oct 26 | Thu 9:42 | about +$157 | +$151 | about +$6 |
| Nov 2 | Thu 9:43 | +$152 (actual) | about −$303 if held | about +$455 |
Estimates use the legs’ last trade price each minute, not quoted mids, so they are approximate. A Thursday-open gap past +$150 fills above the target.
In calm weeks the target is close to a wash: it gives up some of the last day’s decay in some weeks and locks in gains that later faded in others, about −$50 across the four. Its value is concentrated in weeks like November 2. So far the strategy has made its money overnight into Thursday. No live trade reached +$150 before Thursday morning.
What the backtest says about the target
Our nightly backtest re-prices every Monday entry from January 5 to September 11, 2026 under the live rules, at minute-level mid prices. Adding the target didn’t change which 15 weeks traded. It changed how the winners ended.
| 15 trades, $700 floor, gross | Hold to Thursday | +$150 target |
|---|---|---|
| Total P&L | +$1,053 | +$1,618 |
| Winners | 12 (80%) | 13 (87%) |
| Closed at the target | — | 13 of 15 |
| Average win / average loss | +$141 / −$212 | +$160 / −$232 |
| Worst week | −$258 (Apr 13) | −$258 (Apr 13) |
| Max drawdown | $464 | $464 |
Three cautions.
- The target was chosen on these same 33 weeks. A threshold tuned on the data it is tested on flatters itself. The live edge, if there is one, should be smaller.
- It didn’t touch the losing weeks. Both backtest losses, the April 6 and April 13 weeks, never reached +$150. They were held to Thursday and lost the same amount either way. November 5 is the first case, live or backtest, where the target turned a loser into a winner. That is one observation, not a pattern.
- It caps the upside but leaves the downside in place. With the target, wins cluster around +$150–200 while a full loss is still $300–480. The strategy needs to win about three trades in five just to break even. The backtest won 87%, but on 15 trades, with no crash in the data.
The published backtest page also shows an annualized return above 1,000%. As we noted on October 23, that figure scales a small per-trade result by 52 weeks and divides by the capital at risk. It isn’t a realistic expectation, and we don’t use it here.
Operational notes since October 23
- The profit target is live. It was added to the strategy and modeled in the backtest on October 30 and first traded the week of November 2.
- Exits now go out as multi-leg limit orders. The September 17 and October 1 exits were closed with separate single-leg market orders. Each needed a second attempt for one long wing. Since October 22, exits have been multi-leg orders. On October 29, three legs closed together and the 752 put, bid at $0.04, was sold separately ten minutes later. On November 5 all four legs closed in one order.
- The credit floor is still a fixed $700. The strategy skipped October 5 ($680) and October 12 ($648). As our October 23 note explained, a dollar floor drifts with SPY’s price and implied volatility. That is still on our list to revisit.
- Still no stop loss. Positions are held from Monday to Thursday with three nights of gap risk. The worst case is the full wing less the credit, about $300–480 per contract at today’s prices.
What we’re watching
- A gap-down week. The case the target can’t handle. We want to see how often, historically, SPY skipped past the center between Wednesday’s close and Thursday’s open.
- A longer backtest. Thirty-three weeks of 2026 aren’t enough for a strategy whose risk is one bad week. As we said in October, we plan to extend it back to 2022.
- The target out of sample. We will report every target exit, and every week the target doesn’t fire, against holding to Thursday, so the comparison stays honest.
- Daily vs. weekly. Our next note compares the four SPY condor and butterfly strategies, including what each did on October 30 and November 5.
Sources and methodology
- Scope. The Weekly Iron Butterfly (SPY) strategy as it trades today, on our own 1-contract paper account.
- Live. Alpaca paper-trading fills and account history, September 14 – November 6, 2026. The September 21 off-rule trade is shown but excluded from totals. An August 31 test trade that predates the strategy is excluded. Skipped-week credits come from the strategy’s entry logs.
- November 5 counterfactual. Last trade prices of the four legs at 3:40 p.m. ET, Alpaca OPRA data. SPY prices are 5-minute SIP closes.
- Target estimates for earlier trades. Rebuilt each minute from the legs’ last trade prices. Wide-spread wings trade rarely, so the estimates are approximate.
- Backtest. Our nightly backtest builder, which reads the live strategy’s parameters: 33 Monday entries, January 5 – September 11, 2026, minute-level mid prices, no commissions or slippage. The hold-to-Thursday column is the builder’s output before the target was added on October 30.
- Drawdown and worst day. Daily closing equity of the paper account, which holds only this strategy apart from a few small stock fills.
See how every strategy we run is doing, updated each trading day, on the live leaderboard.
Results are from paper-trading accounts and backtests and do not reflect real money. Paper and backtest results may differ materially from live trading. Past performance does not guarantee future results. This note is research and education, not investment advice or a recommendation to trade any strategy or security. Options involve substantial risk and are not suitable for all investors.