Four Months In: Wheel Strategy by the Numbers

Posted on Dec 24, 2025

I started systematically running the wheel strategy at the end of August. After 120 trading days and 140 total transactions, I wanted to see how it’s doing so far.

The Setup

Pure wheel strategy: sell cash-secured puts, get assigned if ITM, sell covered calls until called away. I automated the core execution in late August which eliminated most of the manual tracking overhead. The system has executed 68 opening sells and 71 closes across that period.

Core Metrics

Win rate: 82.8% (24 wins, 5 losses across 29 completed cycles)

Average winning trade: +$93 Average losing trade: -$1,069

Best single trade: BWXT put at $190 strike, netted $207 profit Worst single trade: AMD call at $170 strike, lost $4,842 (had to roll defensively)

The win rate is solid but the risk/reward is asymmetric. Small consistent wins offset by occasional larger losses when positions move against you. That’s the nature of selling premium.

Monthly Breakdown

August (partial month): +$457 September: +$2,577 October: +$7,380 November: +$1,411 December (through 12/19): +$2,355

Total realized P&L: +$15,854

October was an outlier - nearly 3x September’s performance. I’m not modeling future months at that level. The four-month average is about $3,964/month, though that’s skewed by October.

Position-Level Performance

Ranked by net realized P&L on closed positions:

  • GOOG: +$1,229 (19 total trades)
  • NVDA: +$965 (17 trades)
  • AMD: +$922 (17 trades, includes the $4,842 loss)
  • BWXT: +$597 (7 trades)
  • AAPL: +$583 (19 trades)
  • RDDT: +$1,717 (8 trades, position fully closed)
  • NVO: +$235 (6 trades, position fully closed)
  • ENPH: +$189 (7 trades)
  • DXCM: +$64 (5 trades)
  • ORCL: +$61 (4 trades)

RDDT was surprisingly profitable relative to the number of trades. NVDA and GOOG generated the most activity. AMD would’ve been much higher without that one defensive roll.

Current Exposure

Open cash-secured puts on BWXT ($160 strike), NVDA ($160), and ORCL ($170) representing $49K in committed capital if all assign. That’s roughly half my liquid allocation.

Holding shares in AAPL (100 shares), GOOG (100), and GLD (100) with covered calls at $330, $420 strikes respectively. Total stock position value at cost basis is about $73K.

Five active options currently generating $908 in collected premium.

Risk Analysis

Largest single-day exposure was the AMD roll at about 5% of account value. I’ve kept individual position sizing such that no single assignment would exceed 20% of total capital. The wheel concentrates risk in a handful of names, so I’m watching correlation - AAPL, GOOG, and NVDA all move with tech sentiment.

Capital efficiency is decent - I’m keeping about 30% in cash to cover puts while generating returns on the stock positions. Could lever more but I’m not interested in margin calls.

System Performance

The automation handles about 95% of executions. Manual interventions were mainly around defensive rolls (3 instances) and early assignment decisions (2 instances). The code had a bug where it wasn’t recording symbols on some trades - fixed that, but it highlighted the importance of data validation.

Transaction costs are negligible (under $1 per trade on average). The real cost is capital lock-up and opportunity cost when assigned at unfavorable prices.

Forward Expectations

Based on 29 completed cycles, expected value per cycle is roughly +$171 accounting for win rate and average outcomes. At current velocity of about 7 cycles per month, that projects to ~$1,200/month assuming October was an anomaly.

More realistically, I’m modeling $1,500-2,000/month going forward with high variance. Some months will hit $3K+, others might be flat or slightly negative depending on how many defensive rolls I need to execute.

The strategy scales with capital but I’m keeping position count at 5-7 names for now. More positions means more monitoring overhead and correlation risk.

Takeaways

120 days, 140 trades, 83% win rate, +$15,854 realized. The automation works, the math works, but it’s not passive. You’re constantly managing strike selection, expiration timing, and risk when positions move against you.

That AMD loss was 30% of total gains. You can’t avoid losses in this strategy, you can only size them appropriately. One bad roll can erase weeks of premium collection.

Still running it. The data supports continuing, but I’m not scaling up position sizes until I see how this performs through a real drawdown period. Four months of mostly-up market isn’t a stress test.