{
  "status": "Archived source-reported simulation, not independently reproduced; not validated live performance",
  "record_date": "2026-09-09",
  "review_date": "2026-11-16",
  "period_as_recorded": "2008-2025",
  "currency": "USD nominal",
  "rows": [
    {
      "configuration": "SPY buy and hold",
      "annual_compound_growth_pct": 11.11,
      "max_drawdown_pct": -51.9,
      "nominal_annual_premium_budget_pct": null,
      "realised_annual_premium_spend_pct_average_portfolio_value": null
    },
    {
      "configuration": "SPY with budgeted deep-put protection",
      "annual_compound_growth_pct": 14.49,
      "max_drawdown_pct": -30.6,
      "nominal_annual_premium_budget_pct": 2.0,
      "realised_annual_premium_spend_pct_average_portfolio_value": 1.6
    },
    {
      "configuration": "Standing 98% equity / 2% option allocation",
      "annual_compound_growth_pct": 7.98,
      "max_drawdown_pct": -46.2,
      "nominal_annual_premium_budget_pct": null,
      "realised_annual_premium_spend_pct_average_portfolio_value": 11.37
    }
  ],
  "budgeted_configuration": {
    "otm_pct": 30,
    "entry_dte": [
      90,
      180
    ],
    "exit_dte": 30,
    "entry_frequency": "Every two months",
    "funding": "From portfolio, as stated in the archived research record; not external contributions",
    "annual_budget_pct": 2.0
  },
  "standing_configuration": {
    "equity_weight_pct": 98,
    "option_weight_pct": 2,
    "otm_pct": 30,
    "entry_dte": [
      90,
      120
    ],
    "exit_dte": 60
  },
  "included_costs_as_recorded": [
    "Option premiums",
    "Quoted bid-ask spreads"
  ],
  "excluded_costs_as_recorded": [
    "Taxes",
    "Commissions",
    "Additional slippage",
    "Capacity constraints"
  ],
  "derived_differences_pct_points": {
    "full_period_cagr": 3.38,
    "maximum_drawdown_reduction": 21.3
  },
  "recorded_subperiod_results": {
    "post_2011_spy_cagr_advantage_pct_points": 0.37,
    "post_2011_qqq_cagr_advantage_pct_points": -1.62,
    "post_2011_qqq_drawdown_improvement": false
  },
  "limitations": [
    "Retained research record, not a simulation rerun for this website.",
    "The option-quote upstream source is undocumented.",
    "The source reports intrinsic-value fallbacks on roughly 10% of deep-put holding days, which can distort drawdowns.",
    "Exact daily portfolio path, drawdown observation frequency and endpoints are not independently verified; the period is retained as recorded.",
    "The live Federico article and the canonical reproduction script fetched at the recorded engine revision use a different setup and do not independently substantiate this archived table.",
    "No pre-2008 synthetic results included; no future performance or optimal allocation established."
  ],
  "source_record": {
    "repository_path": "content/ai-comparison/evidence/recorded-spy-research.md",
    "sha256": "0335539fca7b4ab0e653879b769618d2272d919e14d9423142c48eddd5f3aa32",
    "text": "# Recorded SPY crash-protection research\n\nSource record dated 9 September 2026.\nThe figures below come from the supplied research record and have not been rerun for this website.\nResults are historical simulations in USD.\n\n## Tested strategies\n\n| Configuration, 2008-2025 | Annual compound growth | Maximum drawdown | Realized annual premium spend |\n| --- | ---: | ---: | ---: |\n| SPY buy and hold | 11.11% | -51.9% | None |\n| Puts 30% out of the money, bought with 90-180 days remaining, sold at 30 days remaining, entries every two months, nominal 2% annual premium budget | 14.49% | -30.6% | 1.60% of average portfolio value |\n| Original 98/2 standing allocation, puts 30% out of the money, bought with 90-120 days remaining, sold at 60 days remaining | 7.98% | -46.2% | 11.37% of average portfolio value |\n\nThe transcript uses the second configuration.\nThe 2% figure describes an annual premium budget funded by the portfolio.\nIt does not describe a permanent 2% allocation to options.\nPremium costs and quoted bid-ask spreads are included.\nTaxes, commissions, additional execution slippage and capacity constraints are excluded.\n\n## Dependence on strategy and period\n\nChanging the exit rule from 30 to 60 days remaining changes the result substantially.\nFor the documented 30% out-of-the-money configuration, the post-2011 SPY comparison reduces the annual return advantage from 3.38 to 0.37 percentage points.\nThe post-2011 QQQ comparison reports an annual return disadvantage of 1.62 percentage points and no drawdown improvement.\nThese comparisons limit conclusions about a universal protection or return benefit.\n\nMuch of the deep-put strategy's benefit comes from selling option time value during volatility spikes, sometimes while the option remains out of the money.\nAn expiry-only payoff calculation cannot reproduce that behaviour.\nThe record identifies unhedged intervals after gains are realised, suppressed entries while existing options retain high market value, and smaller replacement positions when premiums rise.\n\n## Quote and valuation limits\n\nThe source reports purchases at ask and sales at bid.\nUndocumented upstream option-quote sourcing limits independent verification.\nWhen a held contract is missing from a day's option chain, the engine uses an intrinsic-value fallback.\nThe source reports this affecting roughly 10% of deep-put holding days, which can affect measured drawdowns.\nRaw-data hashes support reproducibility but do not verify the upstream quotes.\n\nThe source's pre-2008 paths use synthetic extrapolation.\nThey are separate from the observed 2008-2025 option-quote simulation used in the transcript.\nThe record does not establish a guaranteed loss floor, a future return premium or ten-year outperformance in EUR.\n\n## Source and verification record\n\nThe recorded engine is [`lambdaclass/options_portfolio_backtester`](https://github.com/lambdaclass/options_portfolio_backtester) at revision `e53ef86`.\nIts public repository publishes the canonical SPY raw-data hashes.\nThe [original research article](https://federicocarrone.com/series/leptokurtic/the-tail-hedge-debate-spitznagel-is-right/) supplies the reproduced comparison.\n\n| Research source | Recorded SHA-256 |\n| --- | --- |\n| Strategy driver | `75bc312f2e6c6c43606f5c2b30b507e8af2dade6e82997a6b0e72965588d4bbc` |\n| Exporter | `cc84ef910ddf2c9ac3c6f3818d9a9a9a5c2687c161040e48eebaeb434f5db1dc` |\n\nThe supplied record reports a successful parameter-description command.\nA subsequent verification attempt timed out after 120 seconds.\nEarlier claims of seven oracle tests and 1,332 unit tests remain source-reported receipts.\nThey were not rerun for this website.\n"
  },
  "external_source": "https://federicocarrone.com/series/leptokurtic/the-tail-hedge-debate-spitznagel-is-right/",
  "engine_reference_in_record": "https://github.com/lambdaclass/options_portfolio_backtester/tree/e53ef86928777de6ee0721424762ea3dc133f993",
  "source_version_note": "Retrieved 16 November 2026: the live article uses 40-45% OTM options and external funding; this archived table uses the distinct 30% OTM portfolio-funded setup stated in the September record."
}
