Option Fair Value Performance Report Summary: 2/01/2025

VecViz’s analytic performance reports document the behavior and performance of our analytics. This post focuses on the “VecViz Option Fair Value (OFV) Performance Report” as of February 1, 2025, Here we summarize the key findings, and explain how to use the report to investigate the performance of specific tickers. You can find this report on the “Reports” page of the VecViz website, along with updated reports over time. Vector Model Option fair value estimates as of the prior day’s close can be found on the “Dashboards” page of the VecViz website.

The OFV report should be of interest to active participants in the options market, including those involved in convertible bonds. It may also be of interest to participants in the credit markets, given the conceptual linkages between option pricing and default risk.

Performance Metrics Utilized in the OFV Report

The OFV Performance Report compares the behavior and performance of OFV;s calculated using VecViz’s Vector Model to the same as calculated with VecViz’s implementation of the Black Scholes option pricing model utilizing the Sigma Model for the volatility assumption.

Fair values (FV’s) for out of the money put and call options are calculated for both models and presented in standardized, “percent of model date price” terms. They are calculated for a standardized set of strikes that are also defined in “percent of model date price” terms, and for our standard 6 forward time horizons.

Assuming no arbitrage and that all transactions occur at fair value with zero transaction costs, selling calls and puts on the same tickers in equal quantities, with strikes equidistant from (and near) the model date price should generate an average combined profit at expiration (over a long period of time and across many tickers) that approximates the risk free rate (RFR)¹. Thus, the primary metric utilized to compare the two sets of OFV’s is the proximity of sell side average profit at expiration, across calls and puts, to the RFR, assuming all sales occur at model fair value with no transaction costs.

Given the same level of average P&L proximity to the RFR, the model that suffers less severe losses is preferable. Therefore, severity of average losses by model date across tickers is another metric we evaluate these models upon, considering it in conjunction with the proximity metric². Finally, consistency in loss severity across tickers is another criteria we consider, though shortfalls here can be largely overcome with sufficient diversification.

Alongside these profitability metrics we also provide some detail on average fair values to facilitate comparison with the blogs summarizing Expected Body, OaR and VaR.

OFV Summary Results

We summarize our comparison of Vector Model OFV to Sigma Black Scholes OFV in the table below, for each of the six time horizons (denominated in trading days, where 21d ~= 1 month, 252d ~= 1 year).  We provide values only for the extremes of the time horizon spectrum, all stated in terms of % of model date ticker price. Where the Vector Model better fits the criteria we enter a “V”, and where Sigma Black Scholes is superior, we enter an “S”.  For sake of brevity, we present here results only for the entire out of sample performance record (“ALL TMD” = all ticker model dates), but the complete report also specifically addresses the trailing 1 year, 3 month, and 1 month time lookback windows.

OFV Estimate BEHAVIOR, All TMD (1/31/2022 – 1/31/2025)	1d	10d	21d	63d	126d	252d	1d V / S FV	252d V /S FV	page(s)
Larger NTM3 Call & Put Fair Value	V	V	V	V	V	V	0.83 / 0.58	16.20 / 14.94	11-16
Larger DOOTM Call & Put Fair Value	V	V	V	V	V	V	0.11 / 0.02	10.06 / 8.16	11-16
Larger NTM Call Fair Value	V	V	V	V	V	V	0.90 / 0.59	22.52 / 16.92	11-16
Larger NTM Put Fair Value	V	V	S	S	S	S	0.76 / 0.57	9.89 / 12.95	11-16
Larger DOOTM4 Call Fair Value	V	V	V	V	V	V	0.15 / 0.02	15.60 / 10.54	11-16
Larger DOOTM Put Fair Value	V	V	V	V	S	S	0.07 / 0.01	4.53 / 5.79	11-16

OFV Estimate PERFORMANCE, ALL TMD (1/31/2022 – 1/31/2025)	1d	10d	21d	63d	126d	252d	1d V / S	252d V /S	page(s)
NTM Call & Put Sale Avg P&L RFR Proximity	S	S	S	V	V	V	0.40 / 0.15	0.19 / -1.08	42-47
DOOTM Call & Put Sale Avg P&L RFR Proximity	S	S	S	S	V	V	0.11 / 0.01	1.09 / -0.81	42-47
Less Severe Losses on NTM Call & Put Sales By Date (on Average Across Tickers)	V	V	S	V	V	V	-1.65 / -1.78	-7.57 / -9.47	60-75
Less Severe Losses on DOTM Call & Put Sales By Date (on Average Across Tickers)	V	V	V	V	V	V	-0.18/ -0.31	-5.85 / -8.35	60-75
Less Severe Losses on NTM Call & Put Sales By Ticker (on Average Across Dates)	S	S	S	S	S	S	-0.99 (GME) / -0.33(SIVBQ)	-100.87 (MSTR) / 83.83 (MSTR)	112-147
Less Severe Losses on DOOTM Call & Put Sales By Ticker (on Average Across Dates)	V	S	S	S	S	S	-0.20 (SIVBQ) /-0.22 (SIVBQ)	-98.4 (MSTR) / -82.9 (MSTR)	112-147

Summary Conclusions:

Before we state the Conclusions, we just want to state that (1) obviously options are very risky, (2) see our Terms and Conditions for important disclosures, and (3) just because the Vector Model outperforms Sigma in many respects in this study, it doesn’t mean it will generate profits when applied to live option market data.

All that said, over the period studied, which includes both the bear market of 2022 and the bull market that followed up through most of January 2025, Vector Model Option Fair Values:

• exceeded Sigma Black Scholes fair values for calls generally, and for shorter horizon puts. Vector Model option fair values tended to be smaller than SBS fair values for longer horizon puts.
• had average put & call sale P&L that more closely approximated the RFR than Sigma Black Scholes (S) for longer horizons, but less closely approximated the RFR for shorter horizons.
  • note that the differential in average profitability between Vector Model (V) and Sigma was consistent with the differential in average FV’s. For example, 1d NTM profit differential of 0.25 (0.40 for V vs. 0.25 for S) equaled the 1d NTM FV’ differential of 0.25 (0.83 for V vs 0.58 for S). Likewise, the 252d NTM profit differential of 1.27 (0.19 for V vs -1.08 for S) nearly equaled the 252d NTM FV differential of 1.26 (16.20 for V vs 14.94 for S).
• incurred less severe losses by date (on average across tickers), for both NTM and DOOTM options, across nearly all horizons.
  • note that the differential in max average loss by date (across tickers) exceeded the differential in average FV’s for longer time horizons, suggesting a timing advantage by the Vector Model for such horizons. This is fairly consistent with ROVBC / ROOBC / ROEUB/ ROEDB alpha being positive on an average “by ticker across model dates basis”, particularly for longer horizons.
    • For example, the 252d NTM loss differential of 2.10 (-7.57 for V vs. -9.67 for S) exceeds the difference in corresponding average FV of 1.26.
    • That said, the 1d NTM loss differential of 0.13 (0.13%= -1.65 for the V vs. -1.78 for S) did not even match the differential in corresponding average FV of 0.25.
    • Looking at these same dynamics for DOOTM strikes, we see that the max loss differential exceeded the differential in FV for both 1d and 252d time horizons.
• incurred more severe losses by ticker (on average across dates), for both NTM and DOOTM options, across nearly all horizons. The tickers for which losses were most severe relative to Sigma were concentrated in the “Failed Banks” (SIVBQ, SBNY, FRCB) and the “Crypto/ Meme” stock groups (MSTR, GBTC, GME, AMC). This perhaps suggests, a Vector model “blind-spot” with regard to social media activity.
  • the differential between Vector Model and Sigma max losses by ticker exceeded the differential in average FV’s. This is consistent with ROVBC / ROOBC / ROEUB/ ROEDB alpha being negative on an average “by date across ticker basis”.

Footnotes:

1. both models utilize the same RFR assumption for all periods: 4.00%. ↩︎
2. loss severity should be considered in conjunction with RFR proximity because a fair value model that generated systematically inflated fair values of sufficient magnitude would never generate any losses. ↩︎
3. “NTM” is an abbreviation for “near the money” strike prices. Specifically, we refer to the bars for the +1% out of the money strike for calls, and -1% out of the money strike for puts. ↩︎
4. “DOOTM” is an abbreviation for “deep out of the money” strike prices. Specifically, we refer to the bars for +20% out of the money calls and -20% out of the money puts, though the report presents results for strikes as much as +/-50% out of the money. ↩︎