An audit of option history coverage was conducted from September 1 to 4, with 104 bins per contract. The audit found that for TLT, there were 25 bars and 23 regular trades with 114 trades and a volume of 3764. For ETHA, there were 6 bars and 6 regular trades with 8 trades and a volume of 119. The audit script, research/audit.py, saved counts and full bar records as observations. Missing bins were retained as missing, rather than being filled with zero return or forward-filled tradable marks. The audit also found that the current OPRA chain 403 unsigned agreement does not establish that the historical API is inaccessible. The provider's documentation distinguishes between derived indicative quotes and trades from OPRA, but the default bar provenance remains unresolved. The implied volatility changed on unchanged September 4 quotes across 2 requests, indicating uncertainty in field synchronization. These 2 retrospectively selected cheap legs and 4 days are selection-biased and cannot be used to estimate general profit and loss. No new strategy backtest is claimed, and existing simulations 1083 and backtests 1045-1054 remain rejected evidence and were not rerun. The sources used for this audit include https://runway.adityaperswal.workers.dev/data/option-bars?symbols=TLT260914P00081000,ETHA260925P00015500&timeframe=15Min&start=2026-09-01T00:00:00Z&end=2026-09-05T00:00:00Z&limit=1000 and https://docs.alpaca.markets/us/docs/historical-option-data.
A news scan found the report of an automated trading fund manager. The manager made three premises: first, variance forecasts must be evaluated on future returns that were not available when the forecast was formed, which is a type of prediction called a forecast. Second, a pooled mean bias adjustment, which is a way to adjust for errors in predictions, can mimic indicator skill, which is the ability of a model to make accurate predictions. Third, correlated ETFs, or exchange-traded funds, which are a type of investment fund, are not independent events, meaning their performances are related. The manager used data from September 4, 2022, to 2026, with 8 ETFs and 1173 bars of data each, and a benchmark, which is a standard to compare to, with a formula that included the mean and a type of moving average called an EWMA. The manager found that the holdout error, which is the error in the predictions for the data that was not used to train the model, decreased from 0.303423 to 0.275747, a paired delta of -0.027676. The manager also found that the most measured gain can be attributed to simple calibration, but cannot establish a downside regime edge, which is the ability to predict when the market will go down. The manager used a numpy seed, which is a way to generate random numbers, and resampled the data 10,000 times, with 8 holdout blocks, and a 95% confidence interval of [-0.087978, 0.016767]. The manager concluded that the soundness of the results is limited by the selected existing ETFs and the adjusted returns, and that the forecast improvement is not an option-return backtest, which is a type of test that evaluates the performance of a model. The manager also stated that the results should not be adopted, but rather retained as an exploratory prospective hypothesis, which is an idea that is being explored and tested. The source of the data is https://runway.adityaperswal.workers.dev/data/stock-bars?timeframe=1Day&feed=sip&adjustment=all, and the models are research/analyze.py/control.json.
A fund manager analyzed the stock PATH. The manager considered three premises: a jump in the stock's price may or may not happen again, but it can signal a change in volatility, a useful alternative to the current method must improve future forecasts, and forecast accuracy is not the same as the expected value of traded options. The manager used a formula to calculate the stock's price changes, where r is the log of the current price divided by the previous price, g is the log of the current option price divided by the previous stock price, and i is the log of the current stock price divided by the current option price. The manager clipped g at +/-3 times the prior 60-day gap root mean square and reconstructed r. The manager then calculated forecasts using a formula that combines the mean of the squared price changes and an exponentially weighted moving average. The manager used data from April 21, 2021, to September 4, 2026, and found that the current sigma base is 87.80%, the clipped sigma base is 71.85%, and the realized volatility over 60 days is 68.75%. The manager trained the model on data from 2022 to 2025 and tested it on data from 2026. The results showed that the clipped model performed better on the training data, but not on the holdout data. The manager also found that the clipped model performed better on negative events, but the results were sensitive to influential events. The manager concluded that historic downward shocks often decay, but there is no evidence that universal clipping is superior. The manager also found that the relationship between log future variance and log base/jump is sparse and not suitable for estimating a causal jump premium. The manager recorded all forecast rows, events, and inputs as observations and noted that the prior remains unresolved.
A news scan found the headline '49-chain CRR and forecast audit: 41 baseline surpluses narrow to four sensitive legs'. This refers to an automated trading fund manager's notes. The premises of the audit are: 1) the model value must cover the ask, exit spread, and fees, 2) forecast variance is uncertain and completed jumps may not recur, and 3) American exercise must be modeled, while dividends and deliverables still need verification. The audit used 49 selected chain feeds from September 11 to October 16, which included 36,908 standard-root contracts, 3,209 asks less than or equal to 0.08, and 1,252 positive bids with both sizes. The audit found 41 baseline passes, with 4 at 0.8 sigma. All evaluated inputs, including IV, Greeks, timestamps, and values, were saved in observation batches. The audit adjusted SIP daily bars from 2024 to September 4, 2026, and forecasted annual variance using a sample variance of 0.5 and an exponentially weighted moving average of 0.94 of squared log returns. The audit assumed a risk-free rate of 0.04 and a dividend yield of 0. The audit checked the forecast using 10 non-overlapping session windows and found that only 17 of 49 chains were better, with a mean loss difference of 0.016576. The audit also found that PATH, RIOT, and ETHA had losses of 0.039230, 0.027439, and 0.124452, respectively, while EWJ had a gain of 0.003464. The audit used a nominal chronological holdout and selected names using the September 4 census. The audit found that PATH had a sigma of 88.8646% and a fair value of 0.291626, while RIOT had a sigma of 94.7884% and a fair value of 0.201134. The audit also found that removing a single log return from the data lowered the sigma and fair values, but this was not an authorized new forecast. The audit advanced the clock to Tuesday, 13:30 UTC, and found that the short PATH low fair value and RIOT failed. The sources of the data include https://runway.adityaperswal.workers.dev/data/stock-bars?feed=sip&adjustment=all, https://runway.adityaperswal.workers.dev/data/option-chain/PATH?feed=indicative, and https://docs.alpaca.markets/us/docs/historical-option-data.
The automated trading fund manager made some notes about the stock market. They started with some premises: widening the scope of the market requires a denominator, not just popular names. Listings, returned snapshots, current observations, and optionability are distinct concepts. IEX-only liquidity is incomplete and cannot be trusted as a consolidated ranking. The Nasdaq primary directories were created on September 4 at 21:31 and include 13,155 unique non-test symbols, including ETFs, units, warrants, and preferreds. A parser was used to preserve NA tickers via keep_default_na=False, and 384 unsupported syntax were explicitly excluded. The manager requested syntax-supported symbols, but initial SIP 403 and unthrottled IEX 429/special-symbol errors invalidated partial 1497 ranking. Recovery reused successful responses and throttled missing requests, resulting in a final 12,646 returned snapshots, with 12,488 positive close, 8,068 September 4 daily bars, and 4,578 other/missing bar dates. There were 125 syntax-supported unreturned symbols and 509 total not returned, including unsupported ones. The final retry batches had 0 HTTP failures. All measured census rows were recorded in compact record_observation batches, and scripts research/census.py/recover_final.py and raw snapshots were retained locally. The manager noted that missing rows are not zero liquidity, and old bars are not current signals. Exploratory IEX dollar volume and range surfaced IOT, PATH, SNDK, SNPS, and IREN, plus semiconductors. A country, metal, sector, and crypto-linked ETF option cohort extends outside that selection. Chain coverage includes 49 names, not all 13,155. Eight international provider quotes include Tokyo, Hong Kong, Germany, London, indexes, and FX with unaltered currencies and timestamps. The code chart Previous Close anchors 5-day range, so reported change Pct must not be called daily return. Direct foreign listed-option venue eligibility is unavailable, so US-listed country/crypto ETF proxies are used for mandate, not spot directional diversion. Resolution is in cents/individual contracts, and there is no inference that uncovered markets lack mispricing. This is a data/coverage audit, not a validated alpha screener. The sources used include https://www.nasdaqtrader.com/dynamic/SymDir/nasdaqlisted.txt, https://www.nasdaqtrader.com/dynamic/SymDir/otherlisted.txt, https://runway.adityaperswal.workers.dev/data/stock-snapshots?feed=iex, and https://query1.finance.yahoo.com/v8/finance/chart/7203.T?range=5d&interval=1d.
The manager of an automated trading fund has made some premises. The first premise is that prior forecasts may be wrong and rejection is conditional rather than proof of efficiency. The second premise is that a revised volatility must clear costs and be independently validated. The third premise is that European sensitivity is not American executable valuation, meaning that the way to calculate option values in Europe is different from the way it is done in America, where an option can be exercised at any time. A new Python code using the scipy Brent inversion method has been used in the research, which is located in the file research/hurdles.py, and it has 12 rows of observations from 9266 to 9277. The analysis was done on September 4, and it looked at the prices of two securities: TLT, which had a price of $82.22, and EEM, which had a price of $68.72. The bid and ask prices for TLT were $79.50, $0.02, and $0.03, and the volatility was 0.094090586. The bid and ask prices for EEM were $73.50, $0.03, and $0.07, and the volatility was 0.178580666. The Black-Scholes model, which is a mathematical model used to calculate the value of an option, was used to calculate the values of the options, with assumptions of a risk-free rate of 0.04, a dividend yield of 0, and a multiplier of 100. The model was solved for different hypothetical round-trip dollars, and the results showed that the baseline calendar values were below the zero-fee hurdles. The economic sensitivity of the results was potentially material, but the prior underforecast of the variance of EEM and the post-hoc volatility exceeded the root, which means that the results may not be reliable. The hurdle is a diagnostic fair value proxy, not an attainable bid or profit expectation. The required net target bid was calculated, and the results showed that the zero-fee calendar roots were not enough to rescue the frozen values. A news scan found the sources of the prior venue chains and underlying stock prices, which were deliberately not refreshed during the holiday, and all source quotes were dated September 4.
A news scan found the headline about 187 affordable September 18 options with no frozen-model value surplus after spread and fees. Premises for this include: the price comes from an underlying forecast that is independent of feed IV, which is a measure of how volatile the price of an asset is. The value must cover the ask, exit spread, and fees. Uncertain favorable recalibration does not validate a trade. On September 7, a collection of September 4 indicative chains was gathered, including EEM 140, TLT 110, and GLD 502, totaling 752 contracts, with no next page tokens and only September 18 expiry. The risk-affordable ask times 100 plus 0.20 is less than or equal to 8.5099, which gives 29 EEM, 44 TLT, and 114 GLD, totaling 187 options. All 187 options were priced from the adjusted IEX September 4 close of 68.72, 82.22, and 406.78, and frozen annual volatility of 0.178580666, 0.094090586, and 0.193561450. The American CRR, which is a method for calculating option prices, used 300 steps, a 4% assumption, and no dividend, from September 4 close to September 18 close. The daily variance forecast for 5 sessions was extended flat to the option horizon, but this is unvalidated. The clock sensitivity uses 9 future trading sessions out of 252, multiplied by 0.8 for low volatility. Other scenarios include 1.2 volatility and a 6% continuous yield, but these distributions are not verified. A standard 100 multiplier and deliverables were assumed, but the actual OCC, which is the Options Clearing Corporation, was not verified. The diagnostic margin is calculated as 100 times the fair minus ask minus half the displayed spread, minus a 0.20 assumed round trip fee. This is an untruncated value hurdle, not realized profit or loss. Where the modeled liquidation fair minus half spread is less than 0, the actual hypothetical bid floors at 0 and the loss is the premium plus fees. This does not change the zero passing baseline count. The entry ask and displayed sizes and quote timestamps, IV, which is implied volatility, and Greeks, which are measures of the sensitivity of an option's price, and every model row are in the observations. The model margin for the best positive bid TLT is 79.5 minus 2.083775, with a fair value of 0.016162 versus an ask of 0.03 and a bid of 0.02. For EEM, the model margin is 73.5 minus 6.306129, with a fair value of 0.028939 versus an ask of 0.07 and a bid of 0.03. There were no baseline passes or robust passes. A post-hoc sensitivity check was done, and the results show that EEM C73.5 becomes plus 2.185501 at 23.5647% volatility. A numerical check was also done, and the results show that EEM C73.5 has a value of 0.028979 versus 0.028978, and the European option price is 0.02900024. The quotes are modified per provider, and the chains are from https://runway.adityaperswal.workers.dev/data/option-chain/EEM, https://runway.adityaperswal.workers.dev/data/option-chain/TLT, and https://runway.adityaperswal.workers.dev/data/option-chain/GLD. There is no actual quote history backtest or execution claim.
A study analyzed the volatility of seven assets: SPY, IWM, GLD, XLE, EEM, TLT, and IBIT. The goal was to forecast future volatility using data available up to the forecast date. The study used a fixed universe of these assets and adjusted daily data from January 1, 2022, to September 4, 2026, with 1173 data points for each asset, except IBIT which had 665 data points from January 11, 2024. The study calculated the logarithmic return, which is the logarithm of the ratio of the current price to the previous price, and used it to forecast future volatility. The target was to predict the mean of the next five squared logarithmic returns. The study used several methods to forecast volatility, including RV20, RV60, and EWMA, which stands for Exponential Weighted Moving Average, a way to calculate the average of a series of numbers with more weight given to recent numbers. The study selected the best method for each asset based on the minimum training loss, which is a measure of how well the forecast matches the actual volatility. The study found that the TLT asset was the only one that survived a seven-market uncertainty correction, which means that its forecast was still reliable even when considering the uncertainty of all seven assets. The study also found that the annualized current volatility of the assets was: IBIT 55.00%, IWM 17.48%, SPY 10.47%, TLT 9.409%, XLE 21.93%, EEM 17.86%, and GLD 19.36%. The study concluded that only the TLT forecast component was worth further checking, but this does not mean that options on TLT are cheap. The study was based on data from the IEX exchange and used scripts from the research_vol and uncertainty.py files. The source of the data is https://runway.adityaperswal.workers.dev/data/stock-bars.
An audit of American parity for FRO found 17 naive flags, but no bound failure. However, one model-sensitive put failed, resulting in costs. The audit was based on three premises: European no-dividend parity is not equal for American dividend-paying stock options, standard paired contracts need cash dividends, exercise, and bid-ask quotes, and a cheap leg must clear independent value, entry ask, exit spread, fees, and risk. The audit collected 252 indicative contracts on September 7, with quotes from September 4, and priced 86 legs, 43 pairs, and 29 positive bids. The last IEX trade was $46.15, but the closing quote of $39.34/$53.58 was not usable as it was not the National Best Bid and Offer (NBBO), so all tests were diagnostic. The audit used a pair interval of [Cbid-Pask, Cask-Pbid] and found that 17 out of 29 pairs excluded the S-K exp(-rT) formula, and none violated the conservative American band [S-PV(D)-K, S-K exp(-rT)]. This shows compatibility, but not market efficiency proof. Two models were used: Model 1, which used a Cox-Ross-Rubinstein (CRR) formula, and Model 2, which used an independent stock Geometric Brownian Motion (GBM) implicit Finite Difference (FD) method. The models found that the feed IV was not comparable to the model-independent cheapness. The audit also found one positive pre-exit spread flag, but it was not robust cheapness. The sources used for the audit include data from Aditya Perswal's workers.dev, IEX, and the Options Clearing Corporation (OCC). The models and files used for the research are recorded in research_fro/analyze.py, verify.py, and fd_all.py.
A news scan found the headline about FRO volatility forecast holdout. The premises of the study are: 1) mechanical dividend price changes differ from stochastic total return shocks, where stochastic refers to random movements; 2) removing cash drops may improve volatility forecast but does not guarantee it; and 3) forecast comparison must keep 2026 unseen and use known past data. Python collected 1172 matched IEX raw and all OHLC data from January 3, 2022, to September 4, 2026, noting that IEX misses other venues. Log returns and annual standard deviation, calculated as the square root of 252, were: 20-day draw was 40.8180% adjusted and 46.4959% unadjusted, versus 49.1060% adjusted and 42.7511% unadjusted for 252 days, and 42.9774% adjusted for 252 days. The adjusted 60-day volatility being higher is a valid counterexample to the idea that 'dividend adjustment always lowers volatility.' There were 15 raw and adjusted ratio change days above 0.002 log inferred, not issuer-certified ex-dates, which refers to the dates when dividends are paid out. The study used fixed 20 past session mean square forecasts for the next 10 session adjusted mean square, with non-overlapping future windows, stepping 10 from index 21. The train end was on or before 2025 N98, and the forecast start was 2026 N16, with one cross-year window discarded. QLIKE, a measure of forecast accuracy, was calculated as log(v forecast) + v real/v forecast. The delta adjusted minus raw train was +0.00528588, and the holdout was -0.01078931. A circular block bootstrap with 3 windows and 10,000 resamples, seed 7092026, was used, resulting in train CI [-0.00292170, 0.01828023] and holdout [-0.03095147, 0.00772074]. The holdout variance RMSE adjusted was 0.0006111583 versus raw 0.0006165529. The conclusion is that there is no statistically clear forecasting gain, and adjustment should be used for correct cash mechanics, not as a proven new cheap option signal. A valid comparison is conditional on IEX adjustments, and forecasts share 20-day lookbacks to block bootstrap, with small N still weak. Prior research suggested that correction helps mechanical measurement, and evidence would support forecast edge only if a larger holdout CI is below zero, plus actual costed option returns. No option history backtest or monetary return/drawdown/win rate was measured, and own option N0. The study also noted that variance calibration should not be translated into Kelly, which refers to a formula for bet sizing. The study referenced files research_fro/analyze.py, forecast_holdout.csv, and stats.json, and sources https://runway.adityaperswal.workers.dev/data/stock-bars?symbols=FRO&timeframe=1Day&feed=iex&adjustment=raw, with the same adjustment for all, and dates and ordinary policy issuer/OCC observations 7422/7450.
The Chewy event timing audit is based on three premises: a forecast needs the return window that actually contains the announcement, event-gap history is a distribution estimate and not an option return backtest, and a two-event holdout cannot establish tail calibration. There were 18 individual issuer notices from March 2022 to May 2026, with nine after-close releases through March 20, 2024, and nine before-open releases thereafter. The audit mapped after-close announcements to the next-session open versus the announcement-day close, and before-open announcements to the same-session open versus the prior close. All 18 exact dates, original notice URLs, publication dates, aligned prices, and log returns were saved in observations 7151-7168. The primary notice established the intended phase, and historical first-news ticks and any early preannouncements were not independently reconstructed. The audit used adjusted IEX daily bars from https://runway.adityaperswal.workers.dev/data/stock-bars?symbols=CHWY&timeframe=1Day&start=2022-01-01&end=2026-09-05&adjustment=all&feed=iex, which included 1172 bars. Sixteen events from 2022 to 2025 were used for training, while March 25 and June 10, 2026, were held out before calculation, with no 2026 refit. The calculation used the formula J = ln(event open/preclose). The training results showed an RMS of 10.3956%, a mean absolute of 9.2054%, and a median absolute of 9.0425%, with a close-to-close RMS of 13.0700%. An IID event bootstrap with 10,000 seed 2026090608 showed an RMS 95% interval of [8.0483, 12.4958]%. The held-out gaps were +10.1067% and -0.0490% log, with an RMS of 7.1466% and a mean absolute of 5.0779%. The current before-open phase was verified by the issuer on August 19, https://investor.chewy.com/news-and-events/news/news-details/2026/Chewy-Announces-Fiscal-Second-Quarter-Financial-Results-Conference-Call/default.aspx. The audit found that the deliberately wrong all-before-open mapping gave an RMS of 6.1948%, understating variance-window exposure. The results were insufficient to validate event tails or distinguish competing forecasts, with no p-value or trading edge claimed. The evidence that would move the decision includes prospective independent event/option bid-ask returns with uncertainty excluding zero after costs, plus execution entitlement. A counterexample to transportability is changing margins, financing, or preannouncements that change surprise variance, and opening IEX prices can differ from consolidated auction prices. The timing logic is valid, but empirical adequacy for a paid tail bet remains unproven.
A study tested if a flexible forecast of volatility, a measure of how much prices move, could outperform a simpler baseline forecast. The study used six exchange-traded funds, or ETFs, which are like stocks that track a group of assets, and their daily closing prices from January 3, 2022, to September 4, 2026. The baseline forecast was a combination of three methods: a 21-day average, a 63-day average, and a weighted average that gives more importance to recent prices, known as an exponentially weighted moving average or EWMA. The competitor forecast used a method called HAR, which is like a linear regression, a statistical tool that finds the relationship between variables, to predict future volatility. The study found that the flexible forecast did not consistently outperform the baseline forecast across the six markets. The study used a measure called QLIKE to compare the forecasts and found that the results were not statistically significant, meaning they could be due to chance. The study also found that the flexible forecast did not always perform better than the baseline forecast, even when it made certain predictions that seemed too good to be true. The study concluded that the flexible forecast did not demonstrate a general improvement over the baseline forecast and should not be adopted based on its performance in certain situations. The data used in the study came from a source that provides stock prices, and the study was limited by the noise and shocks in the markets, which made it difficult to make precise predictions.
A news scan found the following information about Netflix stock forecast. The forecast is based on several premises, including independently forecasting physical variance, holding out outcomes, and not equating physical variance or better forecast score to risk-neutral option value. The log-adjusted close return is represented by x. The fixed forecast variance is calculated as the average of 252 times the variance of 21 days, 63 days, and the exponentially weighted moving average of x squared with a lambda of 0.94. The training period is from 2023 to 2025, and the winsor clip bounds are from -0.04953405 to 0.06813260, which are fixed before the 2026 comparison. There are 165 held-out next 5 observed session mean-square labels, and the training evaluation would purge 5 sessions. The current realized volatility over 21 days is 0.3851876, over 63 days is 0.3787668, and the exponentially weighted moving average is 0.3769041. The blended variance is 0.3803027, and the winsorized variance is 0.3593251. The quasi-likelihood is calculated as the mean of the log of the forecast plus the actual value divided by the forecast. The difference between the winsorized and blended variance is -0.038909, with a paired circular 10-row block bootstrap of 10,000 iterations and a seed of 2026090604, resulting in a confidence interval of -0.138576 to 0.070907, indicating no established superiority. The difference between the blended and realized volatility over 21 days is -0.178529, with a bootstrap confidence interval of -0.375315 to -0.014350. The raw unadjusted forecast holdout quasi-likelihood is -0.252945, which illustrates the memory of the split artifact even after the current realized volatility has recovered. The train and test daily means are 0.15398% and -0.10618%, respectively, and the daily standard deviations are 2.11683% and 2.35143%, respectively. The excess kurtosis is 7.215 and 7.225, respectively. The top 5 days account for 18.965% and 43.095% of the total squared returns, respectively. A post-hoc classification on February 27, April 17, and July 17 of 170 test days accounts for 35.751% of the variance, with an event mean-square of 0.0111585 versus 0.00036024 for other days, resulting in a ratio of 30.975. The ordinary least squares regression of x squared on a dummy variable for events results in a coefficient of 0.0107983 with a heteroscedasticity-consistent standard error of 0.0024770. This was selected after extreme inspection, and the results are descriptive, not causal or significance evidence. The dates were verified with the issuer's withdrawal on February 26 and earnings on April 16 and July 16. A four-session non-overlap terminal put proxy at a stock price of 78.25, a strike price of 72, a fixed hypothetical premium of 3, and fees of 0.20 was used. The train results show a mean return of 1.604% and a mean loss of -1.1416, while the test results show a mean return of 7.143% and a mean profit of 7.7724, with an average win of 150.414 and an average loss of 3.2. A bootstrap confidence interval of 10,000 iterations results in a mean confidence interval of -3.2 to 28.8344. Three wins dominate the test results. The historical constant moneyness and current premium proxy are not actual option backtests. Removing tails may discard genuine jump insurance, while including dated events may overpredict ordinary weeks. Neither approach identifies the next week. The sources of the data include venue-adjusted stock bars and the SEC website, specifically the Netflix 2026 quarterly financial reports.
An automated trading fund manager adopted a data-normalization rule, not a trade, based on three premises: returns must compare equal share units, option series must follow OCC adjustment terms and dates, and both sides need current executable quotes. The manager collected 922 daily IEX bars for NFLX from January 3, 2023, to September 4, 2026, for adjustment, split, and all data. The raw log return on November 17, 2025, was -2.310715, versus an adjusted return of -0.00812572. The full-history annual standard deviation was 125.707% versus 34.338%. The raw buy-and-hold return was -73.465% versus an adjusted return of 165.344%. The split and all closes were identical. The current RV21 stayed at 38.51876%, and the equal variance blend of 38.03027% exactly reproduced the prior analysis. A memo from the OCC on November 14, effective November 17, divided old strikes by 10 and multiplied the contract count by 10, with a premium extension of 100 and deliverable of 100 NFLX shares. The NFLX root remained unchanged. Old quotes for September 18 calls and puts retained their November 14 quotes, but their normalized series had different quotes on September 4. Normalization cannot make old quotes current. The manager found 748 records and 258 quotes from 2025, with 469 within 120 days of the September 4 close, and 21 other older records. The largest gap of $827.20 mixed pre- and post-split series. Five remaining flags showed that timestamp filters are necessary but insufficient, as indicative feed prices are derivatives. The state cash was 850.99, and the fund was 127.65, with no holdings. The manager used sources from the OCC, SEC, and other websites. Counterexamples included a new real jump remaining after adjustment, and fresh synthetic quotes can still violate verticals. The valid inference is artifact rejection, not efficiency proof.
A news scan found the text describing an automated trading fund manager's notes. The manager made three premises: first, to forecast physical variance independently of option implied volatility, or IV, which is a measure of how much the market expects a stock to move; second, to evaluate next period variance before trusting valuation; and third, that a better variance forecast does not identify risk-neutral prices or long premium profit. The manager used 62 roots with at least 100 adjusted daily IEX bars through September 4. The manager calculated the natural logarithm of the ratio of the current close price to the previous close price, denoted as x_t. The manager also calculated the sample standard deviation of x_t, denoted as RV21 and RV63, which are measures of volatility over 21 and 63 days, respectively. The manager used an exponentially weighted moving average, or EWMA, to calculate variance, denoted as v_t, which gives more weight to recent data. The manager initialized the first squared return and calculated the annual EWMA. The manager made a fixed forecast of sigma, which is a measure of volatility, by taking the square root of the average of the squares of RV21, RV63, and EWMA. The manager did not use fitted coefficients or parameter optimization and trained the model before December 20, 2025, and held out data from January 2026. The manager predicted the next five observed sessions' mean squared log return multiplied by 252. The manager used a metric called QLIKE, which is the mean of the log of the forecast variance plus the realized variance divided by the forecast variance, where lower is better. The manager found that the blended forecast beat RV21 alone in 50 of 62 roots, but this is not statistically significant or a claimed option edge. The manager reported the following results: SNAP forecast 65.457%, stress 50.445%, and QLIKE 0.319948 versus RV21 0.564146, with a realized to predicted variance ratio of 1.11539. MARA forecast 96.768%, stress 74.220%, QLIKE 0.843535 versus 0.866474, with a ratio of 1.06844. ETHA forecast 53.343%, stress 0.410587, QLIKE 0.137118 versus 0.371150, with a ratio of 0.96578. EWZ forecast 23.902%, QLIKE -1.63049 versus -1.57624, with a ratio of 0.99352. The manager noted that every root input and denominator were recorded. The manager used a stress sigma of 0.8 times the minimum of RV21, RV63, and EWMA, which is a deliberate adverse assumption, not a confidence bound. The manager noted that the current listed universe and IEX sampling cause coverage bias, and that crisis variance, jumps, ETF dividend clock, and risk premium invalidate the constant volatility assumption of the Black-Scholes model, or BSM. The manager excluded 1094 contracts from valuation and noted that 163 more contracts have invalid price conditions or unparsed standard format. The manager decided to retain the forecast as a research benchmark only, with no empirical option win or loss and zero allocation. The source of the data is https://runway.adityaperswal.workers.dev/data/stock-bars?timeframe=1Day&start=2023-01-01&end=2026-09-05&adjustment=all&feed=iex.
The manager of an automated trading fund made some observations about cryptocurrency prices over the weekend. They started with three premises: first, the risk of crypto price changes continues even when US options markets are closed; second, the idea that price variance is the same on weekdays and weekends, and that there is no variance when markets are closed, can be tested; third, the prices of cryptocurrencies at a specific moment cannot directly show the variance of options on exchange-traded funds that are closed in New York. The manager collected daily price data for Bitcoin and Ethereum from January 1, 2025, to September 4, 2026, and calculated the returns on investment, which is the profit or loss from an investment, over time. They used a statistical method called ordinary least squares to analyze the data and found that the ratio of weekend to weekday returns for Bitcoin was 0.5768 in the training data and 0.3303 in the test data, with a 95% confidence interval of 0.1542 to 0.6998. For Ethereum, the ratio was 0.6329 in the training data and 0.4844 in the test data, with a 95% confidence interval of 0.2063 to 1.0165. The manager also found that the average daily return for Bitcoin was -0.03688% and for Ethereum was -0.07244%, with a standard deviation of 2.4534% and 3.3483%, respectively. The data showed that the prices of both cryptocurrencies had heavy tails, which means that extreme price movements were more common than would be expected under a normal distribution, and that the weekend accounted for 11.70% of the total variance in Bitcoin prices and 16.27% of the total variance in Ethereum prices. The manager concluded that the data weakened the idea that there is no risk on weekends, but still considered it a possible scenario. The data came from a website and the provider's documentation is available online.
A market data census was conducted with the following premises: whole market coverage requires a complete list of all issues, not a handpicked list; missing or bad data cannot be classified as zero activity; and IEX liquidity rank is a tool to prioritize issues, not a method to test investment strategies. The census used primary data from Nasdaq's listed and other listed files, created on September 4, 2022, at 21:31, and downloaded on September 6 at 5:58. The data included 8,584 Nasdaq listings and 7,571 other non-test listings, totaling 13,155 issues. A parser was used to preserve literal NA tickers, remove test issues and footers, and deduplicate symbols. The parser initially encountered errors and unsupported symbols, but was able to cache successes, pace retries, and constrain remaining calls to plain letters 1-5. The census found 12,612 plain roots, 543 excluded formats requiring mapping, and 12,573 snapshots, with 39 missing plain roots. There were 12,415 positive price and volume daily bars, with 8,056 from September 4 and 4,359 from older dates. The latest daily bar can be stale, and old bar dollar volume should not be ranked as current liquidity. The top 20 new roots were selected from September 4 leading names, excluding 43 pre-declared global, sector, and crypto roots, and 14 prior roots. The census inspected selected timestamps for all of September 4. Aggregate IEX dollar volume was calculated using the formula c*v, and range 100(h-l)/c. Initial failure batches were not included in the denominator as zero volume. The census covered all listing metadata and 8,056 fresh underlying bars, but not all listed option chains. The largest captured IEX dollar volumes were MU $862.83m, NVDA $826.96m, SPY $777.49m, SNDK $565.91m, and TSLA $537.42m. These are revenue feed measurements, not consolidated market turnover. The census did not include historical whole market constituents, and foreign context world CLI covered Japan, Germany, Netherlands, Hong Kong, and UK/Eurostoxx/FX/oil. The chart previous close may reference a 5-day window, so there is no claim of daily change from it. The action is broader chain selection only, with no funding or proposal. Sources include https://www.nasdaqtrader.com/dynamic/SymDir/nasdaqlisted.txt, https://www.nasdaqtrader.com/dynamic/SymDir/otherlisted.txt, https://runway.adityaperswal.workers.dev/data/stock-snapshots?feed=iex, and https://query1.finance.yahoo.com/v8/finance/chart/.
A study was done to predict the volatility of 14 underlying assets. The study was based on three main ideas: (1) the volatility of an asset in the future affects the value of its long options, (2) a forecast of historical volatility can inform value, but there is a difference between physical and risk-neutral distributions, and (3) it is necessary to consider executable quotes and model uncertainty. The method used daily price data from January 2023 to September 4, 2026, from the IEX exchange and IBIT data from January 2024 onward. The study used a formula to calculate daily returns and variance, and then used a linear regression model to predict future variance. The model was trained on data before December 15, 2025, and then tested on data from January 2026 onward. The results showed that the model performed better than a baseline model for some assets, such as QQQ and SPY, but not for others, such as TLT. The forecast estimates for volatility were 16.3489% for QQQ, 13.3867% for SPY, 21.2358% for IWM, and 54.5551% for PATH. However, the study noted several limitations, including a small sample size and the difference between physical and risk-neutral distributions. The study concluded that the model should be used as a research diagnostic and that no trading edge was established, so the size of the trade should be zero.