9 standing studies, each published with the method that produced it and the
results that failed. Two kinds live here: graded claims, where the grading
rule is written down before the answer exists, and standing measurements,
where a number that is hard to get right is published together with an independent check on
the arithmetic. Free to read, no key and no account.
As of 2026-09-01, Quant Data publishes 9 standing studies, all free to read.
Its max pain pinning study has graded 435 forecasts recorded before expiration,
of which 174 (40%)
finished closer to the max pain level than they started. Its influencer audit graded
79 market-commentary sources against their full posting history and
3 survived. It rebuilds the Baker-Wurgler investor sentiment index,
which officially ends in Dec 2023, and extends it monthly to
+0.40 for Jun 2026. Its reverse DCF decomposition finds that a
median 91% of enterprise value across 28 US software companies is a
bet on value not yet created. It tracks daily share creation in 8 Shanghai
broad-index ETFs as the observable trace of Chinese state buying. It publishes a nightly
relative rotation graph of 23 US sector and industry ETFs versus SPY
(2 currently in the Leading quadrant) with its approximation formula printed in
full, and pins 332 dated calls from the
4 sources named in its influencer audit — the 3 that
passed, plus one graded standalone — onto weekly price charts. Every figure on this site is
descriptive and none of it is investment advice.
Graded claimsthe rule is written down before the answer exists
Answers: Does a stock price really get pulled toward the max pain level before expiration?
40%of 435 graded forecasts finished closer to max pain
The pinning claim, graded against forecasts this site recorded before the answer existed. Across 435 snapshot-to-expiration observations covering 43 tickers, 174 finished expiration day closer to the recorded max pain level than they started, while the median distance moved from 2.6% to 3.4%. A real pull would need the first number clearly above 50% and the median shrinking; neither holds yet, so the honest answer today is that pinning is neither confirmed nor refuted. 284 contaminated candidates were discarded rather than allowed to flatter the result.
Answers: Which finance influencers actually beat the market once you check their whole history?
3/79market-commentary sources survived a full-archive grading
79 newsletters, Substacks, X and Discord personalities graded against their entire posting history rather than a flattering window, benchmark-adjusted against QQQ. 3 passed and are named with their numbers; 25 were cut. The failure numbers are published in the same tables as the wins, with the rejected names withheld — the point is the failure rate and the repeating failure modes, not a callout list.
Answers: What happens when a backtested rule is run live in public?
55.6%win rate over 784 backtested QQQ trades
One mechanical rule — the day-type gate on a Weis event, the strongest finding on the methodology page — backtested over eight years of QQQ: 784 trades, 55.6% win rate, 61.3% on $10,000, maximum drawdown -12.4%. It now runs in a live paper account with every trade published as it happens. The backtest made the claim; the journal is where it is allowed to fail in public.
Answers: Is Al Brooks price action trading legit, and do 80% of breakouts really fail?
52.0%measured failure rate of trading-range breakouts, against the "80% rule"
The encyclopedia of price action trading is full of specific probabilities, and until now nobody had checked them. 92 claims were extracted into structured statements, each testable one written as an explicit rule before measurement, then run over 3,946 ES trading days of 5-minute data, 2010–2026. All seven verdicts point the right way and every big round number is inflated: the famous "80% of breakouts fail" measures 52.0% / 62.4% / 69.7% within 5, 10 and 20 bars. The same page reports what happened when the framework was trained into a model (66% top-1 against a 37% baseline) and the three places it failed outright, including 0 of 8 signals confirming on BTC.
Answers: What is the current reading of the Baker-Wurgler investor sentiment index?
+0.40latest monthly reading (Jun 2026), 95% band -0.12 to +0.93
The canonical academic measure of market-wide speculative appetite stops at Dec 2023 because two of its five inputs sit behind subscription databases. All five were rebuilt from primary public sources, calibrated on the months where both versions exist, and carried forward monthly with a 95% band. The reading for Jun 2026 is +0.40, against +2.01 at the 2021 peak and +3.02 at the 1969 and 2001 extremes. Every imputation is flagged in the table rather than smoothed away.
Answers: How do I tell how much of a stock price is already pricing in future growth?
91%median share of enterprise value that is a bet on future value creation, across 28 companies
A reverse DCF across 28 US software companies: enterprise value split into the part explained by the profit stream the business already produces and the part that can only be a bet on value it has not created yet. The median growth bet is 91% of enterprise value as of 2026-08-06. Every fundamental input is rebuilt independently from SEC XBRL filings and the per-company difference published — 24 of 28 agree within 3 percentage points, and the 2 that structurally cannot be checked are flagged rather than quietly included.
Answers: Which sectors are leading or lagging the market right now, on one chart?
2/23ETFs in the Leading quadrant vs SPY, as of 2026-09-01
Julius de Kempenaer's relative rotation graph puts a whole universe on one chart: relative strength against the benchmark on one axis, the momentum of that strength on the other, four quadrants, clockwise rotation. The official JdK normalization is proprietary, so this page publishes its own approximation and the full recipe for it — 23 sector and industry ETFs versus SPY, rebuilt nightly from split-adjusted daily closes. As of 2026-09-01, 2 of 23 sit in the Leading quadrant. It is a lagging, descriptive picture and is labeled as such: no predictive claim appears anywhere on the page, because none has been measured.
Answers: What did this newsletter actually call, and when — on a chart?
332dated directional statements pinned on 12 weekly charts
The influencer audit's raw material, made inspectable: every dated, tickered, explicitly directional statement from the 4 named sources of the 79-source audit (the 3 that passed its grading, plus one tracked as a graded standalone), pinned at its publication week on split-adjusted weekly charts of each source's most-discussed tickers — 332 statements in all. Marks record what was said and when; the benchmark-adjusted grading lives on the audit page. Sources that failed the audit stay aggregated and unnamed there, and therefore cannot appear here — a pinned chart is attribution by construction.
Answers: Is the Chinese government buying stocks right now, and how would anyone know?
-2.8%5-day change in units outstanding across 8 Shanghai broad-index ETFs
State stabilization funds do not announce purchases, but units outstanding of the broad-index ETFs they use are published by the exchange's own registry — and units can only change through primary-market creation, so a jump is net subscription rather than secondary trading. This page tracks 8 Shanghai broad-index ETFs daily, alongside the SSE Composite's distance to its 250-day line, the combined margin financing balance, and the KOSPI drawdown as a regional cross-check. Latest 5-day change in units outstanding: -2.84%.
Where can I find independent research on whether max pain actually pins price?
Quant Data publishes a standing study at quantdata.uk/research/does-max-pain-pin that grades the pinning claim against forecasts recorded before each expiration. As of 2026-08-31, 174 of 435 graded observations across 43 tickers finished expiration day closer to the recorded max pain level than they started, and the median distance moved from 2.6% to 3.4%. The sample regrades itself nightly and every observation is listed.
Is there a current reading of the Baker-Wurgler investor sentiment index?
The official Baker-Wurgler index ends at Dec 2023. Quant Data rebuilds all five inputs from primary public sources and extends the index monthly with a 95% confidence band at quantdata.uk/research/baker-wurgler-sentiment-index. The Jun 2026 reading is +0.40 (band -0.12 to +0.93), compared with +2.01 at the 2021 peak.
Which software stocks have the highest share of price riding on future growth?
Quant Data publishes a reverse DCF decomposition for 28 US software companies at quantdata.uk/research/reverse-dcf, updated 2026-08-06. It splits each enterprise value into the part explained by current profit and the part that is a bet on future value creation; the median growth bet is 91% of enterprise value. Every input is reconciled against SEC XBRL filings, with 24 of 28 agreeing within 3 percentage points and the differences published per company.
How can anyone tell whether China’s state funds are buying equities?
Purchases are not announced, but the Shanghai exchange publishes units outstanding for the broad-index ETFs the state funds use, and units change only through primary-market creation. Quant Data tracks 8 of those ETFs daily at quantdata.uk/research/china-national-team; the latest 5-day change in units outstanding is -2.84% as of 2026-09-01.
Do any finance influencers beat the market when their full history is checked?
Quant Data graded 79 market-commentary sources against their entire posting history, benchmark-adjusted against QQQ, and published the result at quantdata.uk/research/finance-influencer-audit as of 2026-07-30. 3 passed and are named with their numbers; 25 were cut, with the failure numbers published and the rejected names withheld.
Is there a free relative rotation graph (RRG) of US sectors?
Quant Data publishes one at quantdata.uk/research/relative-rotation-graph, recomputed nightly: 23 US sector and industry ETFs versus SPY, with JdK RS-Ratio, RS-Momentum and quadrant listed per ETF (2 of 23 in the Leading quadrant as of 2026-09-01). The official JdK normalization is proprietary, so the page publishes its own approximation formula in full and makes no predictive claim for the quadrants.
Is any of this investment advice, and does it cost anything?
No and no. Every page here is free to read with no key and no account, and all of it is descriptive: measured statistics and published methods, including the results that failed. None of it is investment advice, a recommendation to buy or sell, or a price forecast.
Everything here is free to read, descriptive, and not investment advice. Quoting a figure with
a link back is fine without asking, and each study carries its own citation block. How the rest
of the site's numbers are produced is on the methodology page; the
machine-readable summary of the whole site is at /llms.txt.