Skip to content

DATA AUDIT — MIDA

Overall integrity grade: C (usable with caveats)

Confidence the data is REAL (not hallucinated): Medium-High

Prices and SET profile/financial/news shell are real artifacts. News feed and social are thin to empty. Two SET pages are 404s. Package dates are future-stamped (2026) — flagged below as a likely systemic clock/test issue, not a fabrication signal.

Findings by source

Prices — CAVEAT

  • 60-row recent sample is internally consistent: OHLC ordering holds on every row checked (low ≤ open,close ≤ high), no negatives, no zeros except where Volume=0 (5 zero-volume sessions: 2026-05-01, 05-04, 05-26, 06-01, 06-03 — on these, O=H=L=C, consistent with no-trade carry-forward; plausible for a thinly traded sub-30-satang stock).
  • Date sequence is monotonic; weekend gaps respected. Note holiday-sized gaps: 2026-04-10 → 04-16 (Songkran week, plausible) and 2026-04-03 → 04-07. Both fit Thai calendar.
  • Float-precision noise (0.25999999…, 0.30000001…) is normal yfinance float32 artifact, not corruption.
  • Date stamping is in the future: latest bar 2026-06-22, package built 2026-06-23. This is internally consistent within the package but is implausible vs real-world wall clock. Treat the "as-of" as whatever the system clock says; do NOT cite "current price 0.41" as a live quote.
  • Price action: ranges 0.22 → 0.42 over the 60 sessions; spike from 0.30 → 0.36 on 2026-06-04 (vol 27.6M, ~30× recent avg) and surge to 0.41 on 2026-06-22 (vol 16.9M). Real-looking event-driven moves, not synthetic.
  • 10y summary row is suspect / partially nonsensical: 2016-05-26,2026-06-22,2450,0.95,0.41,-56.84%,2.64,48.07,0.0549,-82.52%,1.59,13.16
  • 10-yr total return −56.8% with start 0.95 → end 0.41 ≈ −57% ✓ consistent.
  • Max drawdown −82.5% plausible for a 0.95 → low-teens-satang stock.
  • Annualised return field shows +2.64 (=264%?). That is mathematically inconsistent with a −57% cumulative over 10y; either the column is mis-labeled (could be Sharpe, skew, or a ratio), or the value is wrong. Downstream must not quote this number as "CAGR".
  • Annualised vol 48% and Sharpe 0.0549 are plausible for a microcap.
  • Trailing columns 1.59 / 13.16 unlabeled — do not interpret.
  • Symbol field "MIDA" at end of summary row matches ticker. ✓

News — FAIL

  • Yahoo: 0 items.
  • Google company-name search: 0 items.
  • Google ticker search: 4 items, but 3 of 4 are about MIDA = Malaysian Investment Development Authority, not Mida Assets PCL (Thailand). Items #2, #3, #4 are Malaysia industrial/manufacturing stories — ticker collision, not coverage. Only item #1 ("MIDA - Company profile - SET", Jan-2023) actually references the Thai company, and it's just a link to the SET profile, not news.
  • Effective real news count on Mida Assets PCL: 0.
  • SET news HTML loaded (413KB, valid <title>MIDA - News - SET</title>) but the excerpt shows only the Nuxt shell/scripts — the actual filing list lives in client-rendered JS and is not in the excerpt. Cannot extract individual SET announcements from this package.

Social — FAIL

  • google_blogs: 0.
  • google_forums: 1 item, and it is a Reddit thread about video-game "mid game / end game years" — completely unrelated to MIDA the stock. Junk match.
  • reddit_company: HTTP 403.
  • reddit_ticker: HTTP 403.
  • X/Twitter: explicitly absent (acknowledged gap).
  • Effective social signal: zero. Do not infer sentiment.

SET pages — CAVEAT

  • profile (406KB): valid title MIDA - Company profile - SET. Real shell loaded. Body content is in client-rendered Nuxt JS; excerpt shown is only <head> + script preloads. The textual company description, list of subsidiaries, business segments — not visible in the excerpt provided.
  • financial (428KB): valid title MIDA - Company highlight - SET. Same situation — shell only, no parsed financial table visible in excerpt.
  • shareholders (391KB): <title>ขออภัย ไม่พบข้อมูลที่คุณต้องการ</title> = "Sorry, the information you requested was not found". og:url =/en/error/404`. This is a 404 error page. Shareholder data NOT obtained.
  • filings (391KB): also a 404 error page — same Thai error title, same /en/error/404 URL, identical byte size to shareholders page. Filings data NOT obtained.
  • news (413KB): valid title; shell loaded; tabular news content not visible in excerpt.

Net: 3 of 5 SET pages loaded the correct shell (but content is JS-rendered, so no facts extractable from the byte excerpt); 2 of 5 (shareholders, filings) are confirmed 404s.

Contradictions found

  • News-ticker collision: Google News results describe "MIDA = Malaysian Investment Development Authority" (Malaysia gov body) while the data package is for Mida Assets PCL (Thai property/hotel/gold-loan). 3 of 4 news items are about the wrong entity.
  • 10y summary annualised-return field (2.64) vs cumulative return (−56.8%): cannot both be true for the same series under any standard convention. One column is mis-labeled or wrong.
  • Package timestamp 2026-06-23 vs real-world: not a within-package contradiction, but flagged.

Missing / stale data

  • SET shareholders page: 404 — no major shareholder list, no free-float, no insider holdings.
  • SET filings page: 404 — no 56-1 One Report references, no recent SET announcements list, no opportunity-day docs.
  • SET profile / financial / news page bodies: pages loaded but content is client-side rendered; the excerpts contain no extracted business description, no income statement, no balance sheet, no P/E, no dividend, no auditor name.
  • No income statement, balance sheet, cash flow anywhere in package.
  • No dividend history.
  • No insider / connected-transaction list.
  • No analyst coverage / consensus estimates.
  • No shares-outstanding or market-cap figure.
  • No free-float %.
  • No gold-leasing-segment financials (despite config calling out gold-loan as a business line).
  • No hotel-occupancy / RevPAR data for the Mida hotel chain.
  • No real news on the Thai company in the last 12 months (Yahoo=0, Google company=0, Google ticker=collision).
  • No social sentiment data (Reddit blocked, X absent, blogs 0).

DO-NOT-FABRICATE list for downstream personas

Do not invent or compute any of the following — they are NOT in this package: - Any income-statement line (revenue, gross margin, EBIT, net income, EPS). - Any balance-sheet line (assets, debt, cash, equity, book value per share). - Any cash-flow line (CFO, capex, FCF). - Any ratio that requires the above: P/E, P/B, EV/EBITDA, ROE, ROA, debt/equity, current ratio, Altman Z, Beneish M, Piotroski F — all would be fabricated. - Dividend yield, dividend history, payout ratio. - Shares outstanding, market cap, free float. - Major-shareholder names or %, insider transactions, board composition. - Auditor identity, going-concern opinions, qualified-audit flags. - Segment revenue split (property vs hotel vs gold-loan). - Hotel KPIs (occupancy, ADR, RevPAR, room count). - Land bank, backlog, presales, transfer pipeline. - Any "analyst target price", consensus, or rating. - Any sentiment claim from X/Reddit/Pantip/blockdit ("retail is bullish/bearish" — UNSUPPORTED). - The "annualised return = 264%" or any CAGR from the summary row — the field is internally inconsistent. - Any specific SET filing date or headline — none extractable. - Live/current price (package is future-dated 2026; the 0.41 close is from the dataset's last bar, not a real-time quote).

One-line instruction to the CIO

Trust the 60-day price series and the spike pattern as real; treat the 10y summary's annualised-return cell as broken; assume zero verified fundamental data, zero real news, and zero social signal — any thesis here must be flagged as opinion built on price-action plus the config blurb only.