# SearchGrade data study — stratified band: Tranco 10,001-100,000
# 100 domains. Generated 2026-08-01.
#
# WHY BANDS EXIST
#   A top-1k census yields one contestable number per check. Bands yield a CURVE:
#   "llms.txt adoption falls from X% in the top 1k to Y% at 100k-1M" is a defensible finding
#   about the web; "X% of sites have llms.txt", measured only on the head, is not — and it
#   invites the fair criticism that we audited the best-resourced sites on earth and called it
#   the state of the web.
#
# SAMPLING — reproducible, and random rather than contiguous
#   Source     Tranco list 5648N   https://tranco-list.eu/list/5648N  (2026-07-31)
#   Frame      ranks 10,001-100,000, minus any domain already in domains-500.txt
#   Method     Fisher-Yates shuffle with mulberry32, SEED 20260801, then take the first
#              300 draws and keep the first 100 that classify as websites.
#              Random, NOT contiguous: ranks 10,000-10,300 are the top edge of the 10k-100k
#              band, not a sample of it, so a contiguous slice would just be another head
#              sample and would defeat the entire purpose of stratifying.
#   Classify   same single rule as band 1 — a domain is a website only if its APEX serves
#              human-readable content a person could plausibly be seeking. Judge the apex, not
#              the company.
#   Of 300 drawn: infrastructure 34 · website 200 · adult 11 · parked_unknown 52 · mirror 4 · adtech 3
#
# ⚠ Obscurity is not a disqualifier. A German machine-shop site or a Hungarian blog is exactly
#   what the long tail IS, and excluding them would rebuild the head-bias the bands exist to
#   measure. Only genuinely parked / for-sale / link-farm domains were dropped as unknown.
#
# One domain per line. Blank lines and # comments are ignored by scripts/run-study.js.

eu-startups.com
webtar.hu
zam.com
normattiva.it
alertnet.org
airnavradar.com
chargeafter.com
lwn.net
konest.com
realprize.com
richpanel.com
egyptair.com
helloprint.com
obatsuster123.com
ecologi.com
kela.fi
translink.com.au
tranquil.it
datafyhq.com
etf.com
tile-seasons.com
nachdenkseiten.de
dotdash.com
fnal.gov
businesstoday.com.my
birdeye.com
nectarsleep.com
sarkor.uz
nga.cn
umetrip.com
flannels.com
vivaticket.com
bmwfans.info
km.ru
ihadesigns.in
591.com.tw
sunjournal.com
nhqv.com
indiehackers.com
techloq.com
ifs.se
samanage.com
ta3.com
unilim.fr
imgupscaler.ai
city.setagaya.tokyo.jp
mootoon.co.kr
wco.tv
mybharat.gov.in
standsapp.org
feb.es
prenatal.com
foyles.co.uk
jobinja.ir
radios.com.pe
nationalbroadband.pk
chutianyun.gov.cn
technave.com
payworks.ca
photorecoverswift.com
setrahost.com
tencentmusic.com
spreadshirt.de
highradius.com
blackhawknetwork.com
just-eat.es
chatroulette.com
infop.hn
goldsupplier.com
topper.com.ar
nextlevlplay.in
yourok.ru
tensor.ru
catofashions.com
idealo.at
liberapay.com
annsummers.com
backupops.com
mui.com
innoscale.net
shafaq.com
altabank.ru
torun.pl
carnegiehall.org
esciclismo.com
shuminoengei.jp
lootlabs.gg
once.es
looperman.com
mgsu.ru
e-hoi.de
tu-darmstadt.de
mebmarket.com
tamashaweb.com
eb.mil
renaissancecreation.com
dynam-data.jp
aasm.org
learningapps.org
prozeny.cz
