Time Data Mining (TD)
NAS100 — Green / Red Time Data Mining (TD)
Seasonality and intraday bias of the NASDAQ-100 (US100). Green = candle close > open.
Last updated 25 Sept 2026
Source: Yahoo ^NDX (monthly/weekly/daily, 42y, auto-refreshed daily) · FxPro #USNDAQ100 + Pepperstone NAS100 (server time)
Free preview
🔒 You're viewing a free preview — key numbers and days are locked. Unlock the full research from $7.
Unlock full accessWeekly Plan — Sep 28 - Oct 2 2026
Day-by-day signals from 41 years of NASDAQ-100 data (1986-2026): month-by-weekday green odds, first/last-trading-day-of-month effects, and US market holidays. Auto-refreshed every Friday 22:00 (GMT+7).
Mon Sep 28
WAIT50%
Sep Mondays: 50% green
Tue Sep 29
WAIT49.1%
Sep Tuesdays: 49.1% green
Wed Sep 30
🔒 LOCKED🔒
Thu Oct 1
🔒 LOCKED🔒
Fri Oct 2
🔒 LOCKED🔒
💡 Unlock the full pass to read the complete weekly game plan.
Sep by weekday — green candle odds (41 yrs)
Highlights
The standout green and red edges across the study.
Greenest month
Jul 82%
▲ 18/22 years green (22y)
Best weekday
🔒
Members only
Strongest hour
🔒
Members only
Weakest month
🔒
Members only
Worst avg return
🔒
Members only
Method & data
Green candle = close > open. Daily seasonality (sections 1, 2, 6) — Yahoo ^NDX, 22 years (Jan 2005 – Aug 2026, 5,468 trading days). Intrad… 🔒 Unlock to read the rest.
The TD framework
Time Data Mining (TD) dimensions, as used across this page: Big TD - the higher-timeframe clock: week-of-month, weekday, the pre-US 4h bloc… 🔒 Unlock to read the rest.
1) Monthly bias — 22 years
Green = monthly close > monthly open.
| Month | Green % | Red % | Green yrs | Avg return |
|---|---|---|---|---|
| Jan | 59% | 41% | 13 / 22 | +0.31% |
| Feb | 36% | 64% | 8 / 22 | +0.11% |
| Mar | 🔒 | 🔒 | 🔒 | 🔒 |
| Apr | 🔒 | 🔒 | 🔒 | 🔒 |
| May | 🔒 | 🔒 | 🔒 | 🔒 |
| Jun | 🔒 | 🔒 | 🔒 | 🔒 |
| Jul | 🔒 | 🔒 | 🔒 | 🔒 |
| Aug | 🔒 | 🔒 | 🔒 | 🔒 |
| Sep | 🔒 | 🔒 | 🔒 | 🔒 |
| Oct | 🔒 | 🔒 | 🔒 | 🔒 |
| Nov | 🔒 | 🔒 | 🔒 | 🔒 |
| Dec | 🔒 | 🔒 | 🔒 | 🔒 |
TOP 3 green: Jul 82% · Nov 71% · Apr 68%. TOP 3 red: Feb 64% · Jun 55% · Dec 43%. By average return: best Jul +3.08%/mo, worst Sep -0.15%/mo. Negative average return: Sep -0.15% · Dec -0.07%.
1b) Monthly green % across lookbacks
Same monthly stat computed on five windows of Yahoo ^NDX history (to Aug 2026). Green = monthly close > open.
| Month | 4y | 10y | 20y | 26y | 41y |
|---|---|---|---|---|---|
| Jan | 100% | 80% | 60% | 58% | 68% |
| Feb | 25% | 30% | 40% | 35% | 46% |
| Mar | 🔒 | 🔒 | 🔒 | 🔒 | 🔒 |
| Apr | 🔒 | 🔒 | 🔒 | 🔒 | 🔒 |
| May | 🔒 | 🔒 | 🔒 | 🔒 | 🔒 |
| Jun | 🔒 | 🔒 | 🔒 | 🔒 | 🔒 |
| Jul | 🔒 | 🔒 | 🔒 | 🔒 | 🔒 |
| Aug | 🔒 | 🔒 | 🔒 | 🔒 | 🔒 |
| Sep | 🔒 | 🔒 | 🔒 | 🔒 | 🔒 |
| Oct | 🔒 | 🔒 | 🔒 | 🔒 | 🔒 |
| Nov | 🔒 | 🔒 | 🔒 | 🔒 | 🔒 |
| Dec | 🔒 | 🔒 | 🔒 | 🔒 | 🔒 |
Robust across every window: February is the reddest month in ALL five lookbacks. The 4y column has tiny samples (n=3-4) - treat as recent trend only. The 22-year table above uses a 2005-start window, hence small differences.
2) Weekday bias — 22 years
Daily candle, Monday–Friday.
| Day | Green % | Red % | n days |
|---|---|---|---|
| Monday | 56.2% | 43.8% | 1,024 |
| Tuesday | 54.6% | 45.4% | 1,123 |
| Wednesday | 🔒 | 🔒 | 🔒 |
| Thursday | 🔒 | 🔒 | 🔒 |
| Friday | 🔒 | 🔒 | 🔒 |
Greenest: Mon → Tue → Thu. Reddest: Fri → Wed → Thu. Every weekday leans green overall.
2b) Weekday bias - window matters
The same weekday stat on the recent 22-year window vs the full 42-year history. Both verified from the same raw feed.
| Day | 22y green % | 42y green % |
|---|---|---|
| Monday | 56.2% | 53.5% |
| Tuesday | 54.6% | 53.1% |
| Wednesday | 🔒 | 🔒 |
| Thursday | 🔒 | 🔒 |
| Friday | 🔒 | 🔒 |
Monday leads the modern era (2005+); Wednesday leads the full history (Wednesday was very strong pre-2005). In April + July specifically, Wednesday is the clear winner at 60.1% green. Friday is the weakest day in both windows.
3) Session blocks — server time, 4.2 years
Block green = close of last M15 in block > open of first.
| Block (server) | Green % | Red % | n days |
|---|---|---|---|
| 17:00 – 21:00 | 54.9% | 45.1% | 1,092 |
| 21:00 – 24:00 | 51.1% | 48.9% | 1,051 |
The US-session block 17–21 is clearly greener; the late block 21–24 is close to a coin flip.
4) Hourly green % — 17:00–21:00 (server)
Hourly candle, ranked by green probability.
5) M15 green % — 17:00–21:00 (server)
Green % per 15-min candle, in order. Note the 17:45 dip vs the 18:00 / 18:30 strength.
The short version
Best hour
18:00
▲ 55.7% green — greenest in the 12.5y H1 data
Worst candle
🔒
Members only
Late cycle 21–24
🔒
Members only
Streaks
🔒
Members only
5b) H1 candles — 17:00–24:00 cycles (12.5 years, server time)
Hourly candle green/red odds over 2014–2026 (~2,700–2,800 samples per hour). The deepest broker server-time hourly history available.
| Hour | Cycle | Green % | Red % | n |
|---|---|---|---|---|
| 17:00 | 17–21 | 51.9% | 47.9% | 2,804 |
| 18:00 | 17–21 | 55.7% | 44.1% | 2,805 |
| 19:00 | 🔒 | 🔒 | 🔒 | 🔒 |
| 20:00 | 🔒 | 🔒 | 🔒 | 🔒 |
| 21:00 | 🔒 | 🔒 | 🔒 | 🔒 |
| 22:00 | 🔒 | 🔒 | 🔒 | 🔒 |
| 23:00 | 🔒 | 🔒 | 🔒 | 🔒 |
18:00 is the greenest hour in this 12.5-year H1 view (55.7%); the shorter 4.2-year M15 view crowns 20:00. 17:00 and 22:00 are the most red-leaning (47.9%) — but neither is a true sell hour. The 21–24 cycle is uniformly weaker than 17–21.
5c) M15 green % — 21:00–24:00 (server, 4.2 yrs)
The late cycle is mostly a coin flip; only 22:45 shows a mild green edge, and everything after 23:00 leans slightly red.
5d) How long do green / red runs last? (H1, 12.5 yrs)
Consecutive same-color hourly candles, full history 2014–2026 (~14,700 runs per color).
Avg green run
2.07 hrs
▲ median 2 · max 16
Avg red run
🔒
Members only
Runs lasting 4+ hrs
🔒
Members only
P(same color next hour)
🔒
Members only
Streak takeaway
Green run distribution: 1 hr 48%, 2 hrs 25%, 3 hrs 14%, 4+ hrs 14%. Red: 1 hr 53%, 2 hrs 26%, 4+ hrs 10%. After any number of same-color ho… 🔒 Unlock to read the rest.
6) July by week-of-month — 22 years
Which part of the strongest month carries the green.
| Week | Days | Green % | Red % | Green yrs |
|---|---|---|---|---|
| Week 1 | 1–7 | 68% | 32% | 15 / 22 |
| Week 2 | 8–14 | 73% | 27% | 16 / 22 |
| Week 3 | 🔒 | 🔒 | 🔒 | 🔒 |
| Week 4 | 🔒 | 🔒 | 🔒 | 🔒 |
| Week 5 | 🔒 | 🔒 | 🔒 | 🔒 |
Early-to-mid July (W1–W2) is strongest; the last days (W5) flip red.
6b) Week-of-month - all months (weekly candles, 42y)
Weekly candles bucketed by the calendar week the candle starts in (W1 = starts on day 1-7 ... W5 = starts on day 29-31).
| Week | Green % | Avg return | n weeks |
|---|---|---|---|
| W1 (1-7) | 56.5% | +0.25% | 492 |
| W2 (8-14) | 55.9% | +0.40% | 492 |
| W3 (15-21) | 🔒 | 🔒 | 🔒 |
| W4 (22-28) | 🔒 | 🔒 | 🔒 |
| W5 (29-31) | 🔒 | 🔒 | 🔒 |
All-months view: W5 is greenest (small n - weeks rarely start on day 29-31) and W3 is softest. This does NOT contradict the July table above: that one slices July daily data by date range, where July's W5 (29-31) flips red. Scope matters. April + July weekly candles crown W2 (59.8%).
7) Green % — weekday × session block (server, 4.2y)
Each cell = share of days that block closed green. Monday is greenest in every block.
| Mon | Tue | Wed | Thu | Fri | |
|---|---|---|---|---|---|
| 04:00–12:00 | 55.8 | 55.7 | 53 | 54.3 | 51.6 |
| 12:00–15:30 | 59 | 49.3 | 53 | 55.7 | 54.6 |
| 15:30–17:00 |
8) Pre-US block - 4h candles anchored at 01:00 (server)
The 01:00-15:30 stretch sliced into 4h candles starting at 01:00, built from Pepperstone NAS100 M15 (4.2y). MT5 native H4 is midnight-anchored, so these are synthetic.
| Block | All months | Apr + Jul |
|---|---|---|
| 01:00-05:00 | 54.2% | 49.2% |
| 05:00-09:00 | 53.0% | 55.5% |
| 09:00-13:00 | 🔒 | 🔒 |
| 13:00-15:30 | 🔒 | 🔒 |
All-months: the overnight 01:00 block is greenest. In April + July the pattern flips - 01:00 turns red-leaning and 05:00 (Asia morning) leads at 55.5%.
9) M15 green % - 15:30-17:00 (US pre-open ramp, server)
The 90 minutes into the US cash open (~16:30 server). Pepperstone M15, 4.2 years. Mostly coin-flip - the edge lives later, at 18:00.
10) April + July - Super TD (TOP 3 per dimension)
The two strongest months only. TOP 3 greenest and TOP 3 reddest per dimension. Q1-3 from Yahoo ^NDX (41y); intraday from Pepperstone NAS100 server time (H1 12.5y, M15 4.2y).
| Dimension | TOP 3 green | TOP 3 red |
|---|---|---|
| Month | Jul 65.9% · Apr 61.0% | - |
| Week of month | W2 59.8% · W1 54.9% · W4 54.3% | W3 49.4% · W5 46.4% · W4 45.7% red |
| Weekday | 🔒 | 🔒 |
| Pre-US 4h block | 🔒 | 🔒 |
| M15 15:30-17:00 | 🔒 | 🔒 |
| H4 evening | 🔒 | 🔒 |
| H1 17:00-24:00 | 🔒 | 🔒 |
| M15 17:00-24:00 | 🔒 | 🔒 |
18:00 gets even stronger in Apr+Jul (59.3% vs 56.6% all-year) and the 17:45 red dip deepens (58.2% red vs 54.8%). M5 slots are omitted - only ~28 Apr+Jul samples exist, too noisy to rank.
Note
Historical frequencies, not predictions. Daily/weekly/monthly windows now reach back up to 41 years (Yahoo ^NDX); intraday windows cover 12.… 🔒 Unlock to read the rest.
Want the full picture?
Unlock every number, every day, and the complete weekly game plan. $7 for 7 days or $25 for 30 days — pay by card or USDT.
Unlock full access