◎Reading the tape… If this doesn't fill in, the data files haven't loaded — run python refresh.py --all.Regime detail →
Action queueranked by urgency · reason → action → proof
Computed from the current mark against each frozen entry and stop — not from the alert log, which records when a condition first fired and can be weeks stale. Every row shows the numbers behind it: expand proof to audit the arithmetic before you act.
Honest note: This gauge measures recession RISK, not the exact calendar stage. Low risk ≠ early cycle — it means no downturn signal on the 12-month horizon. For position sizing that’s what matters.
Regime Gate
SECTION 1
Market Breadthcomputed · % of stocks above their EMAs
Gate Indicatorsevery number search-verified · source + as-of
Indicator
Value
Read
As-of
How to read:green = supportive, red = a warning sign. Gold-highlighted rows are the ones that decide the verdict. Index comfortably above its 200-day + healthy breadth + risk-on internals = GREEN → full-size permitted. The rich P/E is a caution, not a downgrade.
10-Year ContextIS TODAY'S READING HIGH OR LOW?
How to read: each mini-chart is the recent path; the gold dot is today. The tag says where today sits in its ~10-year range — LOW / MID / HIGH. Valuation (P/E), yields and gold are all at the HIGH end — that's the "rich" caution behind GREEN.
Gate Logic
GREEN index >200D + breadth >55% + RSP/SPY rising → full-size OKYELLOW >200D but internals diverging → half-size, tighten stopsRED index <200D → no new longs; defensive doors only
US Sector Rotation · RRG30 ETFs vs SPYRS 3M · Mom 1M
How Leading / Improving / Weakening / Lagging are computed ▸
Two numbers place every name on the clock, both measured against the benchmark (SPY):
rs = 3-month return − SPY 3-month return → the x-axis: is it stronger than the market over a quarter?
mom = (1-month return − SPY 1-month return) − rs → the y-axis: is that strength accelerating? Subtracting rs is what makes this a rate-of-change rather than a second strength reading.
The quadrant is then just the pair of signs:
Quadrant
rs
mom
Meaning
Leading
≥ 0
≥ 0
stronger than the market and still accelerating
Weakening
≥ 0
< 0
still stronger, but momentum is rolling over — the usual exit warning
Lagging
< 0
< 0
weaker and getting weaker
Improving
< 0
≥ 0
still behind, but turning up — where early entries live
Identical maths on both engines (US vs SPY, MY vs KLCI) so the two clocks are directly comparable. A dot near the centre is barely distinguishable from the benchmark — treat small distances as noise.
The clock: ETFs travel Improving → Leading → Weakening → Lagging. Buy Improving (early), ride Leading, trim Weakening, avoid Lagging. x = strength vs SPY (3M rel), y = whether that strength is accelerating (1M rel − 3M rel).
Rotation Leaderboardrelative to SPY · sorted 1M rel
How to read: every number is the ETF's return minus SPY's over that window — green = beating the index, red = lagging it. Colour intensity ranks each column. quad = its spot on the rotation clock above.
Ranked Universetop of 30-ETF book▲ vs200D · flow IN/OUT
#
Ticker
vs 200D
EMA20/50
3M RS
Value
Flow
Flag
How to read: we only BUY names in an uptrend (ABOVE 200D + bullish EMAs) that are also beating the market (positive 3M RS). TRIPLE ✦ = the best possible setup. BREAKDOWN / value-trap = AVOID, no matter how cheap. Gold row = the one we bought.
3-Month Trend30 ETFs · rebased price + distance to 200-EMA
How to read: each card is the ETF's price over the trailing 12 months. The big number is today's price vs its 200-EMA — green = above (uptrend), red = below (downtrend). Above the 200-EMA is the trend gate; the further above, the stronger the trend.
Flagscomputed from the ranked table
Breakdowns — do not chase
🔄 Reversal Radarwatch-only · never a buy
How to read: a name is ARMED only when it is deeply off its high (≤−20%), a fresh EMA20>50 cross has fired, and relative strength is accelerating (1M rel > 3M rel). This is a watch-only hint — the fund never buys it; the 200-EMA trend gate still governs every real entry.
🧭Two layers in one place. The curated baskets (judgment) sit on top of the systematic 27-sector scanner (quant, equal-weighted vs KLCI). Baskets tell you what to buy; the scanner is the wide radar that feeds them.
Basket Radarcomposite trend · 3M RS vs KLCI · catalyst
Basket
vs 200D
Lead EMA20/50
3M RS
Value
Catalyst this week
How to read: each "basket" = a group of Malaysian stocks in one theme. We want baskets ABOVE their 200D trend, outperforming the KLCI, with a real catalyst (contract, policy). Gold rows = the leaders the fund acted on. Red text = a warning (outflows, weak trend).
Drill-down · Constructiontrend+flow+catalyst
SUNCON — above 200D, higher lows, RM8.2bn order book, foreign +RM19.9m last wk. Gamuda — above 200D, DC award, near support. ⚠ Jul-9 share subdivision — verify basis. MN Holdings — DC-grid pure-play, stacked substation wins (RM216/128/276/122m).
Drill-down · Utilities/Gridstructural tailwind
YTL Power — near 2-yr high RM4.41, above all EMAs, DC-listing value-unlock live. TNB — above 200D, Green Lane DC-connection driver, grid-capex anchor. MN Holdings — bridges grid + construction baskets.
No clean TRIPLE: the trend leaders (Tech, Cons, Grid) are RICH not cheap — valuation is the only non-confirm. Flagged as strongest composite setups (bullish trend + flow/catalyst IN).
🔄 Reversal Radarwatch-only · never a buy
How to read: a name is ARMED only when it is deeply off its high (≤−20%), a fresh EMA20>50 cross has fired, and relative strength is accelerating (1M rel > 3M rel). This is a watch-only hint — the fund never buys it; the 200-EMA trend gate still governs every real entry.
The Scanner27 Bursa sectors · equal-weighted vs KLCI
3-Month Sector Trendself-built index · distance to 200-EMA
How to read: each card is the sector's equal-weight index over the trailing 12 months. The big number is today's price vs its 200-EMA — green = above (uptrend), red = below (downtrend). Above the long-term EMA is the trend gate; the further above, the stronger.
BenchmarksBursa vs Wall Streetrotation is measured vs KLCI
Rotation Leaderboard27 sectors × 6 windows · relative to KLCIclick a row → top-10 by revenue
How to read: every number is the sector's return minus the KLCI's over that window — so green = beating the market, red = lagging it (rotation is relative, not absolute). Colour intensity ranks each column. quad = its spot on the rotation clock. Sort priority is 1M relative strength. Click any sector row to expand its top-10 constituent stocks ranked by revenue.
Rotation Quadrant · RRGRS 3M · Mom 1M
How Leading / Improving / Weakening / Lagging are computed ▸
Two numbers place every name on the clock, both measured against the benchmark (KLCI):
rs = 3-month return − KLCI 3-month return → the x-axis: is it stronger than the market over a quarter?
mom = (1-month return − KLCI 1-month return) − rs → the y-axis: is that strength accelerating? Subtracting rs is what makes this a rate-of-change rather than a second strength reading.
The quadrant is then just the pair of signs:
Quadrant
rs
mom
Meaning
Leading
≥ 0
≥ 0
stronger than the market and still accelerating
Weakening
≥ 0
< 0
still stronger, but momentum is rolling over — the usual exit warning
Lagging
< 0
< 0
weaker and getting weaker
Improving
< 0
≥ 0
still behind, but turning up — where early entries live
Identical maths on both engines (US vs SPY, MY vs KLCI) so the two clocks are directly comparable. A dot near the centre is barely distinguishable from the benchmark — treat small distances as noise.
The clock: sectors travel Improving → Leading → Weakening → Lagging. Buy Improving (early), ride Leading, trim Weakening, avoid Lagging. x = strength vs KLCI, y = whether that strength is accelerating.
Sector Index Trendsrebased 100 · 1Y
Each basket rebased to 100 a year ago (self-built sector index — Bursa's own sub-indices aren't public). Dashed = KLCI & S&P 500. Click a chip to show/hide · double-click to solo.
📰 Malaysia Market NewsThe Edge · The Star · Google News
Virtual Fund 🤖 the automated portfolio
ⓘ How execution works →NO SHORTING · LONG/CASH ONLYSECTION 4 · $100kstruck 2026-07-10 · marked at fill
🤖This is your automated portfolio. The engine built these positions itself from this week's flags. It's paper/virtual (not a real broker); every fill is frozen point-in-time, net of realistic fees.
How the fund is actually doingvs both benchmarks, since inception
Performance live track record since inception
Execution Blotter & Exit Ladderstops provisional · to be tightened to exact 200D/swing-low
Sym
Eng
Shares
Entry
Now
P&L
Net cost $
Wt
Init. stop
Thesis
How to read: these 8 trades were placed automatically from the flags. Entry = price paid. Net cost = what it cost after fees. Init. stop = the price where the system cuts the loss. The exit ladder below is the staged sell plan — all positions are currently rung 0 (healthy).
Exit ladder:0 healthy1 trendline broken → tighten2 swing low broken → trim ½3 EMA cross down → trim to stub4 200D broken → exit· all positions currently rung 0
Valuation — modelled, then stressedjustified P/B · Gordon Growth · click a row for sensitivity
How to read:gold = price · blue = EMA20 · purple = EMA50 · grey thin = 200D (computed from the same 1Y series). Gold dashed = frozen entry · red dashed = stop. Entries are frozen point-in-time — later data never rewrites them.
Conviction ScorecardSIZE = REGIME BASE × MULTIPLIERGREEN base 8% · this is the thesis
Sym
Trend
Mom
RS
Flow
Value
Cat
Score
×Mult
Size
Note
How to read: each pillar is ✓ (pass) or ✗ (fail), after the mandatory Trend gate (200D). Score → multiplier → size, all mechanical. Struck-through/dimmed rows failed the gate or the score threshold — every scored name is shown, passes and fails, because the fails ARE the systematic thesis.
Technical LevelsEMA20/50 · 200D · support sets the stop
Sym
Price
EMA20
EMA50
200D
Cross
Support (→ stop)
Resistance
How to read: EMA20 above EMA50 = up-momentum. Price above 200D = the mandatory trend gate. The Support level is a real recent swing-low — that's where each stop sits, not an arbitrary %.
🚫 Watchlist — dropped this weekno position
Fee Model
US ETFs: $0.0035/share, min $0.35 (IBKR-tiered) — trivial, counted. MY stocks: 0.10% brokerage + 8% SST on brokerage + 0.03% clearing + 0.10% stamp duty ≈ 0.24%/side. FX: 0.20% USD→MYR (rate 4.0770). Total entry cost: $140.82 = 0.14% of fund.
Excluded — firewallno entry
SMH — top RS but below 200D. XLE / GLD / GDX / URA — cheap but downtrending.
Rule: never buy cheapness without a price-trend trigger. Re-arm only on a reclaimed EMA50 with RS turning up.
Backtest
PRICE-ENGINE ONLY
⚠ Scope & Caveats — read firsthonest by constructionPRICE ENGINE ONLY · NOT THE LIVE SYSTEM95% INVESTED (STOP-LOSS JUSTIFIED) · INTRA-MONTH STOPS MODELLED
📋 How the backtest actually buys & sells — full stepsclick to expand
Beating the S&P is worth scrutinising — so here's exactly how every historical fill was decided. No black box.
1
The clock — monthly rebalance
Decisions are made on the last trading day of each month (~112 rebalances since 2017). Between rebalances, every holding's stop is checked on every daily close.
2
Regime gate first
S&P > its 200-EMA → GREEN (base unit 8%). Otherwise RED → 100% to cash, no longs. The regime alone decides whether we invest at all.
3
Score every name (price only)
Gate: price > 200-EMA — fail = skip. Then score = (EMA20>EMA50 ? 1 : 0) + (3M return > benchmark's 3M ? 1 : 0) → 0, 1 or 2. Keep names that pass the gate and score ≥ 1.
4
Rank & size
Rank by score, then by 3-month relative strength. Size = 8% × (score 2 → ×1.5, score 1 → ×1.0), capped at 10% per name. Fill top-down until 95% invested (5% cash), no engine (US/MY) over 52%, max 12 positions.
5
BUY
New qualifiers are bought at that month's close. MY names are priced in RM and converted at the day's USD/MYR. Fees are charged on every turnover.
6
HOLD
A position is carried unchanged to the next rebalance unless it's re-sized or sold.
7
SELL — two ways
(a) Monthly: at the next rebalance if it drops out of the book (fails the gate or gets out-ranked). (b) Intra-month: if any daily close breaches its stop = max(200-EMA at entry, entry × 0.92) → exit to cash that day; it stays out until it re-qualifies. This is the same exit-ladder the live fund uses.
8
What it does NOT use — why the edge is honest
Only the price rules above. It deliberately excludes the weekly flow/value/catalyst reads (those never existed historically → including them would be look-ahead), and uses today's universe (survivorship). So the 18.8% CAGR is the trend + relative-strength core, not a curve-fit story. Verify any single fill in the month-by-month ledger below.
Equity Curvestrategy vs SPY vs KLCI · rebased 100 · monthly
All three curves rebased to 100 at the start. Click a chip to show/hide a series · double-click to solo. MY names and the KLCI are valued in USD so all three share one base currency.
What actually drove the returncompleted round-trips, ranked by contribution
How this is computed: every position is tracked from its buy month to the month it left the book (sell or stopped), using the point-in-time prices already in the ledger below. Contribution = trade return × the weight it was held at, so a +80% winner at 10% weight adds ~8 points. Sorted by contribution, not by headline return — a huge gain on a tiny position did not drive the curve. Read the points as a ranking, not an identity: they are simple return×weight figures, so they will not add up to the headline compounded return — positions overlap in time and gains compound on a moving NAV.
Trade ledger — verify every fillby year → month → holdings
buysellstoppedhold
Every month-end book the price engine struck, expandable down to each fill (symbol · action · native price · weight). Prices are point-in-time; stopped rows exited intra-month on a daily close through the hard stop. Re-check any fill against source history.
10x Compounder 🤖 $10k monthly paper fund
SCREENER + FUND
🚀Long-term compounder screener + a $10k paper fund. Every name is scored /100 on the CIO rubric. The fund holds the top 3 (score ≥85), equal-weight, and rebalances once a month — selling names that fall out, buying new entrants. Paper only, no broker.
Fund · Holdings
Ticker
Name
Score
Shares
Price
Value $
Wt
P&L
How it works: holds the top 3 by score, equal-weight; sells names that fall out of the top 3 and buys new entrants once per calendar month. Next rebalance . Gold dashed = starting $10k. Paper — no real orders.
Screener · Universe
Ticker
Name
Score
Rev CAGR
EPS CAGR
ROIC
D/E
10x
100x
Industry
Score /100 on growth · capital efficiency · balance sheet · business quality · management · industry runway. Green ≥85 buy-ready · amber 80–84 watchlist. ◆ = estimate, verify vs filings.
Fund · Order Loglast 40
When
Action
Ticker
Shares
Price
Value $
Why
Research & World Hub
SECTION 5
📰Research & the wider world. MY research (i3investor), a free US research library, plus Global and China desks — and AI opportunity briefs synthesised from the rotation data. Market news now lives on the US and MY tabs. A headline is a confirmation input, never a trend trigger. Links open the original article in a new tab.
📊 News by Sectorevery headline classified to your baskets
🤖 AI Opportunity Briefssynthesised from PULSE rotation data
Each brief is stamped AI SYNTHESIS — not a bank report. Every number is templated straight from the US & MY rotation files; nothing is asserted that isn't derivable from the data.
🇲🇾 MY Research · i3investorbroker research headlines
🇺🇸 US Industry Research (free)McKinsey · BCG · Bain · Deloitte · IMF · OECD · BIS · World Bank · Fed
LATEST INDUSTRY REPORTS
🇺🇸 Leading Firms · Analyst Researchthe industry read, through its bellwethers