Actionable insights — Weekly ChartStorm, 31 May 2026

1. Track the cluster of extremes as a regime gauge — not a timing signal

1. Maintain a small panel of independent sentiment/positioning extremes — expectations (share expecting higher prices), allocations (household equity weighting), valuations, and the long-term earnings-growth estimate — and check how many are at or near record highs at once.
2. When several light up together, classify the regime as late-cycle euphoria ("the age of speculation") — elevated vulnerability — rather than predicting a top from any single reading.
3. Pair the regime read with a trend filter (here: price above its 10-month moving average) so you don't fight an intact uptrend just because the gauges are stretched. "The trend is still your friend" until it breaks.

• How many gauges hit fresh records simultaneously; the trend filter rolling over (price losing the 10-month MA); whether new extremes are confirmed or diverging across the panel.

2. Don't sell on valuation alone — wait for a catalyst

1. Treat high valuations as a condition, not a trigger: expensive markets can stay expensive for years and rarely crack on multiples by themselves.
2. Identify the catalysts that actually end bull markets — recession, a shock/crisis, or prolonged monetary tightening — and monitor those directly, rather than acting the moment valuations look stretched.
3. Separately note fading structural tailwinds (e.g. buybacks going from major support to near-irrelevant, plus IPO/issuance supply ahead): they raise fragility but still need a catalyst to bite.

• Recession signals, credit/financial shocks, or a tightening cycle; the buyback bid weakening into heavy IPO/issuance supply; any one of these arriving while the gauges are euphoric.

3. Read overbought leaders and crowded shorts as opposing squeeze setups

1. Flag where the crowd has fully piled in: semiconductors "fully overbought," the consensus "sure thing" / "religion of AI" — a spot where the marginal buyer is scarce and disappointment hits hardest.
2. Separately, measure the crowded short side: median stock short interest at decade-plus highs is latent fuel — if the rally broadens, those shorts must cover and assist the move higher (a short squeeze).
3. Use a breadth trigger to tell which way it resolves: an equal-weight S&P breakout says the rally is broadening beyond the mega-caps — conditions for the squeeze to fire rather than the overbought leaders to roll.

• Equal-weight vs cap-weight breadth (confirming/failing breakout); short-interest levels and direction; semis losing momentum from an overbought extreme.

4. Hunt overlooked strength outside the crowded AI trade

1. While everyone crowds the consensus winner (semis/AI), scan the opposite shelf — the left-for-dead and the ignored — for assets quietly turning up. Crowded attention elsewhere is what keeps them cheap.
2. Look for a momentum inflection: a sector going from "slow recovery" to "sharp and surging" (a single big up-day like IGV +6.25% Friday is a tell), or a beaten-down factor still mid-trend.
3. Cross-check the contrarian story against history so you don't mistake mid-run for late-run: an overbought reading isn't automatically a top if past advances ran far further (Cappelleri's IWM +70% vs the +150% 2020–21 run).

• Inflection up-days in left-for-dead sectors; whether an overbought reading sits early or late in the historical analogue; rotation breadth into non-AI corners.

5. Decompose a hot regional trade to test what's really driving it

1. When a region rips, don't assume one cause — break the return into independent drivers and test each: is it concentration (a couple of names/markets), currency (weak-USD flattery), or broad strength?
2. Use a paired-ETF spread to isolate a factor: two near-identical funds that differ in one constituent (here IEMG includes Korea, VWO doesn't) — the YTD gap (IEMG ~2× VWO) is the Korea/AI contribution.
3. Strip out the suspected driver and re-measure: EM ex-Taiwan+Korea, equal-weighted, and EM in local-currency terms. If the trade still works after removing each driver, the thesis is multi-legged and more durable — reiterate it.

• The IEMG-vs-VWO spread (Korea/AI contribution); the EM local-currency index vs USD index (FX vs real strength); whether ex-TW/SK breadth holds if the AI complex cools.

6. Use software and Bitcoin as cross-asset risk barometers

1. Keep a couple of high-beta, speculation-sensitive assets as barometers of risk appetite — they often telegraph the broader tape's tolerance for risk before the index does.
2. Compare them: when one is "sharp and surging" (software/IGV) and another "looks less convincing" (Bitcoin), the divergence is the signal — risk appetite is selective, not uniform.
3. Read the barometers alongside the regime panel: confirming strength supports staying with the trend; a rollover in the barometers is an early caution flag even before the gauges break.

• Software vs Bitcoin momentum agreeing or diverging; either rolling over while the index holds (early warning); broadening vs narrowing risk appetite across the barometers.