Out-of-sample performance of timestamped model portfolios published since 2026-06-26 · unfunded model portfolio — not a client account or independently audited investment product · gross of fees, costs and taxes
Live since 2026-06-26 Updated 2026-07-16 19:32 UTC 15 trading days · re-marked every rerun
CVI's World Portfolio is one systematic global-value book — fundamental value fused with machine learning across 2,900+ companies in 26 markets — from Wall Street and US small-caps to Tel Aviv, Mexico City, Kuala Lumpur and Auckland. Every week it re-ranks the world and the model portfolio updates toward the cheapest quality. This page is the out-of-sample record of the published model outputs — timestamped before the fact, then marked to subsequent prices. It's shown four ways: pre-tax, two illustrative tax scenarios, and versus simply buying-and-holding.
Unfunded model portfolio. These are the published model outputs marked to adjusted closing prices — out-of-sample and timestamped each run, but not a funded client account, not net of fees, costs or taxes, and not independently audited. The record began on 2026-06-26 and is deliberately short — it grows every week. The multi-year hypothetical backtest (2024-07-17 → 2026-07-16) is separate, on the performance factsheet.
+1.2%
Following rebalances
+1.1%
Illustrative tax scenario + TLH
+0.9%
After tax (no TLH)
+1.6%
Buy & hold
+1.1%
MSCI ACWI benchmark
global equity index
+0.1 pts
Excess vs ACWI
-0.4 pts
Value of rebalancing
+0.2 pts
TLH benefit
−0.1 pts
Tax drag
3
Rebalances
live since 2026-06-26
Live results by portfolio tier — which configuration you are viewing
The same live book, shown as each of the five portfolio configurations you can research — from the broadly-diversified Institutional configuration to the concentrated Aggressive. This is the realised result of the selected configuration.
portfolio tier / configuration
Holdings
Following rebalances
Buy & hold
Institutional configuration
The full deep book — broadest diversification
250
+1.2%
+1.6%
Conservative
Top 100 by conviction — steady core
100
+0.9%
+1.7%
Balanced
Top 50 — the mid-concentration tier
50
+0.9%
+2.4%
Growth
Top 25 — higher conviction, larger swings
25
+0.5%
+2.9%
Aggressive
Top 12 — most concentrated
12
+0.0%
+4.4%
These five portfolio configurations — all cut from the same global book, from the broadly-diversified Institutional configuration (250 names) to the concentrated Aggressive (12). Both figures are out-of-sample model results, pre-tax, since inception, marked to the latest close. Following rebalances = holding the weekly-signal book and rotating each week; Buy & hold = holding that tier's inception book untouched — the gap between them is the difference attributable to applying the published model changes.
Click any configuration to update its realised stats and highlight its growth-of-$1 line — each figure is computed from that tier's own realised curve.
Growth of $1 — since inception
The four realised paths of $1 since inception, marked to live prices, out-of-sample — for the selected portfolio tier above.
Applying published model changes (pre-tax).Grows $1 by holding the current published model book until the next published model change, then rotating — before any tax.
Illustrative tax scenario (no harvesting).The same strategy paying a blended 20% global capital-gains tax on net realised gains at each rebalance, with losses left unused — the honest floor.
Illustrative tax scenario + TLH.The same, but realised losses are harvested to offset gains (immediately usable and reinvested), subject to the US 30-day wash-sale rule — so it sits above the no-harvesting line.
Buy & hold.Buys the first published model book and simply holds it, never rebalancing — the gap to the pre-tax line is the difference attributable to applying the published model changes.
Rebalance & mark history — every run
Date
Type
Portfolio
Holdings
Return (day)
Cumulative
2026-07-16
Mark
Institutional configuration · World Portfolio
250
+0.1%
+1.2%
2026-07-15
Mark
Institutional configuration · World Portfolio
250
+0.2%
+1.1%
2026-07-14
Mark
Institutional configuration · World Portfolio
250
+0.1%
+0.9%
2026-07-13
Mark
Institutional configuration · World Portfolio
250
-0.2%
+0.8%
2026-07-12
Mark
Institutional configuration · World Portfolio
250
+0.0%
+1.0%
2026-07-11
Mark
Institutional configuration · World Portfolio
250
+0.0%
+1.0%
2026-07-10
Rebalance
Institutional configuration · World Portfolio
250
+0.3%
+1.0%
2026-07-09
Mark
Institutional configuration · World Portfolio
250
+0.5%
+0.6%
2026-07-08
Mark
Institutional configuration · World Portfolio
250
-1.4%
+0.1%
2026-07-03
Rebalance
Institutional configuration · World Portfolio
250
+1.5%
+1.5%
2026-06-26
Rebalance
Institutional configuration · World Portfolio
250
+0.0%
+0.0%
One row per trading day, marked to that day's close. Return (day) is the book's move that day and Cumulative is growth-of-$1 since inception — so each day compounds cleanly into the next. Rebalance days are when the weekly signal changed the book; Mark days re-value the held book at fresh prices. Portfolio shown is the flagship Institutional configuration (per-tier standings above).
How this record is measured
The strategy.One systematic global-value book — fundamental value fused with machine learning across 2,900+ companies in 26 markets. Each week it re-ranks the world and rebalances toward the cheapest quality. The Balanced configuration is a ~50-name cut of this book.
Timestamped, out-of-sample.Each weekly book is recorded with its date before prices move, then held until the next published model change. Between signals the book is simply re-marked to the latest prices — daily runs don't count as new rebalances — so the record reflects the strategy's real weekly cadence.
Marked to live prices.Each book is valued forward from its recommendation date using adjusted closing prices, FX-adjusted to USD — the figure is what that model book would have been worth at those marks — an unfunded model calculation, not a client account.
Tax assumptions.A blended 20% capital-gains rate is an assumption — jurisdictions differ widely (Germany has no wash-sale rule; several Asian markets levy no CGT at all). Losses are treated as immediately usable against outside gains.
Gross, honest, early.Figures are gross of fees and trading costs, are not independently audited, and the record is deliberately short. Short records can move sharply and are not indicative of future results.
Not the backtest.This page is the live record only. The multi-year, hypothetical strategy backtest is a simulation, shown separately on the performance factsheet.