EnInvs · Quantitative strategies

Backtested portfolios that beat the index — and the reasoning behind them

Three complementary signals — growth-at-a-reasonable-price (FVC), cheap-cyclical upside, and bank forward earnings yield — computed per company, every day, across seven markets. Every claim here is a transparent, reproducible backtest, not a marketing chart.

US Leaders +1.9 pp/yr alpha vs S&P Global Commodities +5 pp/yr vs S&P RU banks +332% since 2023 3 signals · 7 markets

Why these strategies work — and what traditional analysis misses

Each edge below is exploited by a specific signal on this page; the dashed line is the gap in the conventional approach it sidesteps.

US Leaders

The growth-acceleration anomaly

When a company grows faster than 15%/yr, in 66% of cases it keeps growing fast the next year. We buy that acceleration only while it is reasonably priced, and exit at the first sign of a slowdown — sidestepping the IBM/Motorola trap and the Netflix/Amazon multiple bubble.

Traditional gap: investors either chase growth blindly or ignore it — few price the intersection of acceleration and a sane multiple.

Global Commodities

Nowcasting, not a stale DCF

184+ commodity prices refresh every 2–5 minutes and filings are parsed several times a day, so each producer’s EV/EBITDA is repriced in near real time against mid-cycle earnings — catching cheap cyclicals before the consensus updates.

Traditional gap: sell-side DCFs lean on subjective WACC and terminal-growth assumptions, refreshed once a quarter at best.

Commodities · EM

The currency tailwind

An emerging-market producer earns in dollars but pays costs in a depreciating local currency. Devaluation quietly widens margins — a structural edge for Brazilian, South-African and Kazakh exporters that the spot-price headline alone hides.

Traditional gap: single-currency screens miss the cost-side FX boost to margins entirely.

Banks

Forward yield + monthly data

Banks have no EBITDA, so we rank them by ROE / (P/B) = net income / market cap — a forward earnings yield. For Russia we refresh it with monthly CBR data, catching a turn in profitability a full quarter before the IFRS print.

Traditional gap: most screens wait for quarterly statements and value banks on trailing book, missing the inflection.

EnInvs Strategies — recommended portfolios & backtests

Period 2015-01-01 … 2026-06-08. Source: EI1. Generated by Strategies/build_report.py.

How to read this. Backtests are vectorized, equal-weight, monthly rebalance, no transaction costs (matching the PHP engine). Returns are price-only (no dividends), benchmarked against the price index SP500_index. US Leaders universe = point-in-time S&P 500 (top-50 by cap) → FVC≥14 + IsEbitdaBest → top-15 by FVC. Global Commodities = WD universe, EV/EBITDA in [0,20], ranked by EI1 Potential_min3y, top-15.

Caveats. (1) US Leaders: the mechanical preset ≈ S&P (+0–2pp); FVC is a quality filter, not a return-ranking signal. The live fund's larger edge is discretionary + total-return vs price-index. (2) Global Commodities: backtest beats strongly but is upward-biased (delisting losses only partially captured, illiquid penny names equal-weighted, no costs). The live GC_Value made far less. Treat replicas as the mechanical floor, the live series as the discretionary realisation.

Below: the framework extended to Russia / Indonesia / Mexico (total return, dividends included) plus a single-stock signal-efficacy test.

Equity curves

US Leaders

	Total	CAGR	Vol	Sharpe	maxDD
US Leaders (EW)	208.8%	10.3%	18.9%	0.56	-33.0%
S&P 500	258.7%	11.8%	—	0.68	-33.9%

Excess CAGR -1.48% · beta 0.95 · alpha(ann) -0.19%

Recommended portfolio — 2026-06-08 (capped-proportional, cap=150%) (prod = US)

Ticker	Weight	FVC %	Last price	Prod weight
PEP	28.3%	29.0%	141.92	13.9%
NVDA	25.2%	25.8%	205.10	0.0%
MSFT	16.9%	17.3%	416.67	11.6%
IBM	15.2%	15.6%	284.84	0.0%
AMZN	14.3%	14.7%	246.03	9.8%

Calendar-year returns

Year	Strategy	S&P 500
2014	0.0%	—
2015	7.7%	—
2016	1.9%	9.5%
2017	25.8%	19.4%
2018	-7.1%	-6.2%
2019	22.9%	28.9%
2020	5.5%	16.2%
2021	17.6%	27.4%
2022	-2.7%	-19.9%
2023	3.8%	24.5%
2024	5.5%	23.5%
2025	22.3%	16.4%
2026	21.4%	7.7%

Global Commodities

	Total	CAGR	Vol	Sharpe	maxDD
Global Commodities (EW)	340.9%	13.8%	31.0%	0.43	-68.2%
S&P 500	258.6%	11.8%	—	0.60	-33.9%

Excess CAGR +2.00% · beta 0.95 · alpha(ann) +5.05%

Recommended portfolio — 2026-06-08 (capped-proportional, cap=150%) (prod = GC_E)

Ticker	Weight	Potential %	Last price	Prod weight
BRY	6.4%	2231.1%	0.41	0.0%
BATL	6.4%	2070.6%	1.58	0.0%
BRFS	6.4%	551.6%	3.38	11.9%
CRLBF	6.4%	498.9%	0.90	11.9%
AMPY	6.4%	371.7%	4.81	0.0%
TROX	6.4%	286.0%	7.34	0.0%
PBF	6.4%	261.8%	43.21	0.0%
AA	5.8%	135.4%	71.89	10.6%
SBSW	5.7%	132.7%	10.12	10.5%
IPOAF	5.6%	130.9%	50.00	0.0%
CDE	5.3%	125.3%	19.01	0.0%
DAC	5.2%	121.9%	129.17	9.7%
KOS	5.1%	118.6%	2.98	0.0%
BKR	4.1%	95.5%	64.22	0.0%
EPM	3.6%	85.4%	4.41	0.0%
PNRG	3.2%	74.8%	189.60	0.0%
CNQ	3.2%	74.4%	45.70	5.8%
CF	3.1%	73.0%	113.49	5.7%
ANDE	2.7%	62.5%	72.81	4.8%
TECK	2.6%	61.6%	61.67	4.8%

Calendar-year returns

Year	Strategy	S&P 500
2014	0.0%	—
2015	-44.0%	—
2016	58.6%	9.5%
2017	20.6%	19.4%
2018	-18.0%	-6.2%
2019	13.1%	28.9%
2020	-15.5%	16.2%
2021	120.7%	27.4%
2022	67.1%	-19.7%
2023	-9.8%	24.2%
2024	1.7%	23.5%
2025	35.4%	16.4%
2026	14.7%	7.7%

Strategy logic & methodology — the three signal families

Everything on this page rests on three complementary cross-sectional signals, each computed per company and per day, then used both to build portfolios (pick the top names) and to time single stocks (be invested only when the signal is favourable).

FVC — Fundamental Value Created (a quality/growth signal). Rewards companies whose operating earnings are growing and that are not over-priced for that growth.
Commodity-Upside / Potential (a value/mean-reversion signal). Rewards cheap cyclical producers — those trading below a fair EV/EBITDA given mid-cycle earnings.
Banks — return on investment (ROE / P/B) (a profitability signal for financials). Rewards banks with the highest forward earnings yield; for Russia refreshed with monthly RAS data.

1. FVC — Fundamental Value Created

FVC is a port of the platform's getCreationValue (the same number the live US Leaders page shows). For the latest reported period it measures the value a company creates per unit of market cap:

growth_term = (EBITDAnow − EBITDAyear ago) × (EBITDALTM / EBITDAnow)

aim_mult   = ( min(EV/EBITDALTM, 30) + g ) / 2

FVC = ( growth_term × aim_mult + FCFLTM ) / market_cap

In words: take the year-on-year growth in operating earnings, value that growth stream at a blended multiple (the company's own current EV/EBITDA, capped at 30×, blended with a market constant g), add trailing free cash flow, and divide by market cap. A high FVC means "a lot of value created relative to what you pay." The market constant g reflects the normal EV/EBITDA of each market — US/Global g = 12, Russia g = 6.5, Indonesia/Mexico g = 8. Empirically FVC behaves as a quality filter: it reliably keeps you out of value-destroyers, but concentrating by it does not, on its own, rank the future winners (its cross-sectional IC is ≈ 0).

2. Commodity-Upside / Potential

For commodity producers the multiple is fairly stable through the cycle but EBITDA swings with spot prices, so the signal is an EV/EBITDA reversion:

upside = ( fair_mult × EBITDALTM − net_debt ) / market_cap − 1

i.e. what the equity is worth if the name reverts to a fair EV/EBITDA, versus today's price. For Russia and the Global Commodities / Oil&Gas E&P books we use the platform's authoritative Potential_min3y series directly (a richer version that reprices revenue and cost by each commodity's price move and reverts to the name's own historical multiple, with a reserves-life haircut). For Indonesia/Mexico, where no historical potential exists, we compute the simple reversion above from extracted financials. Economically this is a cheap-cyclical signal.

3. Banks — return on investment (ROE / P/B)

Banks have no EBITDA, so neither signal above fits. We rank them by a forward return on investment — what the equity earns per unit of price:

potential = annualized ROE / (P/B) = annualized net income / market cap

The book value cancels, leaving a forward earnings yield: annualized net income over market cap. For Russia we add an edge no other market offers — banks file monthly RAS results (CBR Form 102) weeks ahead of the quarterly IFRS, so we refresh the run-rate with the freshest month. This catches a turn in profitability a full quarter early. Portfolio: each period, long-only, weight every bank by how far its yield exceeds the bank-universe median; benchmark the local index and an equal-weight bank basket.

Market	Window	Signal tilt	Equal-wt basket	Local index	What it shows
Russia	from Jan 2023	+332% (CAGR 53%)	+26…90%	≈ flat (IMOEX)	monthly RAS data — a turn caught a quarter early
United States	2018–2026	+180% (CAGR 13%)	+65% (CAGR 6%)	+178% (S&P 500)	~3× the bank basket; level with the S&P though banks lagged
Mexico	2019–2026	+235% (CAGR +19%)	+174% (CAGR +16%)	+59% (IPC)	signal tilt clears the basket and crushes the index
Indonesia	2019–2026	+28% (CAGR +4%)	+31% (CAGR +4%)	-4% (JCI)	4 near-identical mega-banks — tilt adds nothing over equal weight
Kazakhstan	2019–2026	+412% (CAGR 42%)	+140%	+880% (KASE)*	signal beats the basket; index carried by Kaspi's run
Philippines	2019–2026	−35% (CAGR −6%)	−23%	−26% (PSEi)	the honest counter-example — banks fell, cheap ones lagged
Uzbekistan	2019–2026	+1 676% (CAGR 51%)	+1 209%	— (no index)	signal works again — nominal soum, high inflation, treat with care*

Annualised return of the bank signal portfolio per market. Green = the signal beat its equal-weight basket; amber = on par; red = lagged. Windows differ (Russia from Jan 2023, US from 2018, the rest from 2019); Kazakhstan & Uzbekistan are nominal local currency in high-inflation frontier markets — read the magnitudes with care.

Live equity curves — Indonesia & Mexico

The two markets below are recomputed on every build from Yahoo prices and reported annual net income, so you can see the full path, not just the endpoint. Mexico is the textbook case — the earnings-yield tilt (solid) pulls clear of both the basket and the index; Indonesia shows the limit — four near-identical mega-banks, so the tilt tracks the basket.

The same return-on-investment ranking generalises across markets — the US and Mexico especially. In most markets the bank book beats the local index, and the cross-sectional tilt adds value where the universe is large and dispersed (Russia, US, Mexico, Kazakhstan, Uzbekistan). It does little where the universe is tiny and homogeneous (Indonesia’s four mega-banks), and it actively hurt in the Philippines — the signal is strong but not universal. Russia alone has the monthly-data advantage that lifts the result dramatically.

4. Portfolio construction

At each month-end: take the eligible universe (FVC book = non-financials, top-40 by market cap; Commodity book = the commodity producers), drop names with stale financials, rank by the signal, take the top ~10–15, equal weight. Hold to the next month-end. No transaction costs. Scores are unit-free ratios, so no currency conversion is needed.

5. Total return & benchmarks (dividends included)

Unlike the US/GC backtest above (price-only vs the S&P price index), the multi-market extension is total return: holdings reinvest dividends. Russia builds each price series from EI1 closes plus declared dividends (split days neutralised via share-count jumps) and is benchmarked against MCFTR, the MOEX gross total-return index (IMOEX price-return understates the market by the ~7%/yr dividend gap). Indonesia/Mexico use Yahoo adjusted close vs the ^JKSE / ^MXX price indices, with an equal-weight total-return basket as the fairer yardstick.

6. Single-stock efficacy test (does the signal time / rank?)

For every name we lag the signal one day (no look-ahead) and call it "invested" when the signal is positive. We compare the annualised return while invested to plain buy-&-hold and to the return while out. If the signal flags good moments, invested > buy-&-hold > out. Separately we measure whether it ranks companies — the Spearman correlation between signal and the subsequent 6-month return (the information coefficient, IC). Positive IC ⇒ higher-signal names go on to perform better; IC ≈ 0 ⇒ the signal is a filter, not a stock-picker.

FVC & Commodity-Upside portfolios — Russia / Indonesia / Mexico

Two cross-sectional strategies per market, backtested vs the local index with dividends reinvested (total return). Monthly rebalance, equal weight, no costs. RU uses the platform's authoritative point-in-time potential series (Strategy_Results_Grow.Potential_Curr for quality, Strategy_Results_Comm.Potential_min3y for commodity). ID/MX use a from-financials replica (FVC = getCreationValue port; Commodity-Upside = EV/EBITDA reversion) over the available data window. Generated 2026-06-08

Russia (RU) — vs MCFTR (TR)

Window 2015-01-01..2025-10-07. Authoritative platform potential series, currency RUB. Benchmark is MCFTR, the MOEX gross total-return index — a fair like-for-like vs the dividend-reinvested strategies (IMOEX price-return shown for reference).

strategy	total	CAGR	vol	Sharpe	maxDD	index CAGR	excess	beta	alpha
RU Growth-potential	+275.2%	+13.0%	16.1%	0.61	-57.6%	+13.2%	-0.2%	0.60	+3.4%
RU Commodity-Upside	+2501.0%	+35.2%	18.5%	1.22	-39.7%	+13.2%	+22.0%	0.67	+15.5%

RU Growth-potential — current book (as of 2025-10-07): VSEH (+231%), T (+206%), BELU (+130%), MGKL (+71%), MTSS (+30%), MGNT (+29%), VKCO (+10%), MDMG (-14%), ABRD (-25%), ASTR (-25%)

RU Commodity-Upside — current book (as of 2025-10-07): SMLT (+518%), PIKK (+107%), UGLD (+73%), NKNC (+12%), PLZL (+5%), GMKN (-4%), SELG (-5%), PHOR (-15%), FLOT (-16%), NLMK (-20%)

Indonesia (ID) — vs ^JKSE

Window 2024-09-01..2026-06-08. g=8.0, fair EV/EBITDA=7.0, currency IDR (from-financials replica).

Fair same-treatment benchmark — equal-weight total-return basket of the whole ID universe: CAGR -9.9% (the index ^JKSE is price-return, so part of the excess vs it is the dividend gap). Note the universe is small (selection is closer to a sector basket than a deep cross-section).

strategy	total	CAGR	vol	Sharpe	maxDD	index CAGR	excess	beta	alpha
ID FVC	-24.2%	-14.5%	14.3%	-1.09	-26.1%	-16.6%	+2.1%	0.31	-10.3%
ID Commodity-Upside	+5.5%	+3.1%	27.2%	0.25	-31.5%	-16.6%	+19.7%	0.82	+20.6%

ID FVC — current book (as of 2026-06-08): SMGR (+13%), KLBF (+8%), MIKA (+8%), PTBA (-5%), TLKM (-8%), ICBP (-15%), INDF (-20%), PGAS (-150%)

ID Commodity-Upside — current book (as of 2026-06-08): PGAS (+300%), SMGR (+77%), PTBA (+24%), ANTM (+6%)

Mexico (MX) — vs ^MXX

Window 2023-09-01..2026-06-08. g=8.0, fair EV/EBITDA=8.0, currency MXN (from-financials replica).

Fair same-treatment benchmark — equal-weight total-return basket of the whole MX universe: CAGR +10.9% (the index ^MXX is price-return, so part of the excess vs it is the dividend gap). Note the universe is small (selection is closer to a sector basket than a deep cross-section).

strategy	total	CAGR	vol	Sharpe	maxDD	index CAGR	excess	beta	alpha
MX FVC	+14.3%	+5.0%	13.8%	0.42	-15.9%	+8.2%	-3.3%	0.59	+0.4%
MX Commodity-Upside	+112.9%	+31.3%	26.2%	1.18	-32.0%	+8.2%	+23.1%	1.09	+20.8%

MX FVC — current book (as of 2026-06-08): FEMSAUBD (+88%), GMEXICOB (+70%), AC (+69%), KOFUBL (-3%), ASURB (-3%), WALMEX (-4%), AMXB (-55%), GCARSOA1 (-56%), LIVEPOLC-1 (-124%)

MX Commodity-Upside — current book (as of 2026-06-08): ORBIA (+300%), CEMEXCPO (-95%), GMEXICOB (-95%)

Methodology & caveats

Total return / dividends. RU holdings reinvest declared dividends (Company_Dividend) into the price series; split days are neutralised via share-count jumps. RU is benchmarked against MCFTR (MOEX gross total-return), the correct like-for-like. ID/MX holdings use Yahoo adjusted close (dividends already reinvested); their indices (^JKSE/^MXX) are price-return, so the equal-weight TR basket is the fairer same-treatment yardstick.
RU Commodity-Upside is survivorship- and liquidity-laden. The platform's 31-name RU commodity set includes thinly-traded micro-caps (FESH, AQUA, SMLT, ROLO, UGLD…) that drive a large part of the +22% excess; a tradable, liquidity-filtered version would be materially lower (cf. the global GC backtest, where debiasing cut CAGR ~3pp). Treat the absolute number as an upper bound, the vs-index direction as the signal.
“RU Growth-potential” ≈ the total-return market (−0.2% excess) but at beta 0.60 → positive alpha (+3.4%): a quality filter, not a return amplifier. This series ranks by Strategy_Results_Grow.Potential_Curr (upside), not FVC — a separate “RU FVC (actual)” book below ranks by real value-creation (where MGNT is excluded: FVC −65%).
ID/MX windows are short (data vintage: ID financials from 2024, MX from 2023) and the universes are small (~16 names; MX commodity = only 3), so these are low-power, closer to sector-basket vs index than a deep cross-section. Directionally: commodity-upside beat the index in both; FVC defended in the ID drawdown and modestly beat in MX.
Monthly rebalance, equal weight, no transaction costs. Scores are unit-free ratios so no FX is needed; returns are each market's local currency (RUB/IDR/MXN), compared to a local-currency index.

Signal efficacy — single-stock timing & ranking

Per stock, with the signal lagged one day (no look-ahead) and "invested" = signal ≥ 0 (model sees upside): annualised return while invested vs buy-&-hold vs while out. A signal that highlights good moments shows invested > buy&hold > out. Cross-sectional power = Spearman(mean signal, realised CAGR) and a pooled 6-month forward-return rank-IC. Total return (RU dividends reinvested + split-adjusted; ID/MX Yahoo adjusted close).

US Leaders (FVC/quality) (2015-01-01..2025-12-19, 47 stocks)

Timing — median edge (ann_in − buy&hold) +3.4%; stocks where invested beats buy&hold 51%; invested > out 57%. Avg annualised: invested +16.3% · buy&hold +14.5% · out +13.8% (invested only 28% of the time).

Timed sleeve (cash when signal off): CAGR +4.4% vs buy&hold +14.5%; max drawdown -31.2% vs -42.2%.

Ranking — does the signal prioritise better companies? Spearman(mean signal, CAGR) 0.28; pooled 6-month forward rank-IC 0.03 (127 months).

ticker	% time in	buy&hold	invested	out	edge	timed CAGR	b&h maxDD	timed maxDD
PYPL	43%	-12.2%	+47.7%	-41.1%	+60.0%	+18.5%	-25.8%	-12.2%
TSLA	24%	+33.3%	+86.9%	+19.8%	+53.5%	+16.2%	-73.6%	-45.2%
NFLX	34%	+27.2%	+67.0%	+10.8%	+39.7%	+18.8%	-75.9%	-41.6%
AVGO	37%	+34.1%	+55.8%	+23.0%	+21.7%	+17.6%	-48.7%	-33.3%
HD	22%	+12.8%	+26.7%	+9.1%	+13.9%	+5.3%	-38.0%	-17.0%
GOOG	36%	+23.3%	+35.6%	+17.0%	+12.3%	+11.5%	-44.6%	-24.1%
COST	28%	+19.1%	+28.3%	+15.6%	+9.2%	+7.4%	-31.4%	-19.6%
MSFT	34%	+22.9%	+31.2%	+18.8%	+8.3%	+9.6%	-37.1%	-20.5%
T	26%	+6.0%	+13.9%	+3.4%	+7.9%	+3.4%	-40.3%	-17.9%
JNJ	21%	+8.4%	+16.3%	+6.5%	+7.8%	+3.2%	-27.4%	-18.3%
NVDA	34%	+61.8%	+69.6%	+57.9%	+7.8%	+19.6%	-66.4%	-44.8%
META	41%	+19.2%	+26.2%	+14.5%	+7.0%	+10.1%	-76.7%	-38.0%
NKE	17%	+2.9%	-18.9%	+8.0%	-21.8%	-3.5%	-68.5%	-44.0%
AMGN	34%	-10.3%	—	-32.2%	—	+16.0%	-15.2%	-8.0%
BA	0%	+14.6%	—	+14.6%	—	+0.0%	-23.1%	+0.0%
HON	16%	+14.9%	—	+33.8%	—	-10.0%	-9.5%	-8.5%
INTC	0%	-12.7%	—	-12.7%	—	+0.0%	-23.4%	+0.0%
ORCL	2%	+84.5%	—	+75.8%	—	+6.1%	-9.9%	-1.5%

Global Commodities (potential) (2015-01-01..2026-06-05, 156 stocks)

Timing — median edge (ann_in − buy&hold) +4.4%; stocks where invested beats buy&hold 58%; invested > out 64%. Avg annualised: invested +20.3% · buy&hold +2.1% · out -1.5% (invested only 38% of the time).

Timed sleeve (cash when signal off): CAGR +4.2% vs buy&hold +2.1%; max drawdown -45.0% vs -73.2%.

Ranking — does the signal prioritise better companies? Spearman(mean signal, CAGR) 0.33; pooled 6-month forward rank-IC 0.06 (132 months).

ticker	% time in	buy&hold	invested	out	edge	timed CAGR	b&h maxDD	timed maxDD
CURLF	13%	-4.3%	+343.5%	-24.2%	+347.8%	+21.8%	-96.1%	-43.3%
GTBIF	9%	+1.5%	+334.1%	-12.3%	+332.7%	+14.3%	-87.7%	-30.9%
ALB	5%	+9.8%	+224.5%	+3.6%	+214.7%	+6.2%	-83.9%	-13.3%
YELL	15%	-22.6%	+144.5%	-37.3%	+167.1%	+14.8%	-91.5%	-32.8%
CRLBF	19%	-35.7%	+96.5%	-50.2%	+132.2%	+13.4%	-97.5%	-54.3%
CCJ	8%	+18.9%	+151.0%	+11.5%	+132.1%	+7.5%	-55.1%	-21.1%
PTSI	20%	+8.5%	+122.7%	-9.4%	+114.2%	+17.5%	-64.9%	-32.5%
TCNNF	19%	-2.8%	+88.7%	-16.7%	+91.6%	+12.7%	-94.0%	-47.7%
IIPR	15%	-1.1%	+86.4%	-11.7%	+87.4%	+10.0%	-78.6%	-19.7%
NEM	32%	+16.1%	+70.3%	-2.9%	+54.2%	+18.5%	-62.4%	-27.4%
CVI	49%	+6.9%	+57.0%	-26.4%	+50.1%	+24.9%	-78.1%	-35.5%
ESTE	43%	+1.0%	+48.5%	-24.5%	+47.4%	+18.5%	-93.5%	-50.8%
ODFL	2%	+12.8%	—	+11.8%	—	+1.1%	-68.1%	-4.0%
OROCF	0%	+7.7%	—	+7.7%	—	+0.0%	-52.6%	+0.0%
SNDL	0%	-15.9%	—	-15.9%	—	+0.0%	-54.3%	+0.0%
TELL	0%	-15.7%	—	-15.7%	—	+0.0%	-98.4%	+0.0%
TLRY	0%	-36.1%	—	-36.1%	—	+0.0%	-93.4%	+0.0%
TSP	1%	-0.8%	—	+9.9%	—	-9.6%	-70.8%	+0.0%

Oil & Gas E&P (potential) (2015-01-01..2026-06-05, 24 stocks)

Timing — median edge (ann_in − buy&hold) +2.1%; stocks where invested beats buy&hold 54%; invested > out 54%. Avg annualised: invested +7.4% · buy&hold +7.0% · out +10.3% (invested only 47% of the time).

Timed sleeve (cash when signal off): CAGR +3.1% vs buy&hold +7.0%; max drawdown -54.3% vs -78.5%.

Ranking — does the signal prioritise better companies? Spearman(mean signal, CAGR) 0.52; pooled 6-month forward rank-IC -0.02 (132 months).

ticker	% time in	buy&hold	invested	out	edge	timed CAGR	b&h maxDD	timed maxDD
PXD	37%	+6.2%	+30.8%	-6.0%	+24.5%	+10.4%	-73.0%	-28.0%
CRK	38%	+2.6%	+20.7%	-7.1%	+18.1%	+7.4%	-77.1%	-48.4%
WTI	56%	-4.1%	+9.8%	-19.3%	+13.9%	+5.4%	-87.3%	-79.5%
HES	12%	+8.6%	+19.5%	+7.2%	+10.9%	+2.2%	-60.1%	-32.3%
RRC	37%	+0.3%	+9.8%	-4.8%	+9.4%	+3.5%	-92.9%	-42.7%
NOG	65%	-5.4%	+3.4%	-19.9%	+8.9%	+2.2%	-97.1%	-80.4%
OXY	49%	+0.1%	+8.5%	-7.4%	+8.4%	+4.1%	-88.3%	-49.6%
EQT	43%	+4.0%	+11.5%	-1.2%	+7.4%	+4.8%	-83.5%	-42.9%
APA	64%	-2.8%	+3.7%	-13.6%	+6.5%	+2.4%	-93.5%	-63.3%
EOG	45%	+6.2%	+11.4%	+2.0%	+5.3%	+5.0%	-77.1%	-54.3%
SM	55%	+8.8%	+13.6%	+3.2%	+4.8%	+7.2%	-95.8%	-55.8%
DVN	44%	+5.6%	+7.8%	+3.9%	+2.2%	+3.4%	-88.5%	-59.2%
OVV	37%	+15.9%	+6.0%	+22.2%	-9.9%	+2.2%	-87.8%	-46.1%
MGY	74%	+9.4%	-1.4%	+46.7%	-10.9%	-1.1%	-63.7%	-62.4%
DEN	54%	+17.7%	+5.8%	+33.5%	-11.9%	+3.1%	-35.2%	-30.2%
CNQ	58%	+36.9%	+22.9%	+58.7%	-13.9%	+12.7%	-53.7%	-43.8%
MTDR	64%	+21.5%	+1.5%	+66.3%	-20.0%	+0.9%	-83.1%	-57.6%
MUR	42%	+5.4%	-21.2%	+30.2%	-26.6%	-9.5%	-75.4%	-80.9%

RU Grow/Quality potential (2015-01-01..2025-10-07, 45 stocks)

Timing — median edge (ann_in − buy&hold) +3.5%; stocks where invested beats buy&hold 42%; invested > out 56%. Avg annualised: invested +6.7% · buy&hold -0.9% · out +0.2% (invested only 28% of the time).

Timed sleeve (cash when signal off): CAGR +0.8% vs buy&hold -0.9%; max drawdown -31.5% vs -57.2%.

Ranking — does the signal prioritise better companies? Spearman(mean signal, CAGR) 0.21; pooled 6-month forward rank-IC 0.04 (125 months).

ticker	% time in	buy&hold	invested	out	edge	timed CAGR	b&h maxDD	timed maxDD
ABRD	34%	+2.5%	+60.2%	-18.3%	+57.8%	+17.2%	-56.0%	-29.1%
OZON	31%	+9.6%	+57.4%	-6.9%	+47.8%	+15.1%	-27.7%	-8.6%
QIWI	23%	-11.8%	+35.1%	-22.3%	+46.9%	+7.2%	-88.4%	-37.2%
BELU	42%	+20.4%	+64.7%	-4.4%	+44.3%	+23.5%	-71.7%	-41.7%
OKEY	28%	+5.0%	+49.1%	-8.3%	+44.1%	+11.8%	-47.4%	-33.0%
DSKY	20%	+0.8%	+28.6%	-5.3%	+27.8%	+5.3%	-66.9%	-14.1%
POSI	13%	-11.6%	+15.2%	-15.1%	+26.8%	+1.9%	-66.6%	-10.6%
SOFL	34%	-6.2%	+19.4%	-17.5%	+25.7%	+6.3%	-25.5%	-13.7%
VEON	12%	-12.9%	+11.1%	-15.7%	+24.0%	+1.3%	-90.4%	-21.4%
ASTR	65%	-33.8%	-16.2%	-57.7%	+17.7%	-10.9%	-46.6%	-29.3%
X5	41%	+12.2%	+29.4%	+1.8%	+17.2%	+11.0%	-73.0%	-50.8%
LENT	30%	-1.6%	+10.2%	-6.3%	+11.8%	+3.0%	-77.4%	-26.0%
FIXR	0%	-10.0%	—	-10.0%	—	+0.0%	-62.5%	+0.0%
HNFG	0%	-21.0%	—	-21.0%	—	+0.0%	-27.9%	+0.0%
PIKK	0%	+14.1%	—	+14.1%	—	+0.0%	-67.0%	+0.0%
SVCB	0%	-32.2%	—	-32.2%	—	+0.0%	-35.3%	+0.0%
UWGN	0%	+40.9%	—	+40.9%	—	+0.0%	-52.1%	+0.0%
WUSH	0%	-40.8%	—	-40.8%	—	+0.0%	-74.3%	+0.0%

RU Commodity potential (2015-01-01..2025-10-07, 30 stocks)

Timing — median edge (ann_in − buy&hold) +11.4%; stocks where invested beats buy&hold 77%; invested > out 82%. Avg annualised: invested +28.7% · buy&hold +9.3% · out -1.2% (invested only 50% of the time).

Timed sleeve (cash when signal off): CAGR +9.7% vs buy&hold +9.3%; max drawdown -48.3% vs -62.3%.

Ranking — does the signal prioritise better companies? Spearman(mean signal, CAGR) -0.07; pooled 6-month forward rank-IC 0.12 (125 months).

ticker	% time in	buy&hold	invested	out	edge	timed CAGR	b&h maxDD	timed maxDD
KAZT	21%	+19.5%	+163.6%	-3.0%	+144.1%	+22.4%	-54.6%	-45.6%
GMKN	46%	+44.6%	+107.6%	+6.1%	+63.0%	+40.1%	-44.0%	-28.0%
RASP	26%	+4.6%	+48.0%	-7.4%	+43.4%	+10.7%	-65.5%	-18.5%
AKRN	34%	+24.1%	+63.0%	+7.5%	+38.9%	+18.4%	-50.0%	-39.4%
RUAL	42%	+5.7%	+40.8%	-14.3%	+35.1%	+15.6%	-63.4%	-38.7%
PLZL	71%	+24.2%	+54.6%	-27.2%	+30.4%	+36.2%	-69.0%	-29.7%
TRMK	46%	+16.9%	+43.6%	-2.1%	+26.7%	+18.2%	-66.8%	-44.7%
PHOR	57%	+18.4%	+44.3%	-8.8%	+25.8%	+23.2%	-33.6%	-24.1%
GCHE	56%	+11.5%	+37.2%	-13.9%	+25.7%	+19.2%	-40.6%	-30.8%
FESH	62%	+36.2%	+58.7%	+6.4%	+22.5%	+33.0%	-65.5%	-48.6%
GAZP	41%	+0.4%	+15.5%	-9.0%	+15.0%	+6.1%	-65.1%	-54.8%
MTLR	48%	+0.8%	+15.7%	-11.4%	+14.9%	+7.3%	-80.4%	-60.7%
PIKK	42%	+7.0%	+5.6%	+8.1%	-1.4%	+2.3%	-76.0%	-62.7%
AQUA	59%	+24.6%	+19.7%	+32.0%	-4.9%	+11.3%	-46.0%	-46.2%
POLY	78%	-0.9%	-8.9%	+33.1%	-8.0%	-7.0%	-88.7%	-90.4%
ROLO	42%	+1.7%	-17.7%	+18.6%	-19.4%	-7.9%	-75.3%	-55.2%
NVTK	17%	+11.3%	-12.5%	+17.1%	-23.9%	-2.3%	-58.9%	-57.7%
SGZH	1%	-32.1%	—	-31.5%	—	-1.5%	-85.4%	-7.4%

ID FVC (2024-09-01..2026-06-08, 8 stocks)

Timing — median edge (ann_in − buy&hold) +0.0%; stocks where invested beats buy&hold 12%; invested > out 50%. Avg annualised: invested -38.7% · buy&hold -25.4% · out -12.3% (invested only 35% of the time).

Timed sleeve (cash when signal off): CAGR -19.5% vs buy&hold -25.4%; max drawdown -20.8% vs -37.7%.

Ranking — does the signal prioritise better companies? Spearman(mean signal, CAGR) -0.64; pooled 6-month forward rank-IC -0.80 (5 months).

ticker	% time in	buy&hold	invested	out	edge	timed CAGR	b&h maxDD	timed maxDD
ICBP	28%	-40.2%	-26.7%	-44.8%	+13.5%	-8.4%	-40.0%	-16.5%
KLBF	100%	-56.9%	-56.9%	—	+0.0%	-56.9%	-55.2%	-55.2%
MIKA	100%	-43.2%	-43.2%	—	+0.0%	-43.2%	-43.1%	-43.1%
INDF	28%	-25.1%	-27.8%	-24.0%	-2.7%	-8.8%	-30.8%	-15.2%
PGAS	0%	+1.7%	—	+1.7%	—	+0.0%	-33.1%	+0.0%
PTBA	0%	+5.9%	—	+5.9%	—	+0.0%	-18.6%	+0.0%
SMGR	20%	-44.6%	—	-11.6%	—	-38.9%	-50.6%	-36.6%
TLKM	0%	-0.8%	—	-0.8%	—	+0.0%	-29.9%	+0.0%
MIKA	100%	-43.2%	-43.2%	—	+0.0%	-43.2%	-43.1%	-43.1%
INDF	28%	-25.1%	-27.8%	-24.0%	-2.7%	-8.8%	-30.8%	-15.2%
PGAS	0%	+1.7%	—	+1.7%	—	+0.0%	-33.1%	+0.0%
PTBA	0%	+5.9%	—	+5.9%	—	+0.0%	-18.6%	+0.0%
SMGR	20%	-44.6%	—	-11.6%	—	-38.9%	-50.6%	-36.6%
TLKM	0%	-0.8%	—	-0.8%	—	+0.0%	-29.9%	+0.0%

ID Commodity upside (2024-09-01..2026-06-08, 5 stocks)

Timing — median edge (ann_in − buy&hold) +0.9%; stocks where invested beats buy&hold 40%; invested > out 100%. Avg annualised: invested +10.8% · buy&hold -10.0% · out -26.6% (invested only 60% of the time).

Timed sleeve (cash when signal off): CAGR +4.2% vs buy&hold -10.0%; max drawdown -23.0% vs -51.2%.

Ranking — does the signal prioritise better companies? Spearman(mean signal, CAGR) 0.90; pooled 6-month forward rank-IC 0.62 (10 months).

ticker	% time in	buy&hold	invested	out	edge	timed CAGR	b&h maxDD	timed maxDD
ANTM	50%	+52.7%	+70.1%	+36.9%	+17.4%	+30.6%	-45.2%	-18.3%
SMGR	50%	-41.3%	-39.5%	-42.9%	+1.7%	-22.2%	-64.2%	-38.0%
PGAS	100%	+16.0%	+16.0%	—	+0.0%	+16.0%	-33.1%	-33.1%
PTBA	100%	-3.3%	-3.3%	—	+0.0%	-3.3%	-25.8%	-25.8%
TPIA	0%	-73.9%	—	-73.9%	—	+0.0%	-87.6%	+0.0%
ANTM	50%	+52.7%	+70.1%	+36.9%	+17.4%	+30.6%	-45.2%	-18.3%
SMGR	50%	-41.3%	-39.5%	-42.9%	+1.7%	-22.2%	-64.2%	-38.0%
PGAS	100%	+16.0%	+16.0%	—	+0.0%	+16.0%	-33.1%	-33.1%
PTBA	100%	-3.3%	-3.3%	—	+0.0%	-3.3%	-25.8%	-25.8%
TPIA	0%	-73.9%	—	-73.9%	—	+0.0%	-87.6%	+0.0%

MX FVC (2023-09-01..2026-06-08, 9 stocks)

Timing — median edge (ann_in − buy&hold) +9.6%; stocks where invested beats buy&hold 56%; invested > out 57%. Avg annualised: invested +17.1% · buy&hold +15.2% · out +10.5% (invested only 56% of the time).

Timed sleeve (cash when signal off): CAGR +11.2% vs buy&hold +15.2%; max drawdown -16.7% vs -22.3%.

Ranking — does the signal prioritise better companies? Spearman(mean signal, CAGR) -0.02; pooled 6-month forward rank-IC 0.12 (16 months).

ticker	% time in	buy&hold	invested	out	edge	timed CAGR	b&h maxDD	timed maxDD
AMXB	23%	+21.8%	+55.8%	+13.0%	+34.1%	+10.9%	-18.4%	-6.6%
GMEXICOB	79%	+86.3%	+103.2%	+35.2%	+16.9%	+74.6%	-21.1%	-21.1%
ASURB	90%	+1.2%	+13.3%	—	+12.1%	+12.0%	-22.7%	-12.8%
WALMEX	49%	-6.6%	+4.6%	-16.5%	+11.3%	+2.3%	-22.9%	-15.5%
LIVEPOLC-1	68%	-11.4%	-1.8%	-28.8%	+9.6%	-1.3%	-36.6%	-23.0%
FEMSAUBD	100%	+15.4%	+15.4%	—	+0.0%	+15.4%	-23.3%	-23.3%
AC	33%	+13.9%	+8.2%	+16.8%	-5.7%	+2.6%	-16.6%	-11.3%
GCARSOA1	35%	+9.3%	+3.4%	+12.7%	-5.9%	+1.2%	-21.8%	-20.8%
KOFUBL	28%	+7.1%	-47.9%	+41.5%	-55.0%	-16.6%	-17.3%	-15.9%
WALMEX	49%	-6.6%	+4.6%	-16.5%	+11.3%	+2.3%	-22.9%	-15.5%
LIVEPOLC-1	68%	-11.4%	-1.8%	-28.8%	+9.6%	-1.3%	-36.6%	-23.0%
FEMSAUBD	100%	+15.4%	+15.4%	—	+0.0%	+15.4%	-23.3%	-23.3%
AC	33%	+13.9%	+8.2%	+16.8%	-5.7%	+2.6%	-16.6%	-11.3%
GCARSOA1	35%	+9.3%	+3.4%	+12.7%	-5.9%	+1.2%	-21.8%	-20.8%
KOFUBL	28%	+7.1%	-47.9%	+41.5%	-55.0%	-16.6%	-17.3%	-15.9%

MX Commodity upside (2023-09-01..2026-06-08, 3 stocks)

Timing — median edge (ann_in − buy&hold) +31.9%; stocks where invested beats buy&hold 67%; invested > out 100%. Avg annualised: invested +53.6% · buy&hold +23.2% · out +18.4% (invested only 30% of the time).

Timed sleeve (cash when signal off): CAGR +15.3% vs buy&hold +23.2%; max drawdown -12.1% vs -39.2%.

Ranking — does the signal prioritise better companies? Spearman(mean signal, CAGR) 0.50; pooled 6-month forward rank-IC n/a (too few names) (0 months).

ticker	% time in	buy&hold	invested	out	edge	timed CAGR	b&h maxDD	timed maxDD
ORBIA	16%	-10.5%	+51.2%	-19.2%	+61.7%	+7.0%	-70.5%	-18.2%
GMEXICOB	74%	+54.0%	+56.0%	+48.3%	+2.1%	+39.0%	-21.1%	-17.9%
CEMEXCPO	0%	+26.2%	—	+26.2%	—	+0.0%	-26.1%	+0.0%
ORBIA	16%	-10.5%	+51.2%	-19.2%	+61.7%	+7.0%	-70.5%	-18.2%
GMEXICOB	74%	+54.0%	+56.0%	+48.3%	+2.1%	+39.0%	-21.1%	-17.9%
CEMEXCPO	0%	+26.2%	—	+26.2%	—	+0.0%	-26.1%	+0.0%

Reading it

edge > 0 and invested > out ⇒ the signal genuinely times that stock — the days it says hold are better than average, the days it says avoid are worse.
Timed sleeve maxDD < buy&hold maxDD with similar CAGR ⇒ the signal earns its keep as a risk reducer even when it doesn't add raw return.
Positive rank-IC / Spearman ⇒ across names, higher signal really does sort to higher returns (it prioritises better companies). ~0 ⇒ it's a quality/cheapness filter, not a return-ranking signal.
RU has full history (2015+); ID/MX are short (2023/24+) and low-power — directional only.

RU FVC — backtest on the actual value-creation (same as company pages)

This ranks by the real FVC value-creation shown on /company/RU_* pages (compute_value_creation over RU_*_raw_periods.json, g=6.5, balance-method cash term, cap from EI1 ClosePriceCap), point-in-time, top-12 EW, monthly, dividends reinvested vs MCFTR. This is NOT the same as the “RU Growth-potential” book above, which ranks by Strategy_Results_Grow.Potential_Curr (the growing-strategy upside). The two diverge sharply for debt-funded dividend payers: e.g. Magnit (MGNT) has +30% growth-potential but a −65% FVC — its net debt rose ~797→1097 bn ₽, so the balance-method cash term is negative — so MGNT is correctly excluded from this FVC book. RU raw_periods start ~2018; window 2019-01-01..2026-06-08.

strategy	total	CAGR	vol	Sharpe	maxDD	MCFTR (TR) CAGR	excess	beta	alpha
RU FVC — actual value-creation	+108.2%	+10.3%	24.8%	0.40	-57.2%	+8.7%	+1.6%	0.77	+3.7%

⚠ Survivorship caveat. The data plane holds only the platform's current coverage, so the universe is today's 11 names held back to 2019-01-01. An equal-weight buy-&-hold of all 11 returned +15.0% CAGR vs MCFTR (TR) +8.7% — a +6.3%/yr universe edge unrelated to FVC. The FVC book made +10.3%, i.e. it underperformed naive equal-weight by -4.7%/yr ⇒ the actual FVC adds no return-ranking edge here — a quality filter, same as the efficacy IC≈0.

Current book 2026-06-08: OZON (+48%), MTSS (+44%), YDEX (+43%), X5 (+14%)

US Leaders — backtest on the actual dashboard FVC

Selection ranks by the same FVC shown on /region/us (compute_value_creation over SEC US_*_raw_periods.json, balance-method cash term, cap = Yahoo shares × price), point-in-time, top-15 EW, monthly, dividends reinvested. Differs from the original /strategies US Leaders (EI1 getCreationValue port, restricted to S&P 500 ∩ top-50 ∩ IsEbitdaBest ∩ FVC≥14). SEC data starts ~2018; window 2019-01-01..2026-06-08.

strategy	total	CAGR	vol	Sharpe	maxDD	S&P 500 TR (SPY) CAGR	excess	beta	alpha
US Leaders — actual (dashboard) FVC	+451.5%	+25.7%	19.2%	1.10	-26.2%	+17.5%	+8.3%	0.97	+6.5%

⚠ Survivorship caveat. The data plane holds only the platform's current coverage, so the universe is today's 49 names held back to 2019-01-01. An equal-weight buy-&-hold of all 49 returned +29.4% CAGR vs S&P 500 TR (SPY) +17.5% — a +11.9%/yr universe edge unrelated to FVC. The FVC book made +25.7%, i.e. it underperformed naive equal-weight by -3.7%/yr ⇒ the actual FVC adds no return-ranking edge here — a quality filter, same as the efficacy IC≈0.

Current book 2026-06-08: BA (+103%), ABBV (+85%), MU (+57%), APH (+33%), PEP (+23%), LLY (+22%), ADI (+21%), NFLX (+20%), ORCL (+19%), IBM (+17%), MSFT (+17%), RTX (+16%), JNJ (+16%), GE (+15%), NVDA (+14%)