Family Summaries: Per-Family Findings and Recommendations

1. Core

Family thesis. A market-cap-weighted top-N crypto basket, expressing the simplest possible "I want crypto exposure" allocation. Three variants by size: concentrated, balanced, and broad market beta.

Current-state metrics (4 indices, as of 2026-07-24): Average Sharpe2y -0.473 (OCI10 -0.018, close to BTC's own -0.027 reference; OCI6 -0.085; OCI20 -0.566; OCI50 -1.224, the family's laggard). Average max drawdown around -63.5 percent. Average correlation to BTC 0.87, unchanged from the prior read. Three of the four members (OCI10, OCI20, OCI50) carried degraded coverage (95-96 percent) following the 2026-06-30 TON delisting; a 2026-07-23 stale-tail fix in the pipeline restored their full simulation windows, and TON's history has since been made whole by a verified continuation splice to its successor ticker GRAM; the composition rename itself is pending an index-committee decision (Index Continuity Through Constituent Migrations: Delistings, Rebrands, and the Stale-Tail Problem).

Family finding. Core is still the cleanest illustration of why a market-cap-weighted top-N crypto basket loses to BTC in a sideways tape: every alt-cycle promise breaks down and BTC's market dominance creeps higher. The concentrated variant remains the most BTC-like in shape, and is now close enough to BTC's own -0.027 Sharpe to call the two indistinguishable over this window.

Recommended action. Keep the family. The "crypto index fund" brand expectation requires a Core. Update descriptions to be explicit about the 2-year regime fingerprint. Defer construction changes until 2027 to see whether the regime shifts.

2. Strategy

Family thesis. Factor and styled crypto baskets: low-volatility, smart-beta, barbell, blue-chip, risk-parity, quality, defensive growth, momentum, staking, DeFi income, yield income, reserve. Each expresses a TradFi factor concept translated into the crypto universe.

Current-state metrics (13 indices, as of 2026-07-24): Average Sharpe2y -0.402, still the widest best-to-worst dispersion in the book. The low-volatility expression remains the lone positive contributor (+0.086) and the closest thing the family has to a working factor, though its inverse-volatility candidacy still carries two open items: the TON-to-GRAM composition rename pending an index-committee decision (Index Continuity Through Constituent Migrations: Delistings, Rebrands, and the Stale-Tail Problem), and a concentration question on the computed vector's largest position (see The OLTA Rebalancing Calendar: Effective Dates, Announcements, and Drift Monitoring for the proposed adoption sequencing). The Risk Parity variant sits at -0.170 — better than most of the family; its computed-weight pipeline now proposes an equal-risk-contribution target vector, whose adoption is conditional on a pre-registered comparison gate on the proposed near-term calendar. Average max drawdown holds around -60 percent.

Family finding. Most factor expressions translated from TradFi into the crypto universe still did not deliver on the factor premium over this window. The low-volatility expression, conceptually the closest to a pure factor implementation, remains the sole positive Sharpe contributor. Three Strategy members carrying the family's worst readings were formally retired from the live catalog on 2026-05-24 for failed thesis or duplicate exposure (see Caveats); their history is retained here for continuity but they are no longer live products, and the family's active roster is smaller and somewhat better-behaved than the raw average suggests.

Recommended action. Consider migrating the Strategy family to a "factor crypto" sub-family with one or two Live names and the rest in Second Wave. The construction logic should be revisited (consider strategic-momentum or strategic-volTarget overlays) before re-running the next analysis cycle. The low-volatility and risk-parity members' computed-weight adoptions, should their respective gates clear, are the nearer-term lever and should be evaluated before any broader redesign.

3. Sector

Family thesis. Concentrated thematic exposure to a single crypto sector. Each basket holds 4 to 10 sector-specific tokens.

Current-state metrics (16 indices, as of 2026-07-24): Average Sharpe2y -0.679, with OLM6 the laggard at -1.005. This average excludes a seventeenth Sector member, ODER8, whose simulation window is only 5 days (64 percent coverage, following a 2026-07 Binance listing) and whose resulting headline Sharpe of 11.924 is a short-window statistical artifact, not a comparable family result; it should not be cited as a Sector performance figure until its window extends past 300 days. Average max drawdown remains around -84 percent. Average correlation to BTC holds near 0.72; average beta to BTC around 1.41.

Family finding. Every comparable Sector basket is still negative on Sharpe over the 2-year window. The sector narratives that drove construction (DeFi, L2 scaling, Restaking, ZK proofs, LSTs, oracles, derivatives) continue to fail to translate into token-price performance. The high-correlation, high-beta nature of these baskets means they fail in both bull and bear cases relative to BTC. Three Sector members were formally retired from the live catalog on 2026-05-24 (see Caveats); their history remains in the dataset above for continuity.

Recommended action. Downgrade the family to Watchlist (already done in the live app). Update each index description to explicitly cite the 2-year drawdown range. Defer construction changes until 2027.

4. RWA (Real-World Assets)

Family thesis. Cross-over baskets between crypto and traditional finance: tokenised commodities, treasuries, infrastructure exposure, private credit, gold, broad RWA infrastructure rails. Each addresses a TradFi asset class that has begun to settle on-chain in 2026.

Current-state metrics (8 indices, as of 2026-07-24): Average Sharpe2y -0.147 — a reversal from the window-inflated positive reading of this paper's original edition. OTC5 is now the strongest member (+0.509, 118-day window); OTG2 is the weakest (-1.110). All eight members still carry short windows (104 to 467 days); OTR6, the family's only Live member and the one with the longest history (467 days), sits at -0.455. Average max drawdown around -34 percent.

Family finding. The window-dependent Sharpe inflation flagged in the original edition of this paper has played out as expected: as the post-launch RWA baskets accumulated additional months of history, the short-window premium faded and the family average crossed into negative territory. This confirms the original caution rather than contradicting it — the "recent listings caught the post-launch narrative bid" reading was correct, and the bid has since faded as the sample grew. No RWA member has yet reached the 365-day mark this paper originally set as the threshold for a Live promotion decision.

Recommended action. Continue to wait; the family's negative average confirms this is not yet a promotion decision, and accelerating it would repeat the inflation error rather than correct for it. Re-evaluate at 2027-Q1 as originally planned, once OTR6 and its peers clear the 730-day mark.

5. Equities

Family thesis. Tokenised US equity baskets, routed through Dinari for single stocks and Backed Finance for ETF coverage, settling on Base. Coverage spans mega-cap tech, semiconductors, AI software, AI infrastructure, healthcare, energy, China tech, industrials and defense, consumer, broad ETFs, commodity ETFs, crypto equities, crypto operators, frontier (quantum), and leveraged ETFs.

Current-state metrics (15 indices, as of 2026-07-24): Average Sharpe2y +0.751, still the strongest average of any family and the only one clearly positive; 12 of 15 members are Sharpe-positive. OSEM7 (2.090) and OCOPS3 (2.067, flagged below on turnover) lead the family; OAH7 — the newest addition, an after-hours Mag7 successor — sits close behind at 1.043 with full coverage. OCPT5 (-0.364) is the laggard. Average max drawdown around -45 percent. Average correlation to BTC has drifted up to roughly 0.30 — no longer "effectively zero" as this paper's original edition read it — pulled by newer, more crypto-adjacent entrants (OCPT5 at 0.67, OCOPS3 at 0.32) alongside the traditional near-zero members.

Family finding. Equities continues to dominate the multi-year leaderboard, and the construction thesis — real underlying companies with cash flows, tokenised 1:1, correlation to BTC near zero — still holds for the majority of the family. It no longer holds uniformly: as the family has grown to include crypto-adjacent names, the family-average correlation to BTC has risen from the original edition's near-zero read to roughly 0.30, and three members now carry negative Sharpe. OCOPS3, despite the family's second-best Sharpe, is also the book's single highest-turnover index (roughly 76 percent annualized, 17 percent per quarterly event) and is flagged for a cadence review at the next quarterly cycle. OAH7 has no rebalancing ledger yet — the engine does not currently simulate its after-hours cadence — so its turnover and drift profile cannot be assessed alongside its peers.

Recommended action. Keep the family's construction as validated by the multi-year backtest. Consider adding a Healthcare expansion to capture the GLP-1 cycle more thoroughly, and a Financial Services basket to capture a grouping that is conspicuously absent. Track OCOPS3's turnover question separately from the family-level read, and prioritize an after-hours rebalancing capability so OAH7 is covered by the same governance as the rest of the family.

6. Diversified

Family thesis. Cross-asset baskets that mix crypto, tokenised US equity, gold, and (in one case) tokenised treasuries. Each is designed to deliver Sharpe above 1.0 over the 2-year backtest window. See The Diversified Family: Design, Backtest, and Thesis for the deep dive.

Current-state metrics (10 indices, as of 2026-07-24; 9 Live, 1 Watchlist): The family average Sharpe2y has fallen to +0.216, and the distribution is now bimodal rather than uniformly strong. The original five-member Live core remains solid — OSHARP6 1.058, OQUAL7 0.887, OCRLOW7 0.874, OAILEAD8 0.742, OBAL6 0.730 — but four newer entrants pull the average down hard: the "BTC-plus" trio (OBTCC60 0.018, OBTCR50 -0.247, OBTCA50 -0.456), a set of deliberately high-beta BTC-exposure baskets with correlation to BTC of 0.79 to 0.95 by design, and OHRP12 (-1.222), the family's and the book's second-worst Sharpe, whose computed-weight (HRP) adoption is a pending decision covered in The OLTA Rebalancing Calendar: Effective Dates, Announcements, and Drift Monitoring. Average max drawdown across the 9 Live members is now around -38 percent, up from -27 percent in the original edition, driven by the same four entrants; the original five-member core still runs in the -23 to -36 percent band. ORWY5 remains the family's sole Watchlist member, now at a 307-day window (-0.222 Sharpe) — closer to, but still short of, the year of joint history this paper originally called for. Average correlation to BTC across all 10 members is now approximately 0.78 — materially higher than the "effectively zero" read of the original edition. The multi-asset construction still dampens correlation relative to pure-crypto families (0.85 to 0.96), but the family's composition has shifted toward higher-beta members since March.

Family finding. The flagship thesis still holds for the original five-member core: Sharpe materially above the BTC reference at roughly half the drawdown, with correlation dampened relative to pure-crypto families even where it is no longer literally zero. The four newer entrants complicate a single family-level read. The BTC-plus trio is a deliberate high-beta sleeve, not a construction failure, and inherits BTC's flat-negative regime accordingly — its low Sharpe is the expected output of its own mandate. OHRP12 is a different case: it is the book's flagship example of a label-versus-weights mismatch — a Sharpe of -1.222 that is honestly computed but attached, today, to hand-set weights under an HRP-branded name. Its Sharpe number is not in question; whether that number should carry an HRP label is the open question, and it sits on the proposed near-term adoption calendar.

Recommended action. Present the family in two reads going forward rather than one blended average: the original balanced-diversification core, which remains the flagship, and a higher-beta BTC-plus / computed-weight sleeve whose correlation and drawdown profile should not be averaged into the flagship's headline numbers. Re-evaluate ORWY5 once a full year of joint history is available (currently 307 of 365 days). OHRP12's computed-weight adoption timeline is tracked in The OLTA Rebalancing Calendar: Effective Dates, Announcements, and Drift Monitoring; if the adoption clears its pre-registered gate and ships, its label will match its weights for the first time since launch, and if it fails the gate, the relabel path of From Labels to Computation: HRP, ERC and Inverse-Volatility Weights in Practice applies. Continue to consider the macro-defensive and broader-equity candidates proposed in the original edition.

7. Ecosystem

Family thesis. Concentrated exposure to a specific blockchain ecosystem. Each basket holds the native token plus 5 to 7 protocol tokens from the ecosystem.

Current-state metrics (4 indices, as of 2026-07-24): Average Sharpe2y -0.770 — now the worst average of any family in the book, edging out Sector. OARB6 is the laggard (-1.096); OCOS6 is the least-negative member (-0.634, though at only 85 percent coverage). Average max drawdown around -78 percent. Average correlation to BTC 0.77.

Family finding. Same dynamics as Sector, and now the single worst-performing family in the book on an average-Sharpe basis. A basket of correlated alt-L1 or alt-L2 tokens inherits the full BTC drawdown plus an idiosyncratic alpha-bleed term. The construction is a thematic bet on each ecosystem outperforming BTC, which simply did not happen in this window.

Recommended action. Keep the family as a structural exposure-rotation option (already in Watchlist on the live app). Update each description to explicitly note the drawdown range. Defer until 2027 to see whether any single ecosystem catches a structural bid. Consider adding an Ethereum-native ecosystem basket at the next cycle to broaden coverage.

8. ThematicBeta

Family thesis. Meme-token baskets and other thematic crypto exposure that is closer to "speculation" than "investment".

Current-state metrics (4 indices, as of 2026-07-24): Average Sharpe2y -0.606. OSMC4 is the family's worst member (-0.755, -95.2 percent drawdown) and was retired from the live catalog on 2026-05-24 as untradeable on a volatility-to-Sharpe basis (see Caveats); OMX8 is the least-negative active member (-0.472). Average max drawdown around -90 percent. Average correlation to BTC around 0.75.

Family finding. Predictably hammered, as before. These are speculative-tier instruments by construction, and the family's worst performer has since exited the live catalog on exactly that basis.

Recommended action. Preview-only across the family. Consider renaming the family to "Speculative" rather than "ThematicBeta" to set expectations cleanly.

9. Curio

Family thesis. Deliberately experimental baskets that exist as conversation pieces and product-design exercises rather than as institutional positions.

Current-state metrics (5 indices, as of 2026-07-24): Average Sharpe2y -0.418. OPIZZA3 is the least-negative reading (-0.133) but was retired from the live catalog on 2026-05-24 as a "marketing artifact" (see Caveats) and is retained here for dataset continuity only; among active members, OFOMO6 (-0.491) is the figure to cite. OREKT5 anchors the bottom (-0.734). Average max drawdown around -72 percent.

Family finding. Useful as a UX device. Not investable.

Recommended action. Keep the family for the conversation-piece value. Consider rotating out the worst-performing basket and replacing with a new conversation-piece concept. The Curio family is a useful UX device but is not investable.

10. Stellar and Hedera (new Preview families, partial coverage)

Family thesis. Two chain-native families entered Preview status since this paper's original edition: Stellar (OSTW7, OSTN6 — exposure to the Stellar ecosystem including tokenised treasury and money-market instruments) and Hedera (OHED4 — a single-chain ecosystem basket).

Current-state metrics, with a hard readability caveat:

• Stellar (2 indices): OSTW7 and OSTN6 both read Sharpe2y +0.456, max drawdown -74.4 percent, correlation to BTC 0.52 — but the two figures are identical only because coverage sits at 26 percent and 42 percent respectively, meaning the backtest is effectively pricing a single constituent (XLM) against each basket's full stated weight. Six of the seven distinct constituents across the two baskets have no Binance Spot feed. Separately, three constituents (BENJI, USTBL, EUTBL) show zero trading volume in any available feed; a NAV-accrual methodology decision (yield-bearing, non-rebasing treatment consistent with their real-world-asset-core role) is a prerequisite for an honest backtest of those legs, independent of the data-source gap.
• Hedera (1 index): OHED4 reads Sharpe2y +0.095, max drawdown -77.1 percent, correlation to BTC 0.56, at 85 percent coverage — two constituents (SAUCE, PACK) are missing from the feed.

Family finding. Both families are too early to carry a performance claim. The headline figures above are not comparable to any other family in this paper and should not be cited outside this caveat.

Recommended action. Hold both families at Preview until the missing constituent feeds are backfilled and, for Stellar specifically, the NAV-accrual methodology for the tokenised-treasury legs is settled. Re-evaluate once coverage clears 95 percent and a full quarter of clean joint history is available.

11. Cross-family ranking summary

12. Aggregated recommendations

1. Promote Diversified to flagship family status. Already reflected in the live app — with the caveat (§6) that the flagship read now applies to the original five-member core, not the family's four newer, higher-beta entrants.
2. Maintain the high-Sharpe Equities baskets as flagship Equities offerings.
3. Hold the Sector family at Watchlist until either the regime shifts or the construction is updated.
4. Hold the Ecosystem family at Watchlist for the same reasons; it is now the single worst-performing family in the book on average Sharpe.
5. Confirm the RWA family's original caution: the window-dependent inflation predicted in the March edition has since reversed to a negative average, and no promotion to Live is warranted yet.
6. The Strategy family needs a redesign cycle.
7. Consider adding new Diversified candidates at the next cycle, alongside the macro-defensive and broader-equity ideas proposed in the original edition.
8. Round out Equities thematic coverage with a Financial Services basket.
9. Hold Stellar and Hedera at Preview pending constituent-feed backfills and, for Stellar, an NAV-accrual methodology decision (§10).
10. Track OHRP12's pending computed-weight adoption via The OLTA Rebalancing Calendar: Effective Dates, Announcements, and Drift Monitoring rather than treating its current -1.222 Sharpe as a closed reading under its present label.
11. Note the rising average correlation to BTC in both Equities (0.30) and Diversified (0.78) as newer, more crypto-adjacent members join each family; the "effectively zero" framing of the original edition no longer describes either family precisely.

13. Caveats

• These recommendations are based on a single 2-year window for crypto data and a 5-year window for equities. Different windows would produce different recommendations.
• The visibility taxonomy (Live / Second Wave / Watchlist / Preview) is set with the Sharpe outperformance as a major input. Other factors (geographic restrictions, KYC status, regulator response) also matter and are not captured in this paper.
• The "consider redesign" recommendations are research suggestions, not commitments. Each would require a fresh backtest cycle against a refreshed dataset before production deployment.
• All Sharpe numbers use zero risk-free rate.
• Survivorship: current basket constituents are the constituents the basket would hold today. The bias is small but real.
• This update (2026-07-24) draws on an 84-ticker desk snapshot; the BTC 2-year reference used throughout is Sharpe -0.027, max drawdown -53.0 percent, through 2026-07-21 — a regime still flat-to-negative, and the figure superseding any BTC reference cited in the original edition.
• Ten catalogue members were formally retired from the live app on 2026-05-24: OL2G8, OZK8, OSMC4, ODS5, and OPIZZA3 for failed thesis; OBC8, OML4, OYI8, OIR6, and ODATA8 for duplicate exposure, with successors OCI10, OMM6, none, OTT4, and OOR6 respectively. Their historical performance remains in the family-level statistics above for dataset continuity, but they are no longer live products; where a retired name is materially relevant to a family's headline read (Strategy, Sector, ThematicBeta, Curio), this is flagged inline.

14. References

• Per-index performer tables and the specific visibility-change recommendations that informed this paper: methodology brief on request
• Engine source: lib/backtest.js
• Cross-references: Current-State Analysis: 74 Indices Across the 2-Year Window for full-library context; The Diversified Family: Design, Backtest, and Thesis for the Diversified design philosophy; Outperformance vs BTC: Counting the Winners for the BTC-relative ranking; The OLTA Rebalancing Calendar: Effective Dates, Announcements, and Drift Monitoring for the pending computed-weight adoption and rebalance-calendar program touching Core, Strategy, and Diversified members