1. Introduction
OPEN needs a healthy, inspiring chart backed by a methodology that the community can understand, reproduce, challenge, and maintain.
You can review OPEN v1.1 parameters and history here. OPEN has not rebalanced since Q1 2026 while this methodology work continues.
You are invited to peer review eight backtested methodologies for OPEN (the current v1.1 is included). Nothing in this post is financial advice.
Please critique the eligibility rules, variables, data sources, formulas, portfolio construction, assumptions, and conclusions.
Pending at least 14 days of community review and discussion, the vlSQUILL-OPEN community could choose to advance one of these methodologies or propose something entirely new for OPEN v1.2. This discussion is not itself a governance vote.
As a reminder, the live OPEN v1.1 methodology uses three weighted inputs:
- 60% TVL relative to FDV
- 20% year-over-year TVL growth
- 20% liquidity efficiency
It rebalances quarterly, holds up to 10 constituents, caps a holding at 40%, and funds every selected holding at a minimum 1%. Funding the floor proportionally reduces larger uncapped positions. The historical reconstruction tested here uses the frozen valuation and liquidity proxies described below, so please review that implementation as carefully as the headline results.
2. Summary
- Eight methods combined 10 scored or eligibility inputs, including fees, revenue growth, downside beta, TVL relative to valuation, TVL growth, price-to-sales, Ethereum liquidity inventory, price coverage, and peg continuity. Returns at the Sep 18, 2026 test end ranged from -52% to +29%.
- Jan 1, 2024 is the earliest healthy data common start date for the backtests. Earlier tests lacked complete quarter-end circulating market capitalization, prior-year revenue histories, Ethereum DEX inventory, and continuous rebalance prices for too many candidates. A Jul 2023 start left the P/S method with only three complete candidates. An Oct 2024 start provided healthier breadth but less than two years of returns.
- P/S + Revenue Growth Q-5 (+151%, +29%) finished ahead of ETH (+106%, +11%) at the Aug 22, 2025 ETH peak and Sep 18, 2026 test end, respectively. MKR carried approximately 25% to 34% target weights through Jun 2025, while SYRUP carried as much as approximately 22% during parts of 2025 and 2026. Their gains outweighed losses elsewhere.
- 5-token baskets beat their 10-token counterparts in this period. P/S + Revenue Growth Q-5 finished 35 return points ahead of Q-10. Fee + Revenue Growth S-5 finished 44 points ahead of S-10. Concentration helped here, but it also makes outcomes more dependent on a few assets.
- OPEN remains a high beta thematic asset for open-money believers. The category thesis comes first: either open money matures into an important financial system or it does not. A mechanical index may identify and concentrate the emerging winners sooner, but these assets and their data remain immature. This 33-month backtest compares rules; it cannot judge the theme’s long-term future.
- The analysis required 134,409 canonical observations, 19 frozen primary inputs, seven external source systems, and 177 raw provider responses. A sanitized review package is available below so anyone can inspect or rerun the work.
Figure 1
All lines begin at $100 on Jan 1, 2024. Execution costs are excluded; the complete cost treatment appears beneath the performance table.
3. General eligibility rules
For this backtest, a project entered the candidate universe only when all of the following were true:
- The protocol issued a fungible asset whose token, debt, or rebasing unit explicitly referenced one United States dollar. Variable-price vault shares did not qualify.
- Published minting, redemption, accounting, and risk rules applied across the asset. It could not merely represent one vault, pool, market, maturity, or investment position.
- The asset was transferable on Ethereum, accessible under public smart-contract rules, and usable outside its issuing application.
- An Ethereum-native ERC-20 directly governed the monetary system or captured value generated by it.
- Peg continuity, not collateral certification, governed failure. A stablecoin became ineligible at the next scheduled rebalance after four consecutive UTC closes more than 5% below its peg. It could return after 91 consecutive UTC closes within 1% of its peg. Fully exogenous collateral was neither scored nor required.
These rules deliberately finish the current Open Stablecoin Networks Index mandate. A future discussion could consider a broader Open Money or Open Finance universe, but that would require new eligibility tests and a new backtest.
Candidate eligibility summary
The table preserves the candidate record used by the backtest. Blank depeg cells mean the reviewed record contained no verified event meeting the displayed threshold; they do not prove uninterrupted parity.
| Project and governance token | Applicable stablecoin(s) | Known stablecoin depegs | Eligibility date | Ineligibility or delay |
|---|---|---|---|---|
| Aave (AAVE) | GHO | -5.7%; 08/02/23 to 01/25/24. -6.4%; 03/05/24 to 03/06/24. | 01/01/24 | |
| Abracadabra (SPELL) | MIM | -11.8%; 01/30/24 to 02/02/24. -92.0%; 06/08/26 to test end. | 01/01/24 | Ineligible 06/14/26; recovery pending. |
| Alchemix (ALCX) | alUSD | -8.4%; 04/19/24 to 08/05/24. -8.3%; 05/01/26 to test end. | 01/01/24 | |
| Angle Protocol (ANGLE) | agEUR, USDA | 01/01/24 | ||
| Curve Finance (CRV) | crvUSD | -3.2%; 06/26/25 to 06/27/25. | 01/01/24 | |
| Frax Finance (FRAX, formerly FXS) | legacy FRAX, frxUSD | -12.2%; 03/11/23 to 03/13/23. | 01/01/24 | |
| Inverse Finance (INV) | DOLA | 01/01/24 | ||
| Liquity (LQTY) | LUSD, BOLD | +6.8%; 09/18/22 to 12/13/22. | 01/01/24 | |
| Maker (MKR) | DAI | -11.4%; 03/11/23 to 03/13/23. | 01/01/24 | MKR ineligible after 06/02/25 migration. |
| Reserve Protocol (RSR) | RTokens, eUSD | 01/01/24 | ||
| Lybra Finance (LBR) | eUSD, peUSD | peUSD: -21%; 09/23/23 to 06/24 abandonment. | 04/01/24 | Abandoned 06/24; removed 07/01/24. |
| f(x) Protocol (FXN) | fxUSD | 07/01/24 | 270 fee-reporting days reached. | |
| Origin Protocol (OGN) | OUSD | -84%; 11/17/20 to 01/06/21. | 07/01/24 | First rebalance after 05/28/24 merger. |
| Prisma Finance (PRISMA) | mkUSD, ULTRA | 07/01/24 | Shutdown 12/18/24; removed 01/01/25. | |
| Ethena (ENA) | USDe, sUSDe | 10/01/24 | 270 fee-reporting days reached. | |
| Maple Finance (SYRUP) | syrupUSDC | 01/01/25 | First rebalance after 11/13/24 launch. | |
| Sky Ecosystem (SKY) | USDS | -5.1%; 10/01/24 to 10/06/24. | 07/01/25 | Quarter-end circulating Mcap first available. |
| Usual (USUAL) | USD0, USD0++ | USD0++: -11%; 01/10/25 to test end. | 07/01/25 | |
| ReSupply (RSUP) | reUSD | -44.5%; 06/24/26 to 06/26/26. | 01/01/26 | |
| USD.AI (CHIP) | USDai, sUSDai | 07/01/26 | First rebalance after 04/21/26 launch. | |
| Cap (CAP) | cUSD, stcUSD | Expected 10/01/26 | First rebalance after 06/26/26 TGE. | |
| Frankencoin (FPS) | ZCHF | Not eligible | Swiss-franc reference; Ethereum DEX liquidity absent. |
- Depeg magnitude is the largest observed move from the product’s reference. “Test end” means Sep 18, 2026.
- MIM’s Jan 30 to Feb 2, 2024 intraday event lacked the required daily-close sequence and did not trigger the peg rule.
- SKY replaced MKR under the dated migration policy. MKR remained eligible through Jun 1, 2025; SKY became eligible Jun 2, 2025 and could first enter on Jul 1, 2025.
- Event sources: Pharos history, MIM, crvUSD, USD0++, GHO, Lybra, and OUSD.
- Launch sources: Origin, Maple, Sky, USD.AI, and Cap.
4. Variables and community-verifiable data
OPEN has no permanent analyst, operating team, or recurring data budget. The methodology is only operationally open if an ordinary community member can retrieve, archive, and recalculate its inputs without a privileged account or expensive subscription.
| Variable | What it asks | Calculation | Methods | Approved source - acquisition |
|---|---|---|---|---|
| Fee scale | How much did customers pay? | Trailing 365-day fees; 270-day minimum | 5 | DefiLlama - daily API series |
| 90-day revenue growth | Is retained income accelerating? | Two 90-day windows, one year apart; 60 days each | 4 | DefiLlama - daily API rolling windows |
| Annual revenue growth | Is retained income growing? | Trailing year versus prior year | 2 | DefiLlama - daily API annual windows |
| Resilience | How does it fall with ETH? | Downside beta; 80 paired observations | 1 | DefiLlama + Coin Metrics - UTC daily-return join |
| TVL value | How much capital supports valuation? | TVL divided by fully diluted valuation | 1 | DefiLlama + CoinGecko - daily TVL and quarter-end FDV |
| TVL growth | Is deposited capital expanding? | Annual TVL change; negatives become zero | 1 | DefiLlama - 365-day API comparison |
| P/S | How much valuation per revenue dollar? | Circulating market capitalization divided by trailing 365-day DefiLlama protocol revenue | 2 | CoinGecko + DefiLlama - Pro quarter-end capitalization + daily revenue |
| Liquidity capacity | Can OPEN trade the token? | Ethereum token inventory divided by expected trade size | 3 | Allium + CoW - SQL inventory and live execution check |
| Observed prices | Is trading history sufficient? | 15 price days plus a rebalance mark | 8 | Allium - daily series and documented fills |
| Peg continuity | Did the product hold its dollar peg? | Eligibility screen described above | 8 | CoinGecko - daily closes and mechanical screen |
| Active wallets | How many addresses used the product? | Not scored | 0 | Token Terminal - not frozen |
Important implementation details:
- DefiLlama’s
origin-protocolparent series already aggregates Origin’s business activity. Child series are not added because that would count the same activity twice. - MKR/SKY, MPL/SYRUP, and OGV/OGN follow dated continuity policies.
- When a provider lacks a required observation, the token is excluded from that method at that rebalance. The analysis does not invent a value or turn an unknown observation into zero.
- Allium liquidity measures the constituent token held in covered Ethereum DEX contracts. It is a reproducible capacity proxy, not an executable quote or total pool TVL. Any live proposal still needs a current CoW or Reserve-zapper execution check.
- CoinGecko Pro supplies the quarter-end circulating-market-cap numerator for P/S. The frozen pull produced 194 observations from 21 calls with no failed assets.
- DefiLlama remains the revenue denominator because it provides reproducible daily histories across more of the candidate universe. Token Terminal changed ranks and weights at all 11 rebalances, but complete five-token membership changed at 3. Its incomplete coverage makes it unsuitable as OPEN’s sole production dependency.
If the community cannot independently rerun the basket, the methodology is not operationally open.
5. Testing analysis and results
Method suffixes describe cadence and maximum constituents. Q means quarterly, S means semiannual, and 5 or 10 is the maximum number of holdings.
| Method | Weighting rule | Latest tokens | To ETH peak, Aug 22, 2025 | To test end, Sep 18, 2026 | Annual basket replacement |
|---|---|---|---|---|---|
| P/S + Revenue Growth Q-5 | 60% lower P/S; 30% annual revenue growth; 10% liquidity | SKY 24% · SYRUP 20% · OGN 20% · CRV 18% · FXN 18% | +151% | +29% | 95% |
| Fee + Revenue Growth S-5 | 50% fee scale; 50% 90-day revenue growth; 25% cap | AAVE 25% · SYRUP 23% · SKY 21% · ENA 19% · FXN 14% | +115% | +25% | 88% |
| ETH | Passive benchmark | ETH 100% | +106% | +11% | 0% |
| P/S + Revenue Growth Q-10 | 60% lower P/S; 30% annual revenue growth; 10% liquidity | SKY 17% · SYRUP 14% · OGN 14% · CRV 13% · FXN 13% · AAVE 12% · ALCX 8.2% · ENA 4.0% · FRAX 3.3% · LQTY 2.9% | +113% | -5% | 77% |
| Fee + Revenue Growth S-10 | 50% fee scale; 50% 90-day revenue growth; 25% cap | AAVE 16% · SYRUP 14% · SKY 13% · ENA 12% · FXN 8.7% · CRV 8.7% · LQTY 8.1% · OGN 7.5% · FRAX 6.8% · USUAL 5.6% | +97% | -19% | 70% |
| TVL Value + Growth Q-10 (v1.1) | 60% TVL/value; 20% positive TVL growth; 20% liquidity; funded 1% to 40% bands | CHIP 22% · FXN 17% · SKY 11% · RSUP 11% · AAVE 10% · SYRUP 9.0% · FRAX 6.2% · LQTY 5.3% · ENA 4.5% · CRV 4.2% | +122% | -23% | 101% |
| Fee + Revenue Growth Q-10 | 50% fee scale; 50% 90-day revenue growth; 25% cap | AAVE 16% · SYRUP 14% · SKY 13% · ENA 12% · FXN 8.7% · CRV 8.7% · LQTY 8.1% · OGN 7.5% · FRAX 6.8% · USUAL 5.6% | +58% | -36% | 88% |
| Downside-Aware Fee Growth Q-10 | 33.3% fees; 33.3% revenue growth; 33.3% lower downside beta | SKY 14% · SYRUP 14% · AAVE 14% · ENA 12% · OGN 9.2% · INV 8.3% · LQTY 8.3% · CRV 7.0% · FXN 6.1% · FRAX 6.1% | +45% | -43% | 83% |
| Equal-Weight Fee Leaders Q-10 | 100% fee scale; equal weight | AAVE 10% · SKY 10% · ENA 10% · SYRUP 10% · CRV 10% · USUAL 10% · CHIP 10% · FRAX 10% · RSR 10% · OGN 10% | +50% | -52% | 95% |
Cost treatment: The backtest includes OPEN’s 0.75% annual management fee. It excludes entry, rebalance, exit, gas, spread, price-impact, route-specific CoW execution, and failed-transaction costs. On Aug 18, 2026, OPEN displayed a 0.30% mint fee, or $30 per $10,000, and a 0.75% annual fee, or $75 per $10,000 before performance. An earlier model charged 0.25% per gross traded dollar; removing that assumption raised endpoint returns by roughly 1 to 2 points and keeps the comparison focused on methodology performance rather than hypothetical trade size and routing.
Annual basket replacement is the average share of the portfolio exchanged each year. Replacing $20,000 of a $100,000 basket counts as 20%. Repeated rebalances can push the annual figure above 100%.
Exact scoring and weighting formulas
Each method first scores every eligible token, then converts the highest scores into portfolio weights. Let P(xᵢ) mean token i’s percentile rank among fully scoreable candidates from 0 to 1. Higher is better unless the factor is reversed.
| Method | Token score | Final weight |
|---|---|---|
| Equal-Weight Fee Leaders Q-10 | Rank by Fᵢ | Top 10 receive 1/N each |
| Fee + Revenue Growth Q-10 | Sᵢ = 0.50P(Fᵢ) + 0.50P(G⁹⁰ᵢ) | Score-proportional; 25% cap |
| Fee + Revenue Growth S-10 | Sᵢ = 0.50P(Fᵢ) + 0.50P(G⁹⁰ᵢ) | Same; semiannual |
| Fee + Revenue Growth S-5 | Sᵢ = 0.50P(Fᵢ) + 0.50P(G⁹⁰ᵢ) | Top 5; 25% cap |
| Downside-Aware Fee Growth Q-10 | Sᵢ = ⅓P(Fᵢ) + ⅓P(G⁹⁰ᵢ) + ⅓[1 − P(βᵢ)] | Score-proportional; 25% cap |
| TVL Value + Growth Q-10 (v1.1) | Sᵢ = 0.60(Vᵢ / ΣV) + 0.20[max(Gᵀⱽᴸᵢ, 0) / Σmax(Gᵀⱽᴸ, 0)] + 0.20(Lᵢ / ΣL) | Fund every selected token; 1% to 40% bands |
| P/S + Revenue Growth Q-10 | Sᵢ = 0.60[1 − P(PSᵢ)] + 0.30P(G³⁶⁵ᵢ) + 0.10P(Lᵢ) | Top 10; wᵢ = Sᵢ / ΣS |
| P/S + Revenue Growth Q-5 | Same formula | Top 5; wᵢ = Sᵢ / ΣS |
- Fᵢ: fees reported during the preceding 365 days, requiring at least 270 reporting days.
- G⁹⁰ᵢ: latest 90-day revenue divided by the same period one year earlier, minus 1; each window requires 60 observations.
- G³⁶⁵ᵢ: trailing 365-day revenue divided by the preceding 365-day revenue, minus 1.
- βᵢ: downside beta calculated only on days ETH fell; lower is better.
- Vᵢ: total value locked divided by fully diluted valuation.
- Gᵀⱽᴸᵢ: annual total-value-locked growth; negative growth becomes 0.
- PSᵢ: circulating market capitalization divided by trailing 365-day DefiLlama protocol revenue; lower is better.
- Lᵢ: Ethereum decentralized-exchange token inventory divided by OPEN market capitalization.
For methods with a 25% cap, solve wᵢ = min(25%, λ × max(0.01, Sᵢ)) so all weights total 100%.
For TVL Value + Growth Q-10 (v1.1), cap positions above 40% and redistribute the excess proportionally among uncapped positions. Fund every selected position below 1% at 1% by proportionally reducing larger positions. Repeat until every selected position is between 1% and 40% and the portfolio totals 100%. These bands apply only to v1.1.
Tokens missing any required variable are removed before percentile ranks are calculated. Q means Jan, Apr, Jul, and Oct. S means Jan and Jul.
What appears to have mattered
- Concentration helped in this window. The 5-token variants placed more capital in MKR and SYRUP, two of the strongest contributors. That is evidence about this path, not proof that five holdings will always be superior.
- In the only matched cadence test, Fee + Revenue Growth S-10 beat Q-10 at both checkpoints. The overall winner was quarterly, so the test does not establish an optimal cadence.
- The downside factor did not create defense. Backward-looking downside beta could not neutralize shared altcoin exposure, changing correlations, or token-specific collapses.
- The strongest methods still experienced a large giveback. P/S + Revenue Growth Q-5 reached +151% at the ETH peak and finished at +29%. Fee + Revenue Growth S-5 reached +115% and finished at +25%. These are high-beta baskets, not smooth compounders.
- Provider definitions matter. With the same CoinGecko capitalization numerator, DefiLlama revenue produced higher common-universe ending returns than Token Terminal: +36% versus +25% for Q-5 and +29% versus +19% for Q-10. Token Terminal also lacked complete histories for several candidates, so its coverage would influence portfolio membership.
No method should advance solely because it ranked first in one short backtest. The strongest proposal should combine understandable economics, reproducible inputs, manageable maintenance, and performance that survives reasonable sensitivity tests.
6. Inspect or reproduce the work
The independent-review package contains:
- The 19 frozen primary inputs
- Supporting eligibility and token-continuity policies
- The primary backtest and Token Terminal sensitivity code
- Frozen expected outputs
- SHA-256 file manifests
- A data dictionary, methodology guide, limitations, and audit questions
- A sanitized copy of the full diligence memo
Download: 2026-09-19 - OPEN Independent Review Package v4 - Google Drive
To reproduce the calculations:
- Download and unzip the package.
- Open a terminal in the package folder.
- Run
npm run review. - Confirm that the file hashes match and that both calculations reproduce the frozen expected outputs.
The review command requires Node.js but no package installation, API key, wallet, payment, or network connection.
7. What you are encouraged to do now
These eight methods explore only a small slice of the possible combinations among 10 inputs, factor weights, basket sizes, and rebalance schedules. Given how much of that search space remains unexplored, there is a strong likelihood that better-performing, more durable methodologies are still waiting to be found. Treat this leaderboard as a starting line, not a finish line.
Feed the review package to your preferred LLM and let it chew on the work. Ask it to challenge definitions, change variables or weights, propose new combinations, run the supplied backtest, and compare every result with the frozen baseline. An LLM suggestion is only a hypothesis until its formula, inputs, code, and results can be reproduced by someone else.
Please use this thread to:
- Ask questions about any assumption, rule, input, or result.
- Name your preferred methodology and explain why you prefer it.
- Challenge the eligibility criteria, including which projects should enter or leave the universe.
- Identify a better free or low-cost source for any required variable.
- Check the calculations or rerun the supplied code.
- Run your own backtest with different factors, weights, start dates, constituent counts, or rebalance schedules.
- Share the exact formula, changed inputs, code or configuration, results, and limitations so others can reproduce it.
- Share contradictory results and explain exactly what changed.
- Propose a simpler methodology if it remains objective, reproducible, and maintainable by the community.
After at least 14 days of public review, the vlSQUILL-OPEN community can decide whether one method is ready to advance, needs revision, or should return to research.
