Quick start
Four steps to a useful first comparison.
- 01Start with realistic capital.
Your deposit changes the modeled liquidity share and fee capture. Enter an amount from $1 to $100,000,000.
- 02Use 1M for the first broad scan.
The leaderboard defaults to the last 30 completed days. Compare 1D and 1W to see how more recent conditions differ.
- 03Open one result.
The exact pool, range, capital, timeframe and compounding setting carry into the detailed analysis.
- 04Read period return before annualized rates.
Review fees, ending LP value, IL and the separate HODL benchmark. Check the observed dates and coverage before comparing two results.
Your workspace
Each network keeps its own catalogue, live data and history.
Robinhood Chain
145 pools: blue chips, stocks and RWA, altcoins and memecoins.
Base
103 pools across Uniswap V3, Aerodrome Slipstream and PancakeSwap V3, including stablecoin pairs.
The header switches networks and opens the full contract register or this guide. Scroll through the leaderboard, directory, calculator and detailed backtest. The theme button switches between dark and light; the refresh button requests current market data.
- Market summary
- Catalogue coverage, current volume and accepted live snapshots describe the selected network. The top rolling 24h pool fee APR is a pool-wide trailing screen, separate from a position backtest.
- Bridge & swap access
- Links to the supported network routes are retained. Quotes and transaction signing happen at the destination service, outside this research tool.
Global Pool × Range × Delay leaderboard
Compare the complete selected-network catalogue on a bounded historical window.
| Control | How to use it |
|---|---|
| All pools | Includes every packaged pool on the selected network, regardless of directory search, pagination or liquidity filters. Entries without sufficient verified inputs are accounted for as unavailable. |
| Scan one pool | Search by pair or contract and select an exact entry. Clear it to return to a category or all-pool scan. |
| Historical window | Defaults to 1M (30 days). Select 1D or 1W for a shorter comparison. All entries in that scan request the same window. |
| Auto-compound | OFF keeps fees outside the LP. ON reinvests modeled matching-token fees at rebalance events. The audit identifies whether fee-token direction is observed or assumed. |
| Objective | Balanced considers return and robustness. Highest return favors observed strategy return. Lowest IL favors smaller modeled impermanent loss. |
| Results / confidence | Display the best range per pool or every tested range. Confidence filters affect displayed results, not the catalogue being scanned. |
| Run / stop | Results arrive progressively. Stop keeps completed results; the status records what was completed, unavailable or not yet checked. |
Pool directory
A compact view of the whole catalogue, twelve pools at a time.
Search and filter
Search the entire directory by pair, token or contract. Choose a category, protocol where available, or minimum liquidity. These controls apply before pagination.
Sort and page
Click a column heading to sort; click again to reverse it. Previous and Next navigate the filtered results. Page changes do not change the pool already open in the calculator.
Select a row to open its analysis. Each row has a separate swap-fee badge, alongside price, liquidity and volume, range fit and estimated range APR. Fixed fee tiers are distinguished from current dynamic fees. Full contract details remain available in the selected pool and Pool contracts.
Position Calculator
One pool, one range and one selected historical window.
- 01Choose 1D, 1W or 1M.
The chart, modeled fee rate, range statistics, backtest and optimizer all switch to that selected window. History requests use completed hourly candles.
- 02Enter capital and adjust the range.
Sliders change the lower and upper percentage buffers. Presets set both together. Calculations update locally from the retained verified sample.
- 03Inspect the chart.
The history scrubber selects a displayed candle and timestamp. It changes the chart cursor, not the backtest window or its starting date.
- Observed fee ROI
- Modeled net fees divided by your capital over the displayed sample. It excludes IL and rebalancing costs.
- Annualized fee rate
- The sample fee ROI multiplied by 365 / observed days. A short sample can produce a large annualized number without having earned that return.
- Modeled net / day
- Net modeled fees divided by the actual observed duration.
- Estimated active fraction
- A pair-relative OHLC overlap estimate. It is not exact measured wall-clock time in range.
Historical range and fee backtest
Compare the selected range, alternative ranges and capital scenarios.
1D requests 24 completed hourly candles, 1W requests 168 and 1M requests 720. A newer pool may have a shorter verified sample; its observed dates, days and coverage show that limitation. Missing periods are never filled with invented prices or volume.
The static fee test uses the displayed range, exact capital, pool-relative price path and verified fee inputs. It deducts the modeled 15% MaxFi performance fee. Changing range width changes position liquidity and in-range activity; a wider range can stay active longer while earning a smaller share per unit of capital.
Rebalance delay optimizer
Model when to reposition after a continuous range exit.
The model tests every distinguishable 0–168-hour delay. The timer starts when price leaves the range and resets on re-entry; it is not a recurring rebalance schedule. At hourly resolution, overlapping candles cannot prove a full hour continuously outside the range.
Each trigger uses the pool's verified tick spacing for a no-swap, single-sided repositioning model. Outcomes with identical trades and cash flows are grouped into a delay band. No-exit samples cannot establish a useful delay preference.
- Strategy P/L
- Ending modeled LP inventory plus claimable or external fees, minus starting capital. Compounded fees already inside LP inventory are not counted twice.
- vs HODL
- The same strategy value compared with holding the original token basket. This is a benchmark difference, not account P/L.
- Holdout
- A separate replay on the final 30% of candles checks whether the result remains useful outside the fitting sample. It is not a future forecast.
- Path audit
- Shows entry and exit dates, starting basket, ending inventory, fees, compounding and liquidity-depth basis so the result can be understood.
Read the results together
No single metric establishes a good position.
| Metric | What it answers |
|---|---|
| Period return | How did the modeled strategy change the starting capital over these dates? |
| IL and vs HODL | How did the position compare with holding its starting tokens? |
| Fees / IL | Did modeled fees cover ending IL in this sample? |
| Coverage and confidence | How complete are the prices, fee inputs and liquidity assumptions? |
| Rebalances and holdout | How active was the strategy, and did its result persist in the later sample? |
A repeatable research workflow
Keep your comparisons consistent.
- 01Scan one network at 1M.
Use the same deposit and compounding setting across the full catalogue or your selected category.
- 02Open a shortlist.
Compare period return, IL, holdout and coverage. Review the exact pool contract.
- 03Check recent conditions.
Compare 1W and 1D, then vary range width and capital. Read each sample's observed dates.
- 04Verify the live position.
The retained Open MaxFi and market links let you check the destination pool and current conditions before any action.
Data, model boundaries and security
Know which inputs are observed and which are modeled.
- Contract identity
- Each request resolves by network, exact pool contract and exact asset contract. Matching ticker names do not substitute for matching contracts.
- Price history
- USD and pair-relative candles must be aligned, valid and contiguous. No unclosed candle counts as a complete interval.
- Fee and depth basis
- Aerodrome historical results require complete hourly generated fees, LP retention and depth. Fixed-fee V3 pools use verified historical depth where available, otherwise an explicitly disclosed current concentrated-liquidity pool state assumption. Required missing inputs are withheld.
- Excluded costs and uncertainty
- OHLC data does not reveal the exact intrabar path. Gas, taxes, execution latency, price impact and slippage are not reconstructed. AERO and CAKE incentive emissions are excluded from fee returns. Modeled results are not exact realized wallet returns or predictions.
- Read-only operation
- The site does not connect to wallets, request approvals, sign transactions or move funds. Optional provider credentials stay on the server.
Support the project
An optional, separate flow.
The Buy me a coffee page provides verified public network and token details to copy. It does not open a wallet or submit a transaction.
Common questions
Quick answers while you work.
Why does the rate change when I move a slider?
The same capital buys different position liquidity at different widths. The price path also spends a different estimated fraction inside each range. Both affect modeled fee capture.
Why can 1D show a four-digit percentage?
The one-day fee ROI is multiplied by 365 for the annualized display. It is a short-sample rate, not a forecast of a year's return.
Does All pools scan only the current directory page?
No. It covers the complete selected-network catalogue. Directory pagination and filters do not reduce the scan scope. Unavailable entries are accounted for separately.
Why does a 1M result show fewer than 30 days?
The pool may be newer or the provider may have less complete indexed history. Read its sample dates and coverage. The model annualizes only the observed duration.
Does the chart scrubber change the backtest?
It selects the price and timestamp shown on the chart. Use 1D, 1W or 1M to change the actual backtest window.
Are these exact realized returns?
No. They are estimates using verified inputs and disclosed assumptions. Complete price history cannot reconstruct every trade, liquidity change or execution cost.
