Copy trading · Solana · fee-net accounting
The leaderboard
is gross.
You trade net.
A wallet’s headline profit is measured before the venue fee, the creator fee, the router fee, priority and tip, and slippage. MYCEL subtracts all of it, reconstructs the ledger from chain data, and mirrors only the wallets that are still profitable at the end of that subtraction.
No index has been run. Every figure on this page resolves from chain data or stays a dash.
The subtraction
Six layers sit between a fill and a profit
Published venue rates, read from vendor documentation on 2026-08-17. These are the rates, not your results. Nothing here is measured from a wallet, and the axis below is the axis they are drawn against.
- ABonding curvepump.fun1.25%per side0.95% protocol + 0.30% creator
- BPool swapPumpSwap, post-graduation0.30 – 1.25%per sidedynamic by market cap: 1.25% at 0–420 SOL, 0.30% at ≥98,240 SOL. The base swap component of that is 0.25%.
- CRouterPumpPortal Local0.50%per tradenon-custodial; MYCEL signs and broadcasts
- DPriority + tipleader / Jito0.10 – 0.30%typicalvaries with congestion, not a fixed rate
- ESlippagefill quality0.50 – 2.00%typicalthe largest and least predictable layer
- FMYCELour own takeunsetunsetno band drawn — this rate is not setpricing is being re-derived before execution ships
- 0.000.501.001.502.00percent of trade value
Third-party composite for a full round trip on a migrated pump.fun token, before any copy-tool fee:
3.2% – 6.6%
Cited estimate, 2026-08-17. MYCEL replaces it with your wallet’s own measured figure once the ledger runs.
Signature instrument
The truth ledger
One wallet, one window, every cost. The instrument shows the whole distance between the number a leaderboard prints and the number that reaches a copier’s balance.
The whole rail is hatched because no point on it has been computed. Net resolves to one point in here when the audit runs, and until then the drawing marks nothing — not a midpoint, not a default.
Segment values resolve when an audited wallet is loaded. No wallet is loaded, so every figure is a dash. This instrument has no demo mode.
| # | Wallet | Closed tokens | Gross | Σ costs | Net after fees | Verdict |
|---|---|---|---|---|---|---|
No wallet has been ranked The ordering is MYCEL's own fee-net figure, so a row cannot exist before an audit has produced one. The columns stay so you can see the shape of what a row will hold. | ||||||
No wallet is named on this page, in any state. An exclusion is an accusation, and an accusation belongs in the product with its signatures attached — never in marketing copy.
Wallet audit
Six rules, cheapest first, stop at the first hit
Legitimate trading can resemble wash trading, which is why no rule produces a verdict on its own and why every rule that fires has to hand back the transactions it fired on.
- 01
Same-signature round trip
not runA buy and a sell of the same token inside one transaction.
cost · one parsed transaction
- 02
Zero-risk position
not runNear-identical size bought and sold within a short window at near-identical price.
cost · one wallet history
- 03
Shared funding ancestor
not runThe wallet and its own counterparty trace back to the same non-exchange funder.
cost · n-hop transfer walk
- 04
Bundle co-membership
not runThe wallet shares a bundle with the account on the other side of its fill.
cost · bundle lookup
- 05
Circular volume
not runA quantity cycles through several accounts and returns to the origin.
cost · cycle detection
- 06
Persistent pattern
not runThe same behaviour repeats across multiple tokens over time.
cost · cross-token pass
Copy rules
no account attachedEvery rule is yours, except the one that protects you
Nothing here is pre-filled with a suggested figure. A default position size is a recommendation, and MYCEL does not make one on a page that has never seen your account. Values read unset until an account is attached; they are not zeros and they are not defaults.
Unknown bankroll skips the copy
EnforcedProportional sizing needs the source’s bankroll. When it cannot be estimated, MYCEL skips the trade rather than guessing a size. This control cannot be turned off.
Mode
Share of source position
applies in proportional mode
Fixed size
applies in fixed mode
Round down to whole units
Cap per source wallet
Aggregate cap
across every followed wallet
Max simultaneous sources
Daily loss cap
Mirror only audited wallets
fails the wash ladder, never copied
Minimum pool liquidity
Minimum closed tokens on the source
Token blacklist
Auto-demote on edge decay
stops mirroring when the audited edge falls
Demotion threshold
Paper mode
runs the whole path, signs nothing
Kill switch
halts every task instantly
Six stages, in order. Nothing skips ahead, and stage six refuses to run when stage five came back without evidence.
Method
How the number gets built
Stage 01. Discover
Pull a candidate set from a vendor PnL screen, filtered for arbitrage, thin history and single-token flukes.
Stage 02. Shortlist
Drop the candidates that cannot survive an audit. Full reconstruction is expensive; it is spent on a shortlist.
Stage 03. Reconstruct
Rebuild the wallet's ledger from parsed transactions with a FIFO cost basis, denominated in SOL.
Stage 04. Subtract
Deduct every layer from the fill itself — venue, creator, router, priority, tip, and realised slippage.
Stage 05. Test
Run the wash ladder in cost order and stop at the first rule that fires, keeping its signatures.
Stage 06. Publish
Write the verdict with its evidence, or write nothing. A wallet with no evidence gets no verdict.
Two ledgers, one delta
The vendor screen answers which wallets are worth auditing. MYCEL’s own reconstruction answers which of them survive. Both figures get published, and so does the gap between them.
gross of costs
own reconstruction
the finding
Both ledgers stay unresolved until a wallet has been reconstructed. The delta is not estimated from one side.
- Non-trade balance changes — airdrops, transfers in, dust — are classified and excluded from cost basis rather than quietly absorbed, and the flag is visible.
- Historical fiat conversion does not chain cleanly across days, so the ledger is denominated in SOL and converted at the edge.
- Wash thresholds are ours to set, so they are published as numbers you can argue with, not hidden inside a score.
- No published false-positive rate exists for these detection methods. That is stated on the verdict, not buried in a footnote.
Copy latency
no copy executedMeasured in slots, because Solana has no clock
A millisecond figure would be a claim nobody can check. Block timestamps are second-granularity, so the only honest measurement is the slot the source filled in, the slot MYCEL filled in, and the distance between them — published per copy, with the unknown left as unknown.
Source fill slot
—
slot
MYCEL fill slot
—
slot
Delta
—
slots
No copy has been executed, so there is no slot to report. Landing rate, not detection speed, is what separates one execution path from another — and that gets published the same way.