GOTRADEWARS.COM/METHODOLOGY · LAST REVIEWED AUG 2026
How TradeWars computes everything it shows you.
This page documents the mechanics behind paper execution, verified receipts, rank scoring and crowd intelligence — in plain terms, including where the numbers are estimates rather than certainties. Nothing here is a promise of future results.
01 · Paper execution
02 · Verified receipts
03 · How ranking works
04 · Crowd intelligence
05 · Data provenance
06 · Limitations
01
How paper execution works
Every TradeWars position is simulated against real market data, not a backtest or a random-walk model. No real funds move at any point.
Market data sources
Prices are aggregated from six spot and derivatives exchanges via direct websocket feeds. Each market's displayed price is a volume-weighted composite, not a single exchange's quote — see Data provenance for the exact source list.
Bid / ask and mid price
Order tickets quote the composite bid and ask directly. Chart price and unrealized P&L use the mid price, (bid + ask) ÷ 2, refreshed on every feed tick.
Order timestamps
Every order is timestamped to the millisecond at submission, independent of when the fill simulation completes. This submission timestamp, not the fill timestamp, is what the audit history and any public receipt display.
Fill simulation
Market orders fill against the composite book at submission time. Limit and stop orders fill only once the composite mid price actually trades through the specified level — they are never guaranteed a fill just because price approaches it.
Market order slippage assumption0.02% – 0.06% of notional, size-scaled
Simulated taker fee0.04% of notional
Simulated maker fee0.02% of notional
Funding rate (perpetuals)Pass-through, aggregated across venues
These figures approximate typical retail-size execution on major exchanges. They are estimates, reviewed quarterly against real fill data — not a guarantee that live execution would match exactly.
Delayed or missing data
If a feed drops below four connected exchanges or latency exceeds 2 seconds, that market is flagged "degraded" and new orders are held until the composite stabilizes. Open positions are marked at the last valid composite price and clearly labeled as stale.
02
How verified receipts work
A verified receipt is the permanent record generated for every trade. It cannot be edited or deleted once created.
1
Submission
The order is timestamped and hashed the instant it leaves the client. This hash is what later ties the receipt back to an unaltered submission.
2
Market snapshot
The composite bid, ask, mid and per-exchange contributions at submission time are captured and stored alongside the order — the same data an auditor would need to recompute the fill independently.
3
Fill
The fill simulation applies slippage and fee assumptions to the snapshot and records the simulated fill price, size and cost basis.
4
Close
Closing repeats the snapshot-and-fill process and locks the realized P&L, R multiple, and holding time onto the receipt.
5
Audit history
Every state change — partial closes, stop moves, tag edits — appends to the receipt's history rather than overwriting it. The original submission and fill are never rewritten.
Immutability. Once a receipt closes, its submission hash, market snapshot and fill data are write-locked. Deleting a trade removes it from a trader's public feed but does not remove it from the ladder's underlying score history — a trader cannot quietly erase a bad call from their own ranking.
03
How ranking works
Rank Score is a weighted composite of five categories, recomputed nightly against every trader in the same strategy class.
Performance 30%
Risk-adjusted returns relative to peers in the same class — raw P&L alone is not scored.
Consistency 25%
Variance of results trade-to-trade and week-to-week. Steadier records score higher than streaky ones with the same average.
Risk control 20%
Position sizing discipline, stop usage, and maximum drawdown relative to account size.
Sample confidence 15%
How much verified trade history backs the score. Below the threshold, an account is marked provisional (below).
Activity 10%
Recency and frequency of qualifying trades. Long inactivity discounts the score gradually rather than resetting it.
Provisional accounts
Accounts with fewer than 20 verified qualifying trades are labeled PROVISIONAL. Their Rank Score is calculated and shown, but it is not weighted against established traders when computing family/tier boundaries until Sample Confidence clears the threshold. This keeps a single hot week from vaulting a new account past traders with a long track record.
Rank, Leaderboard Position, and Season Rating are three different numbers
04
How crowd intelligence works
Crowd bias and cohort splits are counts of open positions, not predictions — they describe what traders are doing right now, not what will happen next.
Qualifying usersVerified accounts with ≥ 1 open or closed position in the market
Minimum sample size25 open calls before a bias figure is shown publicly
Open callsCurrently-open positions only; closed positions do not count toward bias
Top-100 cohortThe 100 highest Rank Score accounts, recomputed nightly
Star Admiral cohortAccounts holding the apex rank; shown only when ≥ 5 qualifying accounts hold it, to protect individual privacy
Update frequencyEvery 60 seconds while a market is active
Below the minimum sample size, a market shows "insufficient sample" instead of a bias figure rather than a misleadingly precise percentage.
05
How data provenance works
Every figure on TradeWars can be traced to a source, a timestamp, and an observation count. Below is the live source list for market data.
SOURCE
TIMESTAMP
PERIOD
OBSERVATIONS
Binance
Real-time
Rolling 24h
~4.1M ticks/day
Coinbase
Real-time
Rolling 24h
~2.6M ticks/day
OKX
Real-time
Rolling 24h
~2.1M ticks/day
Bybit
Real-time
Rolling 24h
~1.7M ticks/day
Kraken
Real-time
Rolling 24h
~0.9M ticks/day
Bitget
Real-time
Rolling 24h
~0.5M ticks/day
Figures above are illustrative of typical daily volume and are refreshed continuously; a live per-market breakdown is available from any market detail page.
06
Limitations
Paper fills are a simulation built on real market data, not a guarantee that live execution would produce the same price, slippage or fees.
Rank Score, Leaderboard Position and Season Rating summarize historical, self-reported paper trading activity. They are not a certification of skill and do not predict future performance, in paper or live markets.
Crowd bias and cohort figures describe current open-position counts among TradeWars users. They are not a recommendation to trade in either direction.
No system here is infallible. Feed outages, exchange incidents and edge cases in the fill simulation can all produce results that don't match this document exactly — when they do, we treat it as a bug to fix and disclose, not a rounding error to ignore.
Questions about a specific figure or receipt? Every receipt links back to the exact market snapshot used to compute it.
TRADEWARS
© 2026 TRADEWARS · PAPER TRADING · NO REAL FUNDS