Crypto trading bot glossary

PAPER mode

Trading against live market data with virtual positions. The strategy receives the same Binance feed it would on LIVE; entries and exits are recorded; fees and slippage are simulated; but no real order ever reaches the exchange. Default for every new Quantor bot. See the PAPER guide.

LIVE mode

Trading on a real exchange account via API. Positions are funded by the user's exchange balance; the bot places real buy/sell orders. Cannot be enabled on day one — requires explicit unlock after sufficient PAPER history and 14 risk gates passing.

Regime classifier

A coarse market-state label (CALM, VOLATILE, DANGEROUS) computed from realized volatility and drawdown from peak. When DANGEROUS, Quantor refuses to start new LIVE positions. See the mechanism.

Kill switch

An automated trigger that pauses a bot when its daily loss exceeds a configured percentage of starting capital. Plan-dependent: Starter 3%, Pro 5%, Teams configurable. Once tripped, the bot will not place new orders until the next UTC day rollover.

Signed build

A production binary whose identity is cryptographically attested by an offline private key. Quantor uses Ed25519. The public key is at /.well-known/quantor-pubkey; you verify the running binary in 30 seconds with one openssl command. See the mechanism.

Self-custody

An arrangement where the user retains the technical authority to move their own funds. The bot operator has no withdraw permission and structurally cannot move user money. Contrast with custodial — where you deposit funds with the operator. See the architecture.

Withdraw scope

A permission flag on an exchange API key that grants the ability to initiate a withdrawal. Quantor requires this flag to be OFF on every connected key. Withdraw enforcement is on the exchange side — Binance refuses the operation regardless of what our code does.

Risk gate

An independent pre-trade check that can block a LIVE order. Quantor runs 14 of them in sequence — symbol allowlist, regime, kill switch, plan permissions, exchange key verification, notional ceiling, etc. Each gate emits a machine-readable reason code if it fails. See the full list.

Slippage

The difference between the expected price of a trade and the price at which it actually executes. Caused by order book depth being insufficient at the expected price. Quantor models slippage in PAPER as a fixed 8 basis-point haircut per leg — close to reality for majors like BTC/USDT.

Profit factor

The ratio of gross winning trades to gross losing trades. Above 1.0 means net positive over the sample. Above 1.5 is a solid retail signal. Above 3.0 on real data is suspicious — usually a sign of overfitting or insufficient sample size.

Sharpe ratio

Annualised return divided by annualised volatility. Measures return per unit of risk. Quantor computes 365-day, zero-risk-free-rate. Above 1.0 is good for retail crypto strategies; above 2.0 is exceptional (verify the strategy isn't peeking at the future).

Maximum drawdown

The largest peak-to-trough decline in equity over the observed period, as a percentage. The single most important measure of risk because it tells you how much you'd be down at the worst moment, not just on average. If max DD is bigger than you'd emotionally tolerate at 1× notional, the strategy isn't right for you.

Forward test

Running a strategy on live incoming market data without placing real orders. The strategy can't peek at the future because the future hasn't happened yet. Same as PAPER mode in Quantor terminology.

Backtest

Running a strategy on historical market data offline. Useful for filtering obviously broken strategies; dangerous as a stand-alone signal because it's easy to overfit to a known past. Quantor does not provide a built-in backtester — we prefer the forward-test discipline of PAPER.

Buy-and-hold benchmark

The performance of simply holding the asset over the same period the strategy was active. Used to verify the strategy added value: if BTC went up 40% and your bot delivered +35%, you under-performed the asset and added complexity. The performance dashboard shows this explicitly.

Audit chain

A tamper-evident database table where each row's hash incorporates the previous row's hash. Modifying a historical row breaks the chain at that point; the system detects it within an hour via a Pulse probe. Records every meaningful state change in Quantor: LIVE starts, plan changes, kill switches, webhook fulfillments.

Custodial bot

A trading bot that holds user funds in a wallet it controls. To use the bot you deposit; to withdraw you initiate a transaction through the operator. Quantor is NOT custodial — see self-custody. Examples of custodial: Pionex (the exchange holds your funds).

TWAP

Time-weighted average price. An execution strategy that splits a large order into smaller time-spaced child orders to minimise market impact. Useful for big notionals where placing one order would move the price against you. Quantor doesn't ship a TWAP bot today.

Pine Script

TradingView's scripting language for defining technical-analysis strategies. Quantor imports v1 syntax (the simpler dialect); v2 is on the roadmap. Lets users bring their own strategies without writing Java.

Identity endpoint

A public HTTP endpoint at /api/v1/identity that returns a signed manifest describing the running binary: commit hash, image hash, build timestamp, plus the Ed25519 signature over those fields. The verifiable counterpart to "trust me" — anyone with the public key can confirm the running binary matches the declared binary.

Want to see these in practice?

FREE PAPER plan — no card. Try a strategy, watch the regime classifier do its thing.