Understand the measurement.
34 guides connect financial mechanics, market observations and data quality to the working tools. Read the formula, follow an example and check the assumptions before interpreting a result.
Why crypto arbitrage spreads disappear
See how order-book depth, two trading fees, execution timing and inventory costs can turn a positive Bitcoin price spread into a loss.
Read and apply ↗Research guideFunding APR is not your return on capital
Separate annualized funding on notional from return on committed capital. Work through a spot-perpetual example including fees and margin reserves.
Read and apply ↗Research guideHow funding intervals and negative rates change the calculation
Compare perpetual funding rates with different intervals, understand negative funding and calculate a two-venue funding difference correctly.
Read and apply ↗Research guideOrder-book depth versus trading volume
Learn why daily volume does not determine your fill price. Calculate a depth-weighted BTC purchase and compare liquidity at your intended order size.
Read and apply ↗Research guideBitcoin DCA average price, with a worked example
Calculate weighted Bitcoin acquisition cost from recurring buys. See why averaging purchase prices is wrong and how fees change your break-even price.
Read and apply ↗Research guideBacktesting without look-ahead bias
Build a credible Bitcoin backtest by separating signal time from fill time, modeling costs and rejecting future information and ambiguous candle paths.
Read and apply ↗Research guideMaker versus taker fees
Understand maker and taker fees, why a limit order can still pay taker rates, and how partial fills change the costs used in a crypto trade calculation.
Read and apply ↗Research guideBitcoin virtual bytes versus bytes
Learn the difference between Bitcoin bytes, weight units and virtual bytes. See how transaction vsize connects a sat/vB fee rate to the fee in satoshis.
Read and apply ↗Research guideRealized versus unrealized crypto P&L
Understand realized and unrealized crypto profit, partial closes and fee treatment. Learn why a green position display can differ from final net proceeds.
Read and apply ↗Research guideBitcoin spot versus perpetual versus dated futures
Compare Bitcoin spot, perpetuals and dated futures by ownership, funding, expiry and settlement. Choose the cost model that matches the position you hold.
Read and apply ↗Research guideOrder-book snapshots versus executed trades
A displayed order is an offer to trade, while an executed trade records a match. Snapshot differences alone cannot determine whether removed size was cancelled, filled or moved.
Read and apply ↗Research guideWhy TVL can rise without new deposits
USD-denominated total value locked depends on both token quantities and their prices. A percentage TVL change is not automatically a net-inflow percentage.
Read and apply ↗Research guideBase APY versus reward APY in DeFi pools
A headline pool yield can combine activity-generated yield with incentive tokens. Those components can have different denomination, liquidity and persistence.
Read and apply ↗Research guideStablecoin quote-asset risk in cross-exchange comparisons
USDT, USDC and USD are different settlement assets. Matching the base token name is insufficient to make two prices directly comparable.
Read and apply ↗Research guideLending utilization and the interest-rate kink
A utilization curve can charge a steeper marginal borrow rate above a configured target. The current rate is conditional on both the curve parameters and pool utilization.
Read and apply ↗Research guideLoan-to-value versus liquidation threshold
The maximum borrowing ratio and liquidation threshold serve different purposes. Borrow capacity is an entry constraint; the liquidation threshold is used in collateral coverage tests.
Read and apply ↗Research guideRepay debt or add collateral to repair health factor
Debt repayment changes the denominator of health factor; collateral top-up changes its weighted numerator. Compare their cash requirements under the same target and threshold.
Read and apply ↗Research guideAMM price impact versus slippage tolerance
Price impact is a consequence of trading against the current liquidity curve. Slippage tolerance is a limit supplied to a transaction; it does not remove expected impact.
Read and apply ↗Research guideHow concentrated-liquidity inventory changes across a range
A concentrated-liquidity position changes its token mix as price moves through its selected range. Outside that range, the principal becomes one-sided under the simplified two-token model.
Read and apply ↗Research guideWhen maker rebates do not offset trading costs
A maker rebate reduces one fee component. It does not measure adverse selection, missed fills, price drift or the cost of changing inventory after execution.
Read and apply ↗Research guideCandle-derived VWAP versus trade-level VWAP
Candle-derived VWAP weights an assumed representative candle price by its volume. True executed VWAP requires trade prices and quantities for the same window.
Read and apply ↗Research guideComparing funding rates with different intervals
A funding rate needs its settlement interval and cash-flow sign. Equal displayed percentages can imply different simple cash-flow scenarios when the intervals differ.
Read and apply ↗Research guideOpen-interest units and price effects
Open interest can be reported in contracts, base units or quote value. A rise in quote-denominated open interest may partly reflect a price change.
Read and apply ↗Research guideCorrelation alignment with missing crypto candles
A return comparison requires the same time intervals. Adjacent rows in a file are not necessarily adjacent market periods.
Read and apply ↗Research guideChart timeframes and indicator warm-up
An indicator period counts observations, not a fixed number of clock hours across every chart. Changing timeframe changes the input series and warm-up requirements.
Read and apply ↗Research guideQuote notional versus base quantity in liquidity analysis
Order quantity describes token units. Quote notional multiplies those units by price, allowing differently priced orders to be compared in the same quote currency.
Read and apply ↗Research guideDuplicate market data and resampling checks
Duplicate timestamps or price levels can change statistics if treated as independent observations. Decide whether rows are updates, trades or aggregated snapshots before combining them.
Read and apply ↗Research guideCash-and-carry basis and convergence assumptions
A fixed-expiry basis scenario depends on spot inventory, a derivative position, funding or financing costs and the actual settlement mechanism. A premium is not guaranteed net income.
Read and apply ↗Research guideOracle prices versus exchange prices in DeFi risk
A lending protocol can use an oracle reference that differs from the last trade on one exchange. Liquidation scenarios should use the protocol’s actual valuation rules.
Read and apply ↗Research guideBridge costs, finality and destination liquidity
A bridge fee is one part of the economic and operational comparison. The relevant destination asset, network cost, route liquidity and arrival conditions must also match.
Read and apply ↗Research guideStaking withdrawal queues and opportunity cost
A withdrawal delay changes when capital becomes available for another use. Opportunity cost depends on a chosen alternative and does not mean a guaranteed loss will be realized.
Read and apply ↗Research guidePortfolio concentration through wrapped and pooled assets
A list of different tickers may still contain repeated exposure to the same underlying asset, issuer or protocol. Concentration analysis depends on the grouping rule.
Read and apply ↗Research guideDeFi fees versus protocol revenue
User-paid fees, supply-side distributions and protocol revenue are different accounting categories. A high activity fee total does not automatically accrue to a token holder.
Read and apply ↗Research guideMarket API cache age and source timestamps
Response receipt time tells you when a system received data. Source time tells you when the provider says the observation was made, when that field exists. Cache age adds another distinct delay.
Read and apply ↗