/vectorbt-expert

Crypto Market Transaction Costs

All fee calculations are based on standard crypto exchange fee tiers (base/regular tier). These defaults represent the most common fee structure across major exchanges.

Fee Summary by Segment

Simplified Percentage Fees for VectorBT

VectorBT's fees parameter is a percentage applied to both buy and sell turnover. Use the taker fee for conservative modeling (market orders).

Crypto Spot (Base Tier)

Maker and taker both 0.1% at VIP 0. Most backtest fills simulate market orders (taker).

# Crypto Spot (Base Tier): 0.1% taker fee per side
fees = 0.001             # 0.1% per side (taker)
fixed_fees = 0           # No fixed fee on most crypto exchanges

pf = vbt.Portfolio.from_signals(
    close, entries, exits,
    fees=fees,
    fixed_fees=fixed_fees,
    init_cash=10_000,     # $10K USDT
    freq="1D",
    min_size=0,           # Crypto allows fractional units
    size_granularity=0,   # No rounding needed
)

Crypto Spot (BNB Discount)

25% discount when paying fees with BNB. Reduces 0.1% to 0.075%.

# Crypto Spot (Token Discount): 0.075% taker fee per side
fees = 0.00075           # 0.075% per side (taker with BNB)
fixed_fees = 0

pf = vbt.Portfolio.from_signals(
    close, entries, exits,
    fees=fees,
    fixed_fees=fixed_fees,
    init_cash=10_000,
    freq="1D",
    min_size=0,
    size_granularity=0,
)

USDT-Margined Futures

Maker 0.02%, Taker 0.05% at VIP 0. Use taker fee for conservative backtests.

# USDT-M Futures (Taker): 0.05% per side
fees = 0.0005            # 0.05% per side (taker)
fixed_fees = 0

pf = vbt.Portfolio.from_signals(
    close, entries, exits,
    fees=fees,
    fixed_fees=fixed_fees,
    init_cash=10_000,
    freq="1D",
    min_size=0,
    size_granularity=0,
)

USDT-M Futures (Maker Orders)

For strategies using limit orders, the lower maker fee applies.

# USDT-M Futures (Maker): 0.02% per side
fees = 0.0002            # 0.02% per side (maker)
fixed_fees = 0

pf = vbt.Portfolio.from_signals(
    close, entries, exits,
    fees=fees,
    fixed_fees=fixed_fees,
    init_cash=10_000,
    freq="1D",
    min_size=0,
    size_granularity=0,
)

COIN-Margined Futures

Maker 0.01%, Taker 0.05% at VIP 0. Settled in the base cryptocurrency.

# COIN-M Futures (Taker): 0.05% per side
fees = 0.0005            # 0.05% per side (taker)
fixed_fees = 0

pf = vbt.Portfolio.from_signals(
    close, entries, exits,
    fees=fees,
    fixed_fees=fixed_fees,
    init_cash=1.0,        # 1 BTC for BTC-margined
    freq="1D",
    min_size=0,
    size_granularity=0,
)

Quick Reference: Default Fee Constants

Use these constants at the top of every crypto backtest script:

# --- Fee Constants (Crypto Exchange Standard - Base Tier) ---
# Spot Trading
FEES_CRYPTO_SPOT = 0.001             # 0.1% per side (taker)
FEES_CRYPTO_SPOT_BNB = 0.00075      # 0.075% per side (taker with BNB discount)
FEES_CRYPTO_SPOT_MAKER = 0.001      # 0.1% per side (maker)

# USDT-Margined Futures
FEES_CRYPTO_FUTURES_TAKER = 0.0005  # 0.05% per side (taker)
FEES_CRYPTO_FUTURES_MAKER = 0.0002  # 0.02% per side (maker)

# COIN-Margined Futures
FEES_CRYPTO_COINM_TAKER = 0.0005   # 0.05% per side (taker)
FEES_CRYPTO_COINM_MAKER = 0.0001   # 0.01% per side (maker)

# No fixed fees on most crypto exchanges
FIXED_FEES_CRYPTO = 0

Exchange VIP Tier Fee Schedule (Typical)

Spot Trading

USDT-M Futures

Funding Rate (Futures Only)

Perpetual futures contracts have a funding rate exchanged between longs and shorts every 8 hours. This is a hidden cost that significantly impacts longer-duration futures backtests.

Modeling Funding Rate in VectorBT

Funding rate is NOT modeled by VectorBT's fees parameter. For short-term backtests (<1 week), it can be ignored. For longer holding periods, account for it separately:

# Approximate funding rate impact on a futures backtest
FUNDING_RATE = 0.0001      # 0.01% per 8 hours (typical neutral market)
FUNDING_PERIODS_PER_DAY = 3  # Every 8 hours

# For a position held for N days:
# funding_cost = position_value × FUNDING_RATE × FUNDING_PERIODS_PER_DAY × N_days

# Conservative approach: add estimated daily funding to fees
# 0.01% × 3 = 0.03% per day → for daily bars, add to fees
fees_with_funding = 0.0005 + 0.0003  # taker + daily funding estimate

pf = vbt.Portfolio.from_signals(
    close, entries, exits,
    fees=fees_with_funding,  # 0.08% per side (taker + funding)
    fixed_fees=0,
    init_cash=10_000,
    freq="1D",
)

Funding Rate Notes

• Positive funding rate: Longs pay shorts (bullish market). Hurts long positions.
• Negative funding rate: Shorts pay longs (bearish market). Hurts short positions.
• Typical range: -0.05% to +0.05% per 8h, with 0.01% being the baseline
• In strong trends, funding can reach 0.1%+ per 8h, which is significant
• For backtests with average holding period > 1 day, always consider funding rate impact

Data Source for Crypto Markets

Use yfinance or CCXT for crypto market data:

import yfinance as yf

# Bitcoin (daily)
df = yf.download("BTC-USD", start="2022-01-01", end="2025-01-01", interval="1d")

# Ethereum (daily)
df = yf.download("ETH-USD", start="2022-01-01", end="2025-01-01", interval="1d")

# Benchmark: Bitcoin
benchmark = yf.download("BTC-USD", start="2022-01-01", end="2025-01-01", interval="1d")

For higher resolution data or futures data, consider using CCXT:

# pip install ccxt
import ccxt
import pandas as pd

exchange = ccxt.binance()  # or any supported exchange
ohlcv = exchange.fetch_ohlcv("BTC/USDT", timeframe="1h", limit=1000)
df = pd.DataFrame(ohlcv, columns=["timestamp", "open", "high", "low", "close", "volume"])
df["timestamp"] = pd.to_datetime(df["timestamp"], unit="ms")
df = df.set_index("timestamp")

Crypto-Specific VectorBT Settings

# Crypto allows fractional units - do NOT set min_size=1
pf = vbt.Portfolio.from_signals(
    close, entries, exits,
    fees=0.001,              # Crypto spot taker
    fixed_fees=0,
    init_cash=10_000,        # USDT
    size=0.5,                # 50% per trade
    size_type="percent",
    direction="longonly",    # or "both" for spot + short
    freq="1D",
    min_size=0,              # Fractional crypto allowed
    size_granularity=0,      # No rounding
)

Cost Comparison Across Markets

Best Practices

• Spot trading: Always use taker fees (0.1%) for conservative backtesting. Most signals result in market orders.
• Futures: Model both taker fees AND funding rate for positions held > 1 day
• Token discount: Only apply exchange token discounts if you realistically hold them. Use base fees for conservative modeling.
• Slippage: Crypto markets can have significant slippage on larger orders. Add slippage=0.001 (0.1%) for realistic modeling on altcoins or low-liquidity pairs.
• Fractional units: Unlike stocks, crypto allows fractional trading. Do NOT set min_size=1 or size_granularity=1 for crypto.
• 24/7 markets: Crypto trades 24/7 - no weekend gaps. Use freq="1D" for daily, but note that weekends are trading days.
• When in doubt, use spot taker (0.1%) as a safe default for any crypto backtest
• Default crypto benchmark: Bitcoin (BTC-USD) from yfinance