Resources

Glossary

On this page5

Look up any term from the docs here: a one-line plain definition, how it is written in code and the page that covers it in detail.

Stock selection and backtesting

Term	In code	Meaning
backtest	`report = sim(position, resample='M')`	Evaluate strategy performance using historical data. See Backtesting
position	`position = close > ma20`	The stocks currently held and their weights; `True` means hold that stock on that day. See Stock selection
entry	`entry_signal = close > ma20`	The point at which a strategy generates a buy signal. See Stock selection
exit	`exit_signal = close < ma20`	The point at which a strategy generates a sell signal. See Stock selection
hold until exit	`entry_signal.hold_until(exit_signal)`	Buy when the entry signal is `True` and hold until the exit signal is `True`. See Stock selection
stop loss	`stop_loss=0.1`	Automatically close a position when losses reach a certain percentage; `0.1` means exit at a 10% loss. See Risk management
take profit	`take_profit=0.2`	Automatically close a position when profits reach a certain percentage; `0.2` means exit at a 20% gain. See Risk management
rebalance	`resample='M'`	Periodically adjust position weights to the target allocation; `'M'` means monthly rebalancing. See Backtesting

Term	In code	Meaning
portfolio	`Portfolio({'Strategy A': (report1, 0.5), ...})`	A collection of strategies, each with a weight. See Multi-strategy portfolio
Position	`Position.from_report(report, fund)`	The holdings an order is built from: symbols mapped to quantities (shares in a US account). See Orders
OrderExecutor	`OrderExecutor(position, account=account)`	Creates orders on a broker account to reach a target position; `create_orders(view_only=True)` previews without sending. See Orders
disposition stock	`executor.show_alerting_stocks()`	A Taiwan-market stock under disposition that requires cash or shares to be deposited at the broker before it can be traded; the method lists them among new orders. See Orders

Term	In code	Meaning
DataFrame	`close = data.get('price:adj_close')` after `data.set_market('us')`	The tabular data structure from pandas. In FinLab data, each row is a trading day and each column a stock. See Download data
FinlabDataFrame	`FinlabDataFrame`	A custom class extending pandas DataFrame: it aligns dates automatically in operations and adds methods such as technical indicators. See FinlabDataFrame
universe	`with data.us_universe(index='SP500'):`	Limits the data downloaded inside the block to the selected stocks. See Download data
lookahead bias		A backtest using data that wasn't available yet. A FinLab dataset's dates are the days the data was published and available to you, and `sim()` trades on the next trading day after a signal, so write rules directly. See Backtesting

Term	In code	Meaning
annual return	`annual_return`	Annualized rate of return for the strategy: the average gain per year. In code it is `report.metrics.annual_return()`; in `report.get_stats()` the field is `cagr`. See Backtesting
Sharpe ratio	`daily_sharpe`	Risk-adjusted return; higher is better. See Backtesting
maximum drawdown	`max_drawdown`	Largest peak-to-trough decline in portfolio value. See Backtesting
win rate	`win_ratio`	The share of trades that made money. See Backtesting
MAE / MFE	`report.display_mae_mfe_analysis()`	Each trade's maximum adverse (loss) and favorable (profit) movement while held, used to set stops. See Analysis modules
Edge ratio	`sim(..., mae_mfe_window=...)`	The MFE/MAE ratio at different holding periods; above 1 means gains outweigh losses. See Analysis modules

Standard terms in the docs

When writing docs, asking questions or searching, use the right-hand terms.

Avoid	Use instead	Reason
stop-profit	take profit	Industry-standard term
dataframe	DataFrame	First letters should be capitalized (following pandas convention)
buy, sell (for signals)	entry, exit	"Entry/exit" are more precise
re-balance, roll over	rebalance	Unified terminology
back test	backtest	Follow industry standard

Writing guidelines:

Use the terms above consistently in prose: stop loss, take profit, entry, exit, rebalance, backtest.
Keep code parameters exactly as the API spells them: stop_loss=0.1, take_profit=0.2, resample='M'.
Format code identifiers as code on every mention: sim(), FinlabDataFrame, position.

Was this page helpful?