Price Adjustment (qfq / hfq)

After dividends, bonus shares, or rights issues, a stock's price shows an "ex-rights / ex-dividend" gap. Price adjustment smooths historical prices to account for these events, keeping the K-line continuous both visually and computationally—otherwise moving averages, backtest returns, and the like get polluted by those price gaps.

stock-sdk supports three adjustment modes on its K-line endpoints.

The three modes

Mode	Meaning	Use case
qfq	Forward adjustment: anchored to the latest price, correcting historical prices backward	Charting, technical indicators, backtesting (most common)
hfq	Backward adjustment: anchored to the earliest price, accumulating forward	Long-term cumulative return, true gains across multiple ex-rights events
none	No adjustment: raw prices, gaps preserved	Reconstructing actual historical trade prices, reconciliation

The intuitive difference:

• Forward adjustment keeps the "current price" true; historical prices are pushed down. The right edge of the chart matches today's price—the default in most trading software.
• Backward adjustment keeps the "earliest price" true; prices grow over time. Good for measuring "how many times has it risen since listing".
• No adjustment keeps both ends raw, with a visible gap on ex-rights dates.

Forward adjustment by default

stock-sdk's K-line endpoints default to forward adjustment (qfq). Call them directly to get forward-adjusted daily K:

import { StockSDK } from 'stock-sdk'

const sdk = new StockSDK()

// Forward-adjusted by default
const qfq = await sdk.kline.cn('600519')

// Explicitly specify the adjustment mode
const hfq = await sdk.kline.cn('600519', { adjust: 'hfq' })  // backward
const raw = await sdk.kline.cn('600519', { adjust: 'none' }) // unadjusted

HK and US K-lines work the same way:

const hk = await sdk.kline.hk('00700', { adjust: 'qfq' })
const us = await sdk.kline.us('AAPL', { adjust: 'hfq' })

Parameters follow the implementation

The exact parameter name (e.g. the value set for adjust) and its default follow the final implementation; this page assumes one of qfq / hfq / none, defaulting to qfq. Whether minute K-lines support adjustment likewise follows the implementation.

Backtesting notes

The adjustment mode directly determines whether your backtest is correct, so be deliberate:

1. Backtests generally use forward adjustment (qfq). Forward-adjusted K-lines are continuous and gap-free, so technical indicators (MAs, MACD, etc.) and entry/exit logic aren't disturbed by ex-rights gaps. Since K-line endpoints default to qfq, you usually need no extra config.

2. Don't run trend strategies on unadjusted data. none has a large gap on ex-rights dates, which crossover and breakout strategies will misread as a "crash / surge", producing false signals.

3. Consider backward adjustment for long-term absolute returns. If you care about "the true cumulative return of holding for N years", hfq folds dividend and bonus-share gains into the price, closer to actual holding return than forward adjustment—but hfq prices scale up over time and shouldn't be compared directly with the current price.

4. Keep training / backtest / live modes consistent. Indicator calculation, signal detection, and backtesting should all use K-lines of the same adjustment mode, avoiding mismatches like "signals on forward-adjusted, returns on unadjusted".

5. Adjustment doesn't change volume units. Adjustment only corrects price; volume is unaffected. For volume units, see the note in Indicators & Signals.

Feeding indicator-enriched, forward-adjusted K-lines straight into the backtest engine is the most common and safest combination:

import { StockSDK } from 'stock-sdk'

const sdk = new StockSDK()

// Forward-adjusted by default + indicators attached, ready as backtest input
const kline = await sdk.kline.withIndicators('600519', {
  ma: [5, 20],
  macd: true,
})
// Hand off to the backtest engine in stock-sdk/screener