Linear Regression

Futures Intermediate Singapore SGX FTSE China A50 Index Futures SGX Nikkei 225 Index Futures SGX MSCI Singapore (SiMSCI) Index Futures SGX Commodity Futures (Iron Ore, Rubber)

Trending Markets with Statistical Edge

Learn this and Singapore-market strategies in depth — one-time purchase, lifetime access.
Unlock full hub →

Quick Reference

Strategy Type Statistical Trend Following and Mean Reversion System
Market Outlook Trending Markets with Statistical Edge
Risk Profile Defined Risk Using Channel Boundaries or ATR
Reward Profile Targets Based on Regression Channel Width
Time Horizon Swing to Positional Trading (5-30 days typical)
Capital Requirement Medium (Futures Margin Required)
Margin Type Initial + Maintenance Margin (SGX-DC) for Futures
Best Used When Price bounces from regression channel boundaries in trending market

Payoff Profile

Linear Regression analysis fits a straight line through price data using least squares method, creating a statistically derived trend line. The regression channel adds standard deviation bands above and below the line, providing dynamic support/resistance levels. Price tends to oscillate around the regression line, creating mean reversion and trend following opportunities.

Singapore Market Details

Sgx Applicability Excellent for SGX FTSE China A50, Nikkei 225, MSCI Singapore (SiMSCI) and liquid SGX commodity futures
Mas Compliance Regulated by the Monetary Authority of Singapore (MAS) under the Securities and Futures Act (SFA); SGX-DC clearing rules apply. Futures are Specified Investment Products (SIPs), so retail clients must clear a Customer Knowledge Assessment (CKA) or complete the SGX online education programme before trading.
Contract Sizes US$1 x index per contract (USD-denominated, cash-settled) • JPY 500 x index per contract (Mini contract is JPY 100 x index) • SGD 100 x index per contract; minimum tick 0.05 index points (SGD 5)
Trading Hours SGX FTSE China A50: T (day) session 9:00 AM - 4:30 PM SGT and T+1 (night) session 4:45 PM - 5:15 AM SGT (next day). Singapore observes no daylight saving (SGT is fixed at GMT+8); session times vary by product, so confirm with SGX/your broker.
Expiry Considerations SGX index futures use monthly/quarterly cycles (no proliferation of weekly expiries). For FTSE China A50 the last trading day is the second-last China business day of the contract month, cash-settled to the official close; roll positions before expiry for positional trades.
Tax Implications Singapore has no capital gains tax, so trading gains are generally non-taxable for individual investors. However, if trading is systematic, frequent and profit-seeking such that IRAS assesses it as 'carrying on a trade or business' (via the badges of trade), profits are taxable as income (progressive rates up to 24% for individuals, 17% for companies). Only trading losses (not capital losses) are deductible. SGX derivatives carry no securities or commodities transaction tax; only brokerage commissions and SGX clearing fees apply.
Liquidity Notes SGX FTSE China A50 and Nikkei 225 futures have excellent liquidity across both day and night sessions; SiMSCI is the local benchmark proxy with good depth though lower volume than A50

Frequently Asked Questions

What is linear regression in trading?

Linear regression fits a best-fit straight line through price data using statistical methods. It creates an objective trend line (not subjectively drawn) that shows the average rate of price change. The regression channel adds standard deviation bands for overbought/oversold levels.

What does R² mean and why does it matter?

R² (R-squared) measures how well the regression line fits the data, from 0 to 1. Higher R² means stronger trend. R² > 0.75 = strong trend, R² > 0.65 = tradeable, R² < 0.50 = avoid trend trades. It's a quality filter for signals.

Where are the best entry points?

Best entries are at channel boundaries in the direction of the slope. In uptrend (positive slope), buy at lower channel. In downtrend (negative slope), sell at upper channel. These represent mean reversion opportunities within the trend.

What channel width should I use?

2.0 standard deviations is standard for channel trading. 1.5 SD gives more signals but more noise. 2.5 SD gives fewer but stronger signals. Start with 2.0 SD and adjust based on your instrument and trading style.

How do I set stops with regression?

Standard stop is beyond the channel boundary (lower channel - half SD for longs). This invalidates the trade if the channel breaks. Alternatively, use ATR-based stops for consistent risk measurement across trades.

How do I use multiple regression periods?

Use short period (10-15) for entry timing, medium (20-30) for swing trades, long (50-100) for major trend. Trade when all align: enter on short period channel bounce in direction of all timeframes for highest probability.

What does slope change tell me?

Steepening slope = trend accelerating (strong). Flattening slope = trend decelerating (warning). Slope reversal (crosses zero) = major trend change. Monitor slope changes to anticipate trend shifts before they complete.

How do filters improve regression trading?

Add: ADX > 25 (confirms trend), Volume above average (participation), RSI 30-70 (avoid extremes), Candle confirmation (timing). More filters passing = higher conviction. Score setups and adjust position size accordingly.

How should I size positions based on R²?

High R² (>0.80) = standard or larger position (reliable trend). Moderate R² (0.65-0.80) = normal position. Lower R² (0.50-0.65) = reduced position (50-75%). Below 0.50 = skip (no reliable trend).

Which exit strategy works best?

Depends on trend strength. Strong trend (high R², steep slope): opposite channel or trailing exit. Moderate trend: regression line target. Uncertain: partial exits. Testing showed opposite channel captures most profit in strong trends.

What is R² divergence and how do I use it?

R² divergence occurs when price makes new highs/lows but R² is falling. This indicates trend losing statistical support despite price movement. It's an early warning to reduce positions and tighten stops before visible price reversal.

How do I implement adaptive regression periods?

Adaptive Period = Base Period × (Average ATR / Current ATR). High volatility shortens period (more responsive), low volatility lengthens it (more stable). This keeps sensitivity consistent across different market conditions.

What's the optimal configuration for the SGX FTSE China A50?

Testing showed 20-period regression with 2.0 SD channel and R² > 0.65 threshold is optimal, with only 8% out-of-sample degradation. This provides 58% win rate and PF 1.88, balancing signals with reliability.

How do I build a regression portfolio?

Run regression on 6-8 uncorrelated instruments. Calibrate parameters per instrument or use consistent settings. Weight allocation by R² (stronger trends get more). Maximum 12-15% total portfolio heat. Track aggregate R² for market regime.

When should I pause the regression system?

Pause when: win rate < 50%, PF < 1.3 for 3+ months, R² filter stops helping, or slope signals unreliable. These indicate changing market conditions or degraded system. Investigate before resuming.

Master Singapore trading strategies on AlgoKing

Full guided lessons, quizzes, and a complete strategy library for the Singapore market. One-time purchase. No subscription, ever.

Get Singapore access →