Skip to main content

ObeliskRSI_v6_1 Strategy Analysis

Table of Contents​

  1. Strategy Overview and Design Philosophy
  2. Theory Foundation and Market Understanding
  3. Core Indicator System
  4. Buy/Sell Signal Logic
  5. Profit-Taking and Stop-Loss Design
  6. Trend Judgment and Parameter Adaptation
  7. Parameter Details
  8. Risk Management and Protection
  9. Applicable Scenarios and Pair Selection
  10. Strategy Advantages and Limitations
  11. Code Implementation Notes
  12. Live Application Advice

1. Strategy Overview and Design Philosophy​

ObeliskRSI_v6_1 is an RSI-based quantitative trading strategy released by developer Obelisk in March 2021. The core design philosophy: enter when the market shows oversold pullback, buying the dip and waiting for rebounds. Uses RSI as the primary buy/sell trigger with a dual-threshold system (bull/bear market modes) enabling dynamic parameter adjustment.

The strategy employs 5 minutes as its primary timeframe, suitable for medium-to-short-term trading, capturing intraday opportunities in cryptocurrency markets.

1.1 Core Innovation​

The strategy's key innovation is the long-period RSI-based trend classification — resampling 5-minute data to 1-hour, then determining whether the market is in bull or bear mode. Different thresholds apply in each mode:

  • Bull market: RSI drops to 35 → buy; rises to 69 → sell
  • Bear market: RSI drops to 21 → buy; rises to 55 → sell

1.2 Positioning​

A mean-reversion dip-buying strategy — enters when RSI reaches oversold levels, profits from the subsequent bounce. In trending markets, this works well; in prolonged directional moves, it underperforms.

2. Theory Foundation and Market Understanding​

2.1 Overbought/Oversold Theory​

Based on the technical analysis principle that market prices don't move indefinitely in one direction — extremes must eventually revert. RSI quantifies this deviation from average strength.

2.2 Left-Side Trading​

The strategy uses left-side trading — entering before trend reversal is confirmed, attempting to buy at or near the bottom. vs. right-side trading (waiting for confirmation). Left-side offers better entry prices but higher risk of being wrong.

The dual-threshold system mitigates left-side risk: looser thresholds in bull markets, tighter in bear markets.

2.3 Multi-Period Analysis​

5-minute primary for signal triggering, 60-minute (resampled) for trend classification. "Look long-term, act short-term" filters counter-trend trades and improves win rate.

3. Core Indicator System​

3.1 Standard RSI​

14-period RSI on 5-minute data as primary trigger.

3.2 Long-Period Resampled RSI​

12× resampling: 5-minute × 12 = 60-minute RSI. Used purely for trend classification:

  • When long-period RSI > 60 → bull market mode
  • Otherwise → bear market mode

3.3 Bull/Bear State Flag​

dataframe['bull'] = dataframe[resample_rsi_key].gt(60).astype('int')

4. Buy/Sell Signal Logic​

4.1 Buy Signal​

Dual-threshold system:

Market StateRSI Buy ThresholdMeaning
Bull (bull=1)RSI crosses below 35Moderate pullback in uptrend → buy
Bear (bull=0)RSI crosses below 21Extreme oversold in downtrend → buy
((dataframe['bull'] > 0) & qtpylib.crossed_below(dataframe['rsi'], params['bull-buy-rsi-value'])) |
(~(dataframe['bull'] > 0) & qtpylib.crossed_below(dataframe['rsi'], params['bear-buy-rsi-value']))

4.2 Sell Signal​

Market StateRSI Sell Threshold
BullRSI > 69
BearRSI > 55

5. Profit-Taking and Stop-Loss Design​

5.1 Stepped ROI​

minimal_roi = {
    "0": 0.15,    # Immediate: 15%
    "35": 0.04,   # 35 minutes: 4%
    "65": 0.01,   # 65 minutes: 1%
    "115": 0      # 115 minutes: force exit
}

5.2 Static Stop-Loss: -30%​

5.3 Custom Dynamic Stop-Loss (Key Innovation)​

Uses easeInCubic function to gradually tighten stop-loss over time:

def easeInCubic(t):
    return t * t * t

# Stop ramp: -30% initially → ~0% after custom_stop_ramp_minutes

At ramp start: allows -30% fluctuation. As time passes, stop line rises. After 110 minutes (default), stop ≈ 0 — forced exit.

6. Trend Judgment and Parameter Adaptation​

6.1 Why Dual Thresholds?​

Bull market:

  • Pullbacks are moderate, won't hit extreme oversold — looser threshold (35) catches more opportunities
  • Rising is easier — higher sell threshold (69) lets profits run

Bear market:

  • Downtrends are violent — need extreme oversold (21) to avoid catching knives
  • Rebounds are weak — lower sell threshold (55) takes profits earlier

6.2 Threshold Validation​

RSI > 60 as bull/bear divider means the long-period RSI shows more ups than downs — a reasonable "relative strength" barometer.

7. Parameter Details​

ParameterDefaultDescription
bull-buy-rsi-value35Bull market buy RSI
bear-buy-rsi-value21Bear market buy RSI
bull-sell-rsi-value69Bull market sell RSI
bear-sell-rsi-value55Bear market sell RSI
stoploss-0.30Static stop-loss
custom_stop_ramp_minutes110Dynamic stop ramp time

8. Risk Management and Protection​

8.1 Consecutive Loss Guard​

Recommended configuration:

"protections": [{
    "method": "StoplossGuard",
    "lookback_period": 720,
    "trade_limit": 2,
    "stop_duration": 720,
    "only_per_pair": true
}]

After 2 consecutive stop-losses on one pair in 720 candles, pause that pair for 720 candles.

8.2 Startup Requirements​

startup_candle_count = 240 — needs 240 historical candles.

9. Applicable Scenarios and Pair Selection​

SXP, MATIC, SUSHI, CHZ — moderate market cap, good volatility, sufficient liquidity.

BTC and ETH — too stable, not enough pullback opportunities for this strategy.

9.3 Best Scenarios​

  • Volatile ranging markets: frequent pullback opportunities
  • Moderate volatility: sufficient pullback depth AND rebound space

9.4 Worst Scenarios​

  • Prolonged one-directional bull markets: no pullbacks to buy
  • Prolonged downtrends: bottom-catching fails repeatedly

10. Strategy Advantages and Limitations​

Advantages​

  1. Trend-adaptive: Dual thresholds adapt to market conditions
  2. Risk-controlled: Dynamic stop tightens over time
  3. Dip-buying edge: Better risk/reward than chasing breakouts
  4. Few parameters: Simple, easy to tune
  5. Efficient computation: No complex indicators

Limitations​

  1. Fails in directional markets: Misses trending moves, gets stopped in sustained trends
  2. Parameter-sensitive: Thresholds significantly impact results
  3. Left-side risk: May buy too early, catch falling knives
  4. No trend filter: No clear trend direction indicator

11. Code Implementation Notes​

11.1 Helper Functions​

def easeInCubic(t): return t * t * t
def clamp(num, min_value, max_value): return max(min(num, max_value), min_value)
def clamp01(num): return clamp(num, 0, 1)

11.2 Data Processing Flow​

  1. Resample to 60 minutes
  2. Calculate 14-period RSI on resampled data
  3. Merge back, forward-fill
  4. Calculate bull/bear flag
  5. Calculate standard 14-period RSI on 5-minute data
  6. Generate signals

12. Live Application Advice​

12.1 Backtesting vs. Live Gap​

Author warns: "Backtest looks better than live." Causes:

  • Backtest can't simulate slippage
  • Assumes instant fill at signal price
  • Can't model extreme liquidity crunches

12.2 Capital Management​

  • Per-pair position: 5-10% of total capital
  • 5-10 pairs simultaneously
  • Reserve emergency funds

12.3 Continuous Optimization​

Author welcomes community improvements:

  • Volatility-adjusted thresholds
  • Trend strength filtering
  • Volume confirmation
  • ATR-based dynamic stops
  • Multi-pair correlation consideration

Summary​

ObeliskRSI_v6_1 is a cleanly designed RSI dip-buying strategy. Its core highlights are the long-period RSI for trend classification and dynamic parameter switching, plus a carefully designed dynamic stop-loss mechanism.

Best for ranging markets with moderate-volatility pairs. Not suitable for sustained trending markets or ultra-stable coins. Thorough backtesting and paper trading are essential before going live.

Strategy Author: Obelisk (brookmiles) Document Date: 2026-03-27