v0.04 Kimi_Cursor

Simpleton v0.04 Kimi_Cursor

9-Indicator Dynamic Engine — User Guide

What Is This?

Kimi_Cursor v0.04 is a TradingView Pine Script strategy that fuses the best of two engines: KIMI (5-indicator voting) and CURSOR (12 strategy backtester). It adds ADX regime detection, timeframe-adaptive TP/SL, and a Dynamic mode with soft-weighted voting.

It is the first cross-engine Simpleton variant — designed to beat all timeframes by adapting its behavior to market conditions.

Files

FileTypePurpose
Simpletonv0.04_Kimi_Cursor.pinePine Script v5 StrategyLoad into TradingView to trade or backtest
backtest_kimi_cursor_v04.pyPython ValidatorRun backtests locally, compare to KIMI baseline
simpleton-kimi-cursor-v04-guide.htmlHTMLThis guide

Quick Start (TradingView)

  1. Open TradingView and navigate to BTCUSD (any exchange)
  2. Click Pine Editor at the bottom of the chart
  3. Copy the entire contents of Simpletonv0.04_Kimi_Cursor.pine and paste into the editor
  4. Click Add to Chart
  5. The strategy will appear with default settings: Dynamic mode, Hybrid TP/SL
  6. Open the Strategy Tester tab to see backtest results
Recommended first test: BTCUSD on the 1H or 45m timeframe. These showed the strongest v0.04 improvement over KIMI v0.01 (PF 1.099 vs 0.975).

Strategy Modes

Select a mode from the Strategy Mode dropdown in Settings → Inputs.

Dynamic (Recommended)

Regime-adaptive soft-weighted voting. Boosts trend-following indicators in trends, mean-reversion in ranges. The "beat all timeframes" mode.

Threshold: 4 (TREND/RANGE), 3 (MIXED)

Multi-Indicator

All 9 indicators vote equally. Min 3 confirmations needed (33% agreement). Like KIMI’s Multi mode but with 4 extra indicators.

Individual Strategies

Run any of the 9 indicators solo: RSI-2, SuperTrend, MACD, Triple EMA, BB, Ichimoku, HMA, SFP, or Z-Score.

The 9 Indicators

From KIMI (Original 5)

IndicatorTypeBuy SignalSell SignalRegime
RSI-2Mean ReversionRSI(2) < 30 + 200 EMA filterRSI(2) > 70 + 200 EMA filterRANGE 2×
SuperTrendTrend FollowingDirection flip to supportDirection flip to resistanceTREND 2×
MACDTrend FollowingMACD crosses above signalMACD crosses below signalTREND 2×
Triple EMATrend FollowingFast > Med > Slow alignmentFast < Med < Slow alignmentTREND 2×
Bollinger BandsMean ReversionClose below lower bandClose above upper bandRANGE 2×

From CURSOR (New 4)

IndicatorTypeBuy SignalSell SignalRegime
Ichimoku CloudTrend FollowingTK cross up + above cloud + bullish cloudTK cross down + below cloud + bearish cloudTREND 2×
HMA TrendTrend FollowingHMA turns up + close > SMA50 + RSI < 75HMA turns down + close < SMA50 + RSI > 25TREND 2×
Swing Failure (SFP)UniversalSweeps low, closes above + wick ≥ 25%Sweeps high, closes below + wick ≥ 25%Always 1×
Z-Score MRMean ReversionZ-Score < −2.0 (2 std devs cheap)Z-Score > 2.0 (2 std devs expensive)RANGE 2×

ADX Regime Detection

The engine classifies every bar into one of three regimes using the Average Directional Index (ADX):

RANGE
ADX < 15
MIXED
15–25
TREND
ADX > 25

How It Affects Voting (Dynamic Mode)

RegimeTrend-Following (5)Mean-Reversion (3)SFPMax VotesThreshold
TREND2× weight1× weight144
RANGE1× weight2× weight124
MIXED1× weight1× weight93
Soft weighting, not hard blocking. Unlike hard regime gating (which mutes indicators completely), v0.04 uses soft weighting. Mean-reversion indicators still contribute 1× in TREND — they catch dip-buy opportunities. Only their priority is reduced.

Hybrid TP/SL System

Three modes available under Settings → Inputs → Risk Management → TP/SL Mode:

ModeBehaviorBest For
Hybrid (default)Fixed 3%/2% on ≤1H charts, ATR-based on 4H+ chartsAll timeframes — adapts automatically
ATR Auto-ScaleATR multiples everywhere, scaled by timeframeVolatility-first approach, longer TFs
Fixed PercentageFixed TP/SL % everywhere (KIMI v0.01 style)Simplicity, known behavior

Timeframe Profiles (ATR Auto-Scale)

TimeframeProfileTP (ATR ×)SL (ATR ×)Max Hold
≤5mSCALP1.25×0.75×8 bars
15mINTRADAY1.5×0.75×12 bars
30m–1HSWING-ID2.0×1.0×15 bars
4HSWING2.5×1.25×20 bars
DailyPOSITION3.5×1.75×12 bars
Weekly+MACRO4.5×2.0×8 bars
Max Hold Timeout is only active in ATR mode. It force-closes positions that have been open longer than the max hold period for the current timeframe. This prevents capital from being tied up in stale, directionless trades.

Dashboard

The top-right table on your chart shows real-time information:

RowShows
ModeCurrent strategy mode (Dynamic, Multi-Indicator, etc.)
RegimeTREND / MIXED / RANGE + ADX value
TF ProfileSCALP / INTRADAY / SWING-ID / SWING / POSITION / MACRO
Long VotesCurrent bullish vote count out of max possible
Short VotesCurrent bearish vote count out of max possible
TP/SLActive TP/SL levels (ATR multiples or fixed %)
ATR(14)Current 14-period ATR value
Max HoldMax bars before timeout (ATR mode only)
RSI(2)Current RSI-2 value (green if oversold, red if overbought)
Z-ScoreCurrent Z-Score (green if < −2, red if > 2)
Bars HeldHow many bars the current position has been open

Python Validator

The Python backtester mirrors the Pine Script logic and compares v0.04 against KIMI v0.01 baseline.

Requirements

pip install pandas numpy yfinance

Usage

# Run all timeframes (15m, 30m, 45m, 1H, 4H, 1D)
python backtest_kimi_cursor_v04.py

# Single timeframe
python backtest_kimi_cursor_v04.py --timeframe 4h

# Show individual strategy results
python backtest_kimi_cursor_v04.py --verbose

# Save results to JSON
python backtest_kimi_cursor_v04.py --save

Output

Produces a comparison table for each timeframe showing KIMI v0.01 (fixed 3%/2%), v0.04 Hybrid (9-indicator + hybrid TP/SL), and v0.04 Dynamic (regime-weighted + hybrid TP/SL). Metrics: trades, win rate, profit factor, return %, max drawdown, Sharpe ratio.

Data limitation: yfinance provides max ~730 days of 1H data and ~59 days of 15m data. TradingView has 9+ years. The Python results are directionally accurate but TradingView is the authoritative backtest source.

Key Backtest Results

Python validator on BTCUSD, comparing v0.04 Dynamic vs KIMI v0.01:

TimeframeKIMI v0.01v0.04 DynamicWinner
PFReturn%PFReturn%
15m0.785−9.65%0.718−11.10%KIMI
30m1.212+4.36%0.897−2.60%KIMI
45m0.975−5.20%1.099+21.48%v0.04
1H0.975−5.20%1.099+21.48%v0.04
4H1.073+5.31%0.681−18.78%KIMI
1D0.903−2.49%0.274−23.16%KIMI
v0.04 Dynamic beats KIMI on 45m and 1H — turning a −5.20% loss into a +21.48% gain (PF 1.099 vs 0.975). Load into TradingView for full 9-year validation on 4H and Daily.

Recommended Settings

Conservative (Default)

SettingValueWhy
Strategy ModeDynamicAdapts to regime automatically
TP/SL ModeHybridFixed on short TFs, ATR on long TFs
Min Confirmations333% agreement — filters noise with 9 indicators
ADX Trend Threshold25Standard ADX trend threshold
ADX Range Threshold15Standard ADX range threshold

Aggressive

SettingValueWhy
Strategy ModeMulti-IndicatorMore signals, higher trade count
Min Confirmations2Lower bar = more trades, more risk
TP/SL ModeATR Auto-ScaleVolatility-adaptive everywhere

KIMI-Compatible

To replicate KIMI v0.01 behavior exactly:

SettingValue
Strategy ModeMulti-Indicator
Min Confirmations2
TP/SL ModeFixed Percentage
Fixed TP %3.0
Fixed SL %2.0

This uses only 5 of the 9 indicators for voting (the original KIMI set) because the extra indicators will rarely fire simultaneously with the KIMI ones to reach min 2 on their own. The additional indicators are passive in this config.

What’s New vs KIMI v0.01

FeatureKIMI v0.01Kimi_Cursor v0.04
Indicators59 (+Ichimoku, HMA, SFP, Z-Score)
TP/SLFixed 3%/2%Hybrid (fixed + ATR + auto-scale)
Regime awarenessNoneADX TREND/MIXED/RANGE
VotingStatic (all equal)Soft-weighted (regime-adaptive)
Timeframe profilesNone6 profiles (SCALP–MACRO)
Max hold timeoutNoneTF-scaled (8–20 bars)
Python validatorsimpleton_backtest.pybacktest_kimi_cursor_v04.py

Setting Up Alerts

Two alert conditions are built in:

  1. KC04 Buy — fires when the active mode generates a long signal
  2. KC04 Sell — fires when the active mode generates a short signal

To create an alert: right-click the strategy on your chart → Add Alert → select the condition.

Related Links