TT-Autotune · data 2026-09-26
ICP trading bot: TT-Autotune tuned results by timeframe and regime
Tuned vs default Lorentzian Classification results for ICP across 5 timeframes and six market regimes, and how to automate ICP from TradingView paper first.
ICP with TT-Autotune
TT-Autotune tunes the Lorentzian Classification strategy separately for every token, chart timeframe and market regime, then switches to that regime's champion settings live inside one TradingView strategy alert. This page summarizes what the optimizer found for ICP: 5 trained timeframes (15m, 30m, 1h, 4h, 1d) and 30 regime cells, of which 27 beat the stock default settings.
On tuned ROI averaged across regimes, the strongest ICP timeframe is 4h (+82.3% mean), and its best single regime is Bull+ at +124.4%. ROI here is the sum of per-trade ROI over each cell's walk-forward window, scored with fees and slippage, not a live result.
ICP on the 15m chart
On 15m, 6 of 6 regimes beat the default settings, led by Chop, which is no trend but high volatility: big swings in both directions where breakouts often fail.
- Bull+
- tuned +19.0% vs default -2.4% (beats default by 21.4 pts)
- Bull-
- tuned +23.3% vs default -3.8% (beats default by 27.1 pts)
- Bear+
- tuned +20.1% vs default -6.7% (beats default by 26.8 pts)
- Bear-
- tuned +15.7% vs default -0.0% (beats default by 15.7 pts)
- Chop
- tuned +29.0% vs default -0.4% (beats default by 29.4 pts)
- Quiet
- tuned +26.6% vs default -9.2% (beats default by 35.9 pts)
ICP on the 30m chart
On 30m, 5 of 6 regimes beat the default settings, led by Bull+, which is a strong uptrend: ADX of 25 or more with a strong positive slope, where buyers are in control and pullbacks stay shallow.
- Bull+
- tuned +28.2% vs default +0.0% (beats default by 28.2 pts)
- Bull-
- tuned +14.6% vs default -4.4% (beats default by 19.0 pts)
- Bear+
- tuned +7.8% vs default -1.9% (beats default by 9.7 pts)
- Bear-
- tuned +12.8% vs default +0.6% (beats default by 12.2 pts)
- Chop
- tuned +0.0% vs default +0.0% (trails default by 0.0 pts)
- Quiet
- tuned +27.3% vs default -9.5% (beats default by 36.8 pts)
ICP on the 1h chart
On 1h, 6 of 6 regimes beat the default settings, led by Bear+, which is a strong downtrend: ADX of 25 or more with a strong negative slope, where sellers are in control and bounces fail.
- Bull+
- tuned +16.5% vs default +0.6% (beats default by 15.9 pts)
- Bull-
- tuned +33.2% vs default -18.9% (beats default by 52.1 pts)
- Bear+
- tuned +48.3% vs default -1.3% (beats default by 49.5 pts)
- Bear-
- tuned +35.6% vs default +1.6% (beats default by 33.9 pts)
- Chop
- tuned +42.2% vs default -1.8% (beats default by 44.0 pts)
- Quiet
- tuned +33.0% vs default -5.6% (beats default by 38.7 pts)
ICP on the 4h chart
On 4h, 6 of 6 regimes beat the default settings, led by Bull+, which is a strong uptrend: ADX of 25 or more with a strong positive slope, where buyers are in control and pullbacks stay shallow.
- Bull+
- tuned +124.4% vs default +10.1% (beats default by 114.3 pts)
- Bull-
- tuned +47.0% vs default -17.0% (beats default by 64.0 pts)
- Bear+
- tuned +90.2% vs default -11.5% (beats default by 101.7 pts)
- Bear-
- tuned +58.1% vs default -10.4% (beats default by 68.5 pts)
- Chop
- tuned +116.2% vs default +26.2% (beats default by 90.0 pts)
- Quiet
- tuned +58.0% vs default +21.1% (beats default by 36.9 pts)
ICP on the 1d chart
On 1d, 4 of 6 regimes beat the default settings, led by Chop, which is no trend but high volatility: big swings in both directions where breakouts often fail.
- Bull+
- tuned +38.3% vs default +38.3% (trails default by 0.0 pts)
- Bull-
- tuned +46.3% vs default -8.3% (beats default by 54.6 pts)
- Bear+
- tuned +87.3% vs default -3.6% (beats default by 90.9 pts)
- Bear-
- tuned +118.5% vs default -8.4% (beats default by 126.9 pts)
- Chop
- tuned +152.8% vs default -3.8% (beats default by 156.6 pts)
- Quiet
- tuned +0.0% vs default +0.0% (trails default by 0.0 pts)
How to read these ICP numbers
These are optimizer results on historical data, and every champion had to pass gates before publication: positive Sortino and ROI in its regime, beating the default on both, enough trades spread across walk-forward windows, and a plateau rather than a spike. Longer timeframes cover longer calendar windows, so compare timeframes with care. TensorTrader's forward testnet books have not yet shown an edge after fees, so treat any ICP cell as a hypothesis to paper trade, not a forecast.
Trade ICP from TradingView
- 1Connect a trade-only key on a live-ready venue that lists ICP perpetuals, on testnet first.
- 2Open the TensorTrader extension's TT-Autotune tab, select that key and browse the ICP cells.
- 3Enroll the ICP timeframes you want; each becomes one strategy alert with its regime champions injected.
- 4Watch fills and closed-trade accounting on paper before any live opt-in.
Not financial advice. Figures are TensorTrader optimizer or backtest results with the method stated; past results do not predict future returns.