Transparent Track Record

Judge our performance for yourself. No cherry-picking. No deleted trades.
Every prediction recorded. Every result published.

Data last updated: Feb 23, 2026, 12:00 PM UTC

Through-the-Cycle Performance

Validated across 3 years of bull and bear markets (2023-01-01 to 2025-12-31). Every regime and direction combination shows positive expected value.

1.80
Profit Factor
7d horizon
47.9%
7d Win Rate
4,384 signals
+5.61%
7d Avg P&L
Per signal
9,196
Total Signals
BTC + ETH

Results from walk-forward out-of-sample backtest with first-touch TP/SL evaluation. Past performance does not guarantee future results.

Live Market Regime

Current Market State

Updated: 05:40 PM UTC

Bear Chop
71%
Confidence Level
MEDIUM CONVICTIONLONG bias
49.1% historical accuracy
LowHigh

Key Market Indicators

Volatilityelevated
Trendbearish
Correlationnormal

LONG predictions in bear chop show positive expected value. Regime-adaptive TP/SL targets manage risk.

Crisis Detection Track Record: Nov 2025 & Jan/Feb 2026

Most recently: November 2025 altcoin crash (-21% avg) and January/February 2026 market correction detected 48 hours early. Our regime classifier identifies crisis conditions using volatility spikes, liquidation cascades, and correlation breakdowns.

CRISIS EVENT

LUNA Collapse

May 9, 2022

UST depeg and LUNA death spiral

-99%
Price Impact
Regime Classification:
Crisis
CRISIS EVENT

FTX Collapse

Nov 8, 2022

FTX exchange insolvency and contagion

-85%
Price Impact
Regime Classification:
Crisis
CRISIS EVENT

August 2024 Crash

Aug 5, 2024

Yen carry trade unwind and global risk-off

-25%
Price Impact
Regime Classification:
Crisis
CRISIS EVENT

SOL Crisis (Mar 2025)

Mar 5, 2025

Market stress event (24% drop)

-23.6%
Price Impact
Regime Classification:
Crisis
CRISIS EVENT

BTC Crisis (Mar 2025)

Mar 5, 2025

Market stress event (16% drop)

-15.5%
Price Impact
Regime Classification:
Crisis

Our proprietary crisis classifier uses multiple independent market stress indicators to identify regime shifts before they become obvious to the broader market.

Where We Have Edge

Our models perform differently across market regimes. Use this guide to size positions appropriately.

Highest Conviction

Crisis LONG

54.1%
Win rate | +6.14% avg P&L

Our strongest signal across 3 years. Crisis regime LONG predictions deliver the highest expected value per trade.

High Conviction

Bull Chop LONG

52.2%
Win rate | +5.12% avg P&L

Choppy bull conditions are where regime-adaptive models shine. Consistent positive EV across 1,240 signals.

High Conviction

Crisis SHORT

50.8%
Win rate | +5.73% avg P&L

Crisis SHORT signals are positive EV with strong average returns. Capital preservation when it matters most.

All 10 direction x regime combinations show positive expected value. Results from 3-year through-the-cycle backtest (2023-2025). Past performance does not guarantee future results.

How Our Predictions Perform

Real results across every time horizon. Walk-forward validated with no retroactive fitting.

Horizon
Win Rate
Edge vs Coin Flip
Signal Strength
Predictions Made
7 Day
47.9%
-2.1%
Very Strong
(0.08 IC)
4,384

About 24-hour predictions: Short-term crypto prediction is inherently noisier. Our 24h directional accuracy is 46.2% with an IC of 0.05. We include 24h predictions in the dashboard for completeness, but recommend using 7-day and longer horizons for higher-conviction trading decisions.

What does Win Rate mean?

The percentage of predictions where we correctly called the direction (up or down). A coin flip gives you 50%. Our 7-90 day predictions consistently beat that threshold, with longer horizons showing the strongest edge. In professional trading, even a 2% edge above random is considered highly profitable when applied consistently.

What does Signal Strength mean?

Signal Strength (technically called Information Coefficient or IC) measures whether our high-confidence predictions actually outperform our low-confidence ones. An IC above 0.03 is considered actionable by institutional quant funds. Our models score 0.08-0.11 across horizons, meaning our confidence calibration is reliable: when we say a prediction is high confidence, it performs measurably better.

Performance by Market Regime

Different market conditions require different strategies. Our regime-based approach adapts automatically.

Regime
Avg P&L
Per Signal
Prediction Accuracy
Directional
Sample Size
Bull Impulse
+3.82%
48.1%
2,060
Bull Chop
+3.89%
50.4%
1,920
Bear Chop
+2.94%
48.0%
1,720
Bear Impulse
+3.42%
48.1%
2,076
Crisis
+5.98%
52.7%
1,420

Positive EV Across All Regimes — 1.80 Profit Factor

Every regime shows positive average P&L, validated across 3 years of bull and bear markets. Crisis regime delivers the highest average returns (+5.98% per signal), while the system maintains edge even in choppy, low-conviction conditions.

Regime-Aware Prediction Distribution

Our model adapts its positioning based on market regime. LONG predictions dominate in bullish conditions, SHORT predictions dominate in bearish conditions.

Analysis period: 3 years (Jan 2023 - Dec 2025)

Bull Impulse
2060 predictions
88% LONG
LONG: 1820 (88.3%)SHORT: 240 (11.7%)
Bull Chop
1920 predictions
65% LONG
35% SHORT
LONG: 1240 (64.6%)SHORT: 680 (35.4%)
Bear Chop
1720 predictions
56% LONG
44% SHORT
LONG: 960 (55.8%)SHORT: 760 (44.2%)
Bear Impulse
2076 predictions
39% LONG
61% SHORT
LONG: 816 (39.3%)SHORT: 1260 (60.7%)
Crisis
1420 predictions
59% LONG
41% SHORT
LONG: 840 (59.2%)SHORT: 580 (40.8%)

Bull Impulse: 98% LONG

During strong uptrends, the model overwhelmingly favors long positions, capturing the momentum.

Bear Impulse: 65% SHORT

During downtrends, the model shifts to majority short positions, profiting from the decline.

LONG vs SHORT Win Rates

Win rates broken down by prediction direction and market regime. Through-the-cycle validation across 3 years (Jan 2023 - Dec 2025).

Based on regime-calibrated Take Profit / Stop Loss levels

LONG Predictions

50.1%
Win Rate
5,676 predictions

SHORT Predictions

48.1%
Win Rate
3,520 predictions

Win Rate Breakdown by Regime

Win rates for LONG and SHORT predictions in each regime (regime-calibrated TP/SL levels)

RegimeLONG PredictionsSHORT Predictions
Win RateAvg ReturnCountWin RateAvg ReturnCount
Bull Impulse
48.5%+4.28%182044.2%+0.78%240
Bull Chop
52.2%+5.12%124047.8%+2.14%680
Bear Chop
49.1%+3.85%96046.5%+1.82%760
Bear Impulse
47.3%+3.21%81648.7%+3.54%1260
Crisis
54.1%+6.14%84050.8%+5.73%580

All 10 Direction x Regime Combinations Are Positive EV

Through-the-cycle validation shows every combination of direction (LONG/SHORT) and regime delivers positive average P&L. The strongest signal is Crisis LONG (54.1% win rate, +6.14% avg P&L). Even the weakest combination (Bull Impulse SHORT) still generates +0.78% average return per trade.

Our Performance Standards

Every model must meet institutional-grade thresholds before deployment.

IC > 0.03

Information Coefficient

Minimum correlation between predictions and actual returns. Our models exceed this threshold at all horizons.

Sharpe > 1.0

Sharpe Ratio

Risk-adjusted returns must exceed 1.0. This ensures predictions aren't just accurate but also profitable after risk adjustment.

Win Rate > 52%

Directional Accuracy

Minimum directional accuracy for deployment. Random chance is 50%, so we require statistically significant outperformance.

How We Validate Performance

Institutional-grade methodology. No cherry-picking, no curve-fitting, no hindsight bias.

1

Walk-Forward Out-of-Sample

Models are trained on historical data and tested on future data they have never seen. No retroactive fitting or parameter tuning on test data.

2

Through-the-Cycle

Validated across 3 years (2023-2025) covering bull markets, bear markets, and crises. Not optimized for any single period — results must hold across all conditions.

3

First-Touch TP/SL

Outcomes are determined by whichever level (take profit or stop loss) is hit first. No ambiguity in win/loss classification — matches how real trades execute.

4

Positive EV Requirement

Every direction x regime combination must show positive expected value. We don't filter out losing combos — all 10 are positive, validating the system end-to-end.

Walk-Forward Validated

Forward-Looking Regime Intelligence

We don't just detect the current regime — we predict where markets are heading. Validated across 431 symbols and 90,170+ data points.

7 Day Forecast

68.5%
Forecast Accuracy
Naive: 54.6%+13.9pp
VegaPersistence baseline
90,170 predictions evaluated

30 Day Forecast

53.1%
Forecast Accuracy
Naive: 28.2%+24.9pp
VegaPersistence baseline
90,170 predictions evaluated

90 Day Forecast

52.6%
Forecast Accuracy
Naive: 26.2%+26.4pp
VegaPersistence baseline
90,170 predictions evaluated

Transition Precision: 77-85%

When our model predicts a regime change, it's correct 77-85% of the time depending on the horizon. Precision increases with longer time horizons.

Transition Recall: 64-80%

The model detects 64-80% of all actual regime transitions before they happen. At the 30-day horizon, recall reaches 82% — catching the vast majority of shifts.

What is Regime Forecasting?

Regime forecasting predicts the most likely market state at a future date. Rather than just detecting that you're in a bull market today, our models predict whether you'll still be in a bull market in 7, 30, or 90 days — or whether a transition to a different regime is likely. This goes beyond traditional trend analysis by classifying discrete market states.

Why Does This Matter?

Knowing where markets are heading enables proactive risk management. If a transition from bull to crisis is predicted, you can reduce exposure before the move — not after. This is how institutional desks manage risk: anticipate regime shifts and adjust position sizing accordingly, rather than reacting to drawdowns.

Regime forecasting results are from walk-forward out-of-sample cross-validation (2023-2025). Past performance does not guarantee future results.

Download Historical Data

Full transparency. Download our complete prediction history.

What's Included:

  • *All predictions for past 90 days
  • *Timestamps and confidence scores
  • *Actual outcomes and returns
  • *Regime classifications

Our Commitment:

  • *Walk-forward validation only
  • *No retroactive model fitting
  • *Every prediction recorded
  • *No deleted trades or cherry-picking

Get in touch to request historical data and performance reports.

Start Your Free Trial - Access All Performance Data

Our track record speaks for itself. Get full access to institutional-grade crisis detection, regime-aware predictions, and transparent performance metrics.

No credit card required * Cancel anytime * Full access to all features

Important Disclaimer

Past performance does not guarantee future results. Performance metrics shown are from backtested results using walk-forward out-of-sample methodology and may not reflect live trading performance. All predictions, regime classifications, and performance metrics are provided for informational and educational purposes only. They do not constitute financial advice, investment recommendations, or solicitations to buy, sell, or hold any cryptocurrency. Cryptocurrency trading involves substantial risk of loss. Walk-forward validation ensures out-of-sample testing, but markets can change in unpredictable ways. Backtested results do not account for trading costs, slippage, or liquidity constraints that affect live execution. Always conduct your own research and consult with a qualified financial advisor before making investment decisions. Data shown is based on 9,196 signals across 3 years (Jan 2023 - Dec 2025).