The model estimates the probability that price hits the upper or lower barrier within the next 4h. Read the docs →
ETHUSDT_UP_ethusdt_up_tbm_4atr_0.1
higher threshold → fewer, more precise signals
| Threshold | Signals | Precision | Recall | F1 | Lift |
|---|---|---|---|---|---|
| ≥0.5 | 4,005 | 23.1% | 61.8% | 0.337 | 1.53× |
| ≥0.55Now 57.6% | 2,416 | 27.9% | 44.9% | 0.344 | 1.84× |
| ≥0.6 | 1,462 | 30.5% | 29.7% | 0.301 | 2.01× |
| ≥0.65 | 688 | 34.0% | 15.6% | 0.214 | 2.25× |
| ≥0.7 | 203 | 48.3% | 6.5% | 0.115 | 3.19× |
| ≥0.75 | 37 | 59.5% | 1.5% | 0.029 | 3.93× |
Not necessarily. A UP probability of 90% means that in historical data, market setups similar to the current one resulted in an upward move 90% of the time. It is a base rate from past observations — not a prediction or guarantee about what will happen next.
Because they are produced by two independent models, each answering a separate question: will the price cross the upper barrier? Will it cross the lower barrier? Both can occur within the same candle, neither can occur, or only one. They are not two sides of a single distribution.
Each model outputs the probability that price will cross a barrier set at 4× ATR above or below the current price, within the next 4 hours. The output is a statistical frequency derived from training data — it describes past patterns, not future certainty.
The models are trained on historical price data (OHLCV). They do not incorporate news, social sentiment, macroeconomic indicators, or any real-time external information. The output reflects purely price-based historical patterns.
Probabilities are updated on each candle close. The current models are trained on 15-minute candle data, so a new prediction is produced every 15 minutes. This cadence may vary in the future as models trained on other timeframes are introduced.
No. These outputs are one statistical data point among many. A single model trained on historical data cannot account for all variables that affect markets. Augurly does not recommend using any single tool as the sole input to a decision.
Augurly provides statistical model outputs for informational purposes only. It does not provide financial advice, and nothing on this platform should be construed as a recommendation to buy or sell any asset. All decisions and their results are solely the responsibility of the user.