Chris Amow
8ca624d435
The confluence engine had nothing to work with. Daily moving averages were the only level source, and they sat 163 to 697 points from price, so every cluster had exactly one member and no alert could ever fire. Two new sources, chosen for having a real following — the engine is a bet that many participants watch the same price, which is what makes a level hold: - Prior day high/low/close, from the last *closed* daily bar so mid-session the levels do not silently switch to today's own developing range. Full daily weight rather than the 0.75 average discount: a traded high is structure, not a derived average. - Session VWAP, anchored to the 18:00 ET open like the daily bars. Institutional execution is benchmarked against it, and zero-volume overnight minutes are skipped rather than dividing by zero. Both are stamped 1d, so they get their own colours to stay distinguishable from the daily averages. Prior-day levels draw as price lines, which span the chart and label the axis instead of relying on bar-index interpolation. VWAP re-prices every minute while a daily average carries hundreds of points and changes once a session, so broadcasting the whole level set on the VWAP cadence would have pushed the entire history every minute. Levels now go out as a delta that clients merge by id. Adding the levels then exposed two defects that had been invisible while nothing could cluster: - Cluster identity was sha1(side + round(center / tolerance)), and tolerance derives from ATR, so it changed every bar. The same zone was continually issued a new id, never matched the cooldown table, and the cooldown did nothing. Identity is now the set of converging levels. - Alert suppression keyed on that identity, so a level drifting in or out of a group read as a new zone. It now suppresses by proximity: two zones within an ATR are the same zone, and the strongest is the one reported. Over six replayed sessions at threshold 28 that is 247 alerts, then 54, then 40; raising the cooldown to 4h — which only affects repeats of the same area, never a genuinely new zone — gives 17 total with a worst session of 9. calibrate_alerts.py now sweeps threshold and cooldown together in one pass, since the threshold turns out to be quantised and nearly useless as a control. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
81 lines
2.8 KiB
Python
81 lines
2.8 KiB
Python
from dataclasses import dataclass
|
|
|
|
from app.analysis.confluence import Cluster
|
|
|
|
|
|
@dataclass(slots=True)
|
|
class Alert:
|
|
cluster: Cluster
|
|
message: str
|
|
|
|
|
|
@dataclass(slots=True)
|
|
class _Fired:
|
|
side: str
|
|
center: float
|
|
at: int
|
|
|
|
|
|
class AlertEngine:
|
|
"""Fires once per price zone, then stays quiet until price genuinely leaves.
|
|
|
|
Suppression is by proximity rather than cluster identity. Membership churns
|
|
constantly — a moving average drifts in and out of a group, changing the
|
|
cluster's identity while a human still sees one zone sitting at the prior
|
|
day's close. Keying on identity let every reshuffle through as a fresh
|
|
alert; keying on where the zone *is* does not.
|
|
"""
|
|
|
|
def __init__(self, min_score: float, cooldown_seconds: int = 900):
|
|
self.min_score = min_score
|
|
self.cooldown_seconds = cooldown_seconds
|
|
self._fired: list[_Fired] = []
|
|
|
|
def evaluate(
|
|
self,
|
|
clusters: list[Cluster],
|
|
current_price: float,
|
|
atr15: float,
|
|
now: int,
|
|
symbol: str,
|
|
) -> list[Alert]:
|
|
tolerance = 0.5 * atr15
|
|
if tolerance <= 0:
|
|
return []
|
|
# Two zones within an ATR of each other are the same zone as far as
|
|
# being told about them goes.
|
|
merge_distance = 2 * tolerance
|
|
|
|
# Re-arming needs both elapsed time and real separation. Time alone lets
|
|
# price oscillating on a level alert forever.
|
|
self._fired = [
|
|
entry
|
|
for entry in self._fired
|
|
if not (
|
|
now - entry.at >= self.cooldown_seconds
|
|
and abs(entry.center - current_price) > merge_distance
|
|
)
|
|
]
|
|
|
|
alerts: list[Alert] = []
|
|
# Strongest first, so when several overlapping zones qualify at once the
|
|
# one that survives suppression is the most significant.
|
|
for cluster in sorted(clusters, key=lambda item: item.score, reverse=True):
|
|
if cluster.score < self.min_score or abs(cluster.center - current_price) > tolerance:
|
|
continue
|
|
if any(
|
|
entry.side == cluster.side.value
|
|
and abs(entry.center - cluster.center) <= merge_distance
|
|
for entry in self._fired
|
|
):
|
|
continue
|
|
self._fired.append(_Fired(cluster.side.value, cluster.center, now))
|
|
direction = "BEARISH" if cluster.side.value == "resistance" else "BULLISH"
|
|
timeframes = ", ".join(dict.fromkeys(member.tf.value for member in cluster.members))
|
|
message = (
|
|
f"{direction} ZONE {symbol} {current_price:.2f}\n"
|
|
f"{cluster.side.value.title()} confluence {cluster.score:g} "
|
|
f"@ {cluster.low:.2f}-{cluster.high:.2f}\n{timeframes}"
|
|
)
|
|
alerts.append(Alert(cluster, message))
|
|
return alerts
|