feat: data-driven autopilot profitability tuning + edge profile panel

Analysis of 410 live closed trades found the edge is real but was being
destroyed by execution: gross +$267 vs $245 fees; trades held <1h were net
negative (the <15m bucket alone: -$48 on $66 fees) while 1h+ holds carried
+$78; shorts lost $72 while longs made $94 — and both the prompt and the
engine were manufacturing those losers.

Changes, each tied to the data:
- Prompt: removed the 'MUST open at least one long AND one short every
  cycle' mandate (forced weak-signal shorts); direction is now data-driven
  with 'never open just to balance the book'. Added fee-awareness (round
  trip ≈ 0.1% notional, require expected move ≥ 3x cost) and aligned the
  hold discipline with the backend throttle.
- Forced book-balance opens now require |board z-score| ≥ 0.75 — the engine
  previously force-opened full-size 10x positions on near-neutral signals
  with hardcoded confidence 70. DirectionalCandidates now carries scores.
- Min AI-managed hold raised 45m → 60m (the 15-60m bucket still bled after
  the earlier throttle landed).
- Legacy prompt hygiene: vergex path drops long-only-era custom prompts and
  zh-era configs fall back wholesale to built-in English sections — fixes
  the two long-failing kernel prompt tests.
- New Edge Profile dashboard panel: net after fees by hold-time bucket and
  side, computed from recent closed trades, with an automatic takeaway line
  — the fee/churn regression detector, always visible. Fixed the
  HistoricalPosition timestamp types (epoch ms, not strings).
This commit is contained in:
tinkle-community
2026-07-07 19:21:49 +09:00
parent e593874912
commit c53a563bff
9 changed files with 261 additions and 42 deletions

View File

@@ -877,16 +877,24 @@ func minInt(a, b int) int {
return b
}
// DirectionalCandidates returns bullish (long) and bearish (short) candidate
// symbols from the most recent Vergex signal ranking, each ordered by upstream
// rank (strongest first). Only populated for vergex_signal coin sources, since
// that is the only source carrying a per-symbol directional bias.
func (e *StrategyEngine) DirectionalCandidates() (bullish []string, bearish []string) {
// DirectionalCandidate is a Vergex board candidate with its directional
// signal strength (the board z-score; sign follows the bias direction).
type DirectionalCandidate struct {
Symbol string
Score float64
}
// DirectionalCandidates returns bullish (long) and bearish (short) candidates
// from the most recent Vergex signal ranking, each ordered by upstream rank
// (strongest first) and carrying the signal score so callers can require a
// minimum strength. Only populated for vergex_signal coin sources, since that
// is the only source carrying a per-symbol directional bias.
func (e *StrategyEngine) DirectionalCandidates() (bullish []DirectionalCandidate, bearish []DirectionalCandidate) {
if e == nil || len(e.vergexRankingCache) == 0 {
return nil, nil
}
type ranked struct {
sym string
cand DirectionalCandidate
rank int
}
rankKey := func(r int) int {
@@ -900,20 +908,21 @@ func (e *StrategyEngine) DirectionalCandidates() (bullish []string, bearish []st
if item == nil {
continue
}
entry := ranked{DirectionalCandidate{Symbol: sym, Score: item.Score}, item.Rank}
switch strings.ToLower(strings.TrimSpace(item.Bias)) {
case "bearish", "short", "sell":
br = append(br, ranked{sym, item.Rank})
br = append(br, entry)
case "bullish", "long", "buy":
bl = append(bl, ranked{sym, item.Rank})
bl = append(bl, entry)
}
}
sort.SliceStable(bl, func(i, j int) bool { return rankKey(bl[i].rank) < rankKey(bl[j].rank) })
sort.SliceStable(br, func(i, j int) bool { return rankKey(br[i].rank) < rankKey(br[j].rank) })
for _, r := range bl {
bullish = append(bullish, r.sym)
bullish = append(bullish, r.cand)
}
for _, r := range br {
bearish = append(bearish, r.sym)
bearish = append(bearish, r.cand)
}
return bullish, bearish
}