Files
winning-wallet-finder_github/research/lagbot.py
T
2026-07-23 16:54:53 -04:00

575 lines
24 KiB
Python

#!/usr/bin/env python3
"""lagbot — T9 Stage-2: same-event lead-lag at REAL execution.
PAPER ONLY. No keys, no orders, no bot imports — self-contained (baked
into the wwf-lagbot image). Stage-1 (research/event_leadlag.py) read
+$9.73/$100 buying name-matched siblings at their STALE LAST PRINT after
a leader burst — an entry no real order can have. This box answers the
only question that matters: when the leader moves, is the sibling's
STANDING ASK still stale (edge real) or already repriced (mirage)?
Two legs per episode, both recorded:
observational ask premium over the sibling's stale last print at burst
time (median premium >= 8c over 3 days = mirage,
pre-registered kill-switch — no paper sample needed)
paper $100 FAK walk at the real ask, gated by the pre-declared
premium cap (+4c over stale — derived from Stage-1's
~5c/share measured edge, frozen here, not tuned)
Timing realism is the instrument: RTDS burst detection ~1s + book read
~300ms = the reaction a real bot would have. THE VERDICT BINDS TO THIS
STUDY'S PRE-REGISTRATION ISSUE; entries graded nightly against chain
truth (grade_lag.py). Any semantics-altering fix bumps SEM_VER and
restarts the shakedown clock.
FROZEN spec (from Stage-1 + pre-declared; do not tune here):
burst leader print moves >= 10c within 120s · leader = most-printed
asset of its (event, outcome-name) group · cooldown 600s/event
siblings <= 2 name-matched (same lowercased outcome; yes/no/over/under
excluded) · down-moves target the sibling's complement token
entry $100 FAK walk at asks <= min(stale_print + 0.04, 0.95);
entry-side price band [0.05, 0.95]; fee 3%*sh*min(p,1-p)
Live-only plumbing (never signal): 30-min warmup, worker pool + bounded
queue (throttle counted), map/print trims, registry-style settle passes.
"""
import json
import os
import queue
import ssl
import threading
import time
import urllib.request
import websocket
WS_URL = "wss://ws-live-data.polymarket.com"
CLOB = "https://clob.polymarket.com"
STATE = os.environ.get("LAG_STATE", "/data/lag_state.json")
ATTEMPTS = os.environ.get("LAG_ATTEMPTS", "/data/lag_attempts.jsonl")
SETTLES = os.environ.get("LAG_SETTLES", "/data/lag_settles.jsonl")
SEM_VER = "l1" # bump on ANY semantics-altering change
# ── FROZEN — Stage-1 params + pre-declared execution rule ──────────────────
MOVE_C = 0.10
MOVE_WIN = 120
COOLDOWN_S = 600 # per event
MAX_SIBS = 2
PREMIUM_CAP = 0.04 # ask <= stale + 4c (from Stage-1 edge)
STAKE = 100.0
BAND = (0.05, 0.95)
FEE_RATE = 0.03
SKIP_OUTCOMES = {"yes", "no", "over", "under", ""}
# ── LIVE-ONLY plumbing ─────────────────────────────────────────────────────
WARMUP_S = 1800
BOOK_WORKERS = 4
BOOK_QUEUE = 16
CLOB_PASS_CAP = 25
SETTLED_TRIM = 10000
SKIPS_TRIM = 500
MAP_IDLE_TRIM_S = 86400 # drop groups quiet > 24h
TOP_LEVELS = 5
BID_LEVELS = 3
SSL_CTX = ssl._create_unverified_context()
def log(m):
print(f"{time.strftime('%H:%M:%S')} {m}", flush=True)
def get_json(url, timeout=8):
req = urllib.request.Request(url, headers={"User-Agent": "Mozilla/5.0"})
with urllib.request.urlopen(req, timeout=timeout, context=SSL_CTX) as r:
return json.loads(r.read().decode())
def walk_asks(asks, cap, stake_usd=STAKE):
"""FAK against a CLOB ask list within cap; partial fills kept; top raw
levels recorded for offline replays. asks arrive UNSORTED."""
lv = sorted(((float(a["price"]), float(a["size"])) for a in asks or []))
top = [[p, s] for p, s in lv[:TOP_LEVELS]]
if not lv:
return {"filled": False, "best_ask": None, "top": top}
best = lv[0][0]
spent = shares = 0.0
levels = 0
for pxx, sz in lv:
if pxx > cap or spent >= stake_usd - 1e-9:
break
take = min(sz, (stake_usd - spent) / pxx)
if take <= 0:
break
shares += take
spent += take * pxx
levels += 1
if not shares:
return {"filled": False, "best_ask": best, "top": top}
return {"filled": True, "vwap": spent / shares, "shares": shares,
"usd": spent, "levels": levels, "best_ask": best, "top": top,
"partial": spent < stake_usd - 1e-9}
def fresh_state():
return {
"sem_ver": SEM_VER, "boot_ts": int(time.time()),
"open": {}, "settled": [], "skips": [],
"pnl_realized": 0.0,
"counters": {"bursts": 0, "sib_evals": 0, "attempts": 0, "fills": 0,
"craters": 0, "premium_skips": 0, "throttle_skip": 0,
"book_fail": 0, "settled_total": 0, "settle_clob": 0,
"skip": {"warmup": 0, "no_sibs": 0, "no_stale": 0,
"band": 0, "no_complement": 0}}}
class LagBot:
def __init__(self):
self.state = fresh_state()
if os.path.exists(STATE):
try:
loaded = json.load(open(STATE))
if loaded.get("sem_ver") != SEM_VER:
log(f"⚠ SEM_VER {loaded.get('sem_ver')} -> {SEM_VER} — "
f"shakedown clock restarts here")
loaded["sem_ver"] = SEM_VER
self.state = loaded
log(f"resumed: pnl {self.state['pnl_realized']:+.2f} · "
f"{len(self.state['open'])} open")
except Exception as e:
log(f"⚠ state load failed ({e}) — fresh book")
# live maps (rebuilt from the stream; not persisted — warmup covers)
self.groups = {} # (event, outc) -> {asset: [print_count, last_ts]}
self.cond_tok = {} # cond -> {outc: asset}
self.asset_info = {} # asset -> (event, outc, cond, title)
self.last_px = {} # asset -> (ts, px)
self.lead_hist = {} # leader asset -> [(ts, px)] trailing MOVE_WIN
self.last_burst = {} # event -> ts
self.boot = time.time()
self.lock = threading.Lock()
self.book_q = queue.Queue(BOOK_QUEUE)
# ── stream ─────────────────────────────────────────────────────────────
def on_trade(self, p):
ev = p.get("eventSlug") or ""
outc = (p.get("outcome") or "").lower()
asset = str(p.get("asset") or "")
cond = p.get("conditionId") or ""
try:
px = float(p.get("price") or 0)
except (TypeError, ValueError):
return
if not (ev and asset and cond and 0 < px < 1):
return
now = time.time()
with self.lock:
self.last_px[asset] = (now, px)
self.cond_tok.setdefault(cond, {})[outc] = asset
if outc in SKIP_OUTCOMES:
return
g = self.groups.setdefault((ev, outc), {})
st = g.setdefault(asset, [0, now])
st[0] += 1
st[1] = now
if asset not in self.asset_info:
self.asset_info[asset] = (ev, outc, cond,
(p.get("title") or "")[:60])
# leader = most-printed asset of the group
leader = max(g.items(), key=lambda kv: kv[1][0])[0]
if asset != leader:
return
hist = self.lead_hist.setdefault(asset, [])
hist.append((now, px))
while hist and hist[0][0] < now - MOVE_WIN:
hist.pop(0)
base = hist[0][1]
mv = px - base
c = self.state["counters"]
if abs(mv) < MOVE_C:
return
if now - self.boot < WARMUP_S:
c["skip"]["warmup"] += 1
return
if now - self.last_burst.get(ev, 0) < COOLDOWN_S:
return
self.last_burst[ev] = now
c["bursts"] += 1
sgn = 1 if mv > 0 else -1
sibs = [a for a in g if a != asset][:MAX_SIBS]
if not sibs:
c["skip"]["no_sibs"] += 1
return
for sib in sibs:
self._queue_sibling(ev, outc, asset, sib, mv, sgn, px, now)
def _queue_sibling(self, ev, outc, leader, sib, mv, sgn, lead_px, now):
"""Resolve the tradable target (sibling side in the leader's
direction) + its stale reference, then queue the book read."""
c = self.state["counters"]
c["sib_evals"] += 1
info = self.asset_info.get(sib)
cond = info[2] if info else None
title = info[3] if info else ""
if sgn > 0:
target, stale_src = sib, sib
else:
# down-move: buy the sibling market's complement token
toks = self.cond_tok.get(cond or "", {})
others = [a for o, a in toks.items() if a != sib]
if not others:
c["skip"]["no_complement"] += 1
return
target = others[0]
stale_src = target
lp = self.last_px.get(stale_src)
if lp is None:
# complement never printed: derive stale from the sibling side
lp_s = self.last_px.get(sib)
if lp_s is None:
c["skip"]["no_stale"] += 1
return
stale = 1 - lp_s[1]
else:
stale = lp[1]
if not (BAND[0] <= stale <= BAND[1]):
c["skip"]["band"] += 1
return
job = {"event": ev, "outc": outc, "leader": leader,
"lead_mv": round(mv, 3), "lead_px": round(lead_px, 3),
"sgn": sgn, "asset": target, "cond": cond,
"title": title, "stale": round(stale, 4),
"ts": round(now, 3)}
try:
self.book_q.put_nowait(job)
c["attempts"] += 1
except queue.Full:
c["throttle_skip"] += 1
self._skip(job, "throttle", None)
self.log_attempt({**job, "filled": False, "why": "throttle"})
# ── execution (workers; network OUT of the lock) ───────────────────────
def book_worker(self):
while True:
job = self.book_q.get()
try:
self.attempt(job)
except Exception as e:
log(f"⚠ attempt error: {e}")
def attempt(self, job):
t_req = time.time()
try:
book = get_json(f"{CLOB}/book?token_id={job['asset']}")
except Exception:
with self.lock:
self.state["counters"]["book_fail"] += 1
self._skip(job, "book_fail", None)
self.log_attempt({**job, "filled": False, "why": "book_fail"})
return
lat_ms = int((time.time() - t_req) * 1000)
cap = min(job["stale"] + PREMIUM_CAP, 0.95)
r = walk_asks(book.get("asks"), cap)
bid_top = sorted(((float(b["price"]), float(b["size"]))
for b in book.get("bids") or []),
reverse=True)[:BID_LEVELS]
premium = (round(r["best_ask"] - job["stale"], 4)
if r["best_ask"] is not None else None)
rec = {**job, "cap": round(cap, 4), "latency_ms": lat_ms,
"premium": premium, "top": r["top"],
"bid_top": [[p, s] for p, s in bid_top],
"best_ask": r["best_ask"], "filled": r["filled"]}
with self.lock:
c = self.state["counters"]
if not r["filled"]:
if r["best_ask"] is not None and r["best_ask"] > cap:
c["premium_skips"] += 1
why = "premium"
else:
c["craters"] += 1
why = "crater"
self._skip(job, why, r["best_ask"])
ba = "—" if r["best_ask"] is None else f"{r['best_ask']:.3f}"
log(f"{why.upper()} {job['title'][:36]} stale "
f"{job['stale']:.2f} ask {ba} "
f"(lead {job['lead_mv']:+.2f}, {lat_ms}ms)")
self.persist()
else:
fee = FEE_RATE * r["shares"] * min(r["vwap"], 1 - r["vwap"])
lot = {**job, "price": round(r["vwap"], 5),
"best_ask": r["best_ask"], "premium": premium,
"shares": round(r["shares"], 4),
"cost": round(r["usd"], 2), "fee": round(fee, 4),
"levels": r["levels"], "partial": r["partial"],
"latency_ms": lat_ms}
self.state["open"][f"{job['asset']}:{job['ts']}"] = lot
c["fills"] += 1
rec.update({"price": lot["price"], "shares": lot["shares"],
"usd": lot["cost"], "fee": lot["fee"],
"partial": lot["partial"]})
log(f"FILL {job['title'][:36]} @ {lot['price']:.3f} "
f"(stale {job['stale']:.2f} prem "
f"{premium if premium is not None else 0:+.3f}, "
f"lead {job['lead_mv']:+.2f}, ${lot['cost']:.2f}, "
f"{lat_ms}ms)")
self.persist()
self.log_attempt(rec)
def log_attempt(self, rec):
try:
with open(ATTEMPTS, "a") as fh:
fh.write(json.dumps(rec) + "\n")
except Exception as e:
log(f"⚠ attempts log write failed: {e}")
def _skip(self, job, why, ba): # under lock
self.state["skips"].append(
{"ts": int(job["ts"]), "asset": job["asset"], "why": why,
"stale": job["stale"], "ba": ba, "title": job["title"]})
del self.state["skips"][:-SKIPS_TRIM]
# ── settles: CLOB backstop -> nightly chain truth ──────────────────────
def settle_clob_pass(self):
with self.lock:
now = time.time()
due = list(self.state["open"].keys())
due.sort(key=lambda lid: self.state["open"][lid].get("clob_ck", 0))
due = due[:CLOB_PASS_CAP]
snap = {}
for lid in due:
lot = self.state["open"][lid]
lot["clob_ck"] = now
snap[lid] = (lot.get("cond"), lot["asset"])
verdicts = {}
for lid, (cond, asset) in snap.items():
if not cond:
continue
try:
m = get_json(f"{CLOB}/markets/{cond}")
except Exception:
continue
if not m.get("closed"):
continue
for t in m.get("tokens") or []:
if str(t.get("token_id")) == str(asset):
verdicts[lid] = 1.0 if t.get("winner") else 0.0
if not verdicts:
return
with self.lock:
n = 0
for lid, pay in verdicts.items():
lot = self.state["open"].get(lid)
if lot:
self._settle(lid, lot, pay)
n += 1
if n:
self.persist()
def _settle(self, lid, lot, pay): # under lock
pnl = round(lot["shares"] * pay - lot["cost"] - lot["fee"], 2)
self.state["pnl_realized"] = round(self.state["pnl_realized"] + pnl, 2)
c = self.state["counters"]
c["settled_total"] += 1
c["settle_clob"] += 1
rec = {**lot, "payout": pay, "provisional": True,
"settled_ts": int(time.time()), "pnl": pnl}
self.state["settled"].append(rec)
del self.state["settled"][:-SETTLED_TRIM]
del self.state["open"][lid]
try:
with open(SETTLES, "a") as fh:
fh.write(json.dumps(rec) + "\n")
except Exception as e:
log(f"⚠ settles log write failed: {e}")
word = "WON" if pay == 1.0 else "refund" if pay == 0.5 else "lost"
log(f"SETTLE {word} {lot['title'][:34]} {pnl:+.2f} · "
f"realized {self.state['pnl_realized']:+.2f}")
# ── plumbing ───────────────────────────────────────────────────────────
def persist(self):
tmp = STATE + ".tmp"
json.dump(self.state, open(tmp, "w"))
os.replace(tmp, STATE)
def trim_maps(self): # under lock (minute loop)
now = time.time()
dead = [k for k, g in self.groups.items()
if all(now - st[1] > MAP_IDLE_TRIM_S for st in g.values())]
for k in dead:
del self.groups[k]
if len(self.last_px) > 400000:
self.last_px = {a: v for a, v in self.last_px.items()
if now - v[0] < MAP_IDLE_TRIM_S}
def feed_dict(self): # under lock
stt = self.state
return {
"mode": "paper", "study": "leadlag (T9 stage-2)",
"sem_ver": SEM_VER, "updated": int(time.time()),
"boot_ts": stt["boot_ts"],
"warmup_left_s": max(0, int(self.boot + WARMUP_S - time.time())),
"frozen": {"move_c": MOVE_C, "move_win_s": MOVE_WIN,
"cooldown_s": COOLDOWN_S, "max_sibs": MAX_SIBS,
"premium_cap": PREMIUM_CAP, "stake": STAKE,
"band": BAND, "fee_rate": FEE_RATE},
"groups": len(self.groups),
"counters": stt["counters"], "pnl_realized": stt["pnl_realized"],
"open_n": len(stt["open"]),
"open": [{**l, "id": lid}
for lid, l in list(stt["open"].items())[-200:]],
"settled": stt["settled"][-200:], "skips": stt["skips"][-100:],
"note": "verdict binds to the Study D pre-registration; median "
"ask-premium >= 8c over 3 days is the observational "
"kill-switch (graded nightly from the attempts stream)"}
SUB = json.dumps({"action": "subscribe", "subscriptions": [
{"topic": "activity", "type": "trades", "filters": ""}]})
def run_conn(tag, bot):
backoff = 2
while True:
state = {"fresh": time.time()}
def on_open(ws):
ws.send(SUB)
state["fresh"] = time.time()
log(f"ws[{tag}]: connected")
def ping():
while ws.keep_running:
time.sleep(5)
try:
ws.send('{"action":"ping"}')
except Exception:
break
if time.time() - state["fresh"] > 20:
try:
ws.close()
except Exception:
pass
break
threading.Thread(target=ping, daemon=True).start()
def on_message(ws, raw):
state["fresh"] = time.time()
try:
m = json.loads(raw)
except Exception:
return
if m.get("topic") == "activity" and m.get("type") == "trades":
try:
bot.on_trade(m.get("payload") or {})
except Exception as e:
log(f"⚠ on_trade error: {e}")
try:
app = websocket.WebSocketApp(WS_URL, on_open=on_open,
on_message=on_message)
app.run_forever(sslopt={"cert_reqs": ssl.CERT_NONE})
except Exception as e:
log(f"ws[{tag}]: {str(e)[:60]}")
time.sleep(backoff + (1 if tag == "b" else 0))
backoff = min(backoff * 2, 30)
if time.time() - state["fresh"] < 60:
backoff = 2
def serve_feed(bot, port=8080):
from http.server import BaseHTTPRequestHandler, ThreadingHTTPServer
class H(BaseHTTPRequestHandler):
def log_message(self, *a):
pass
def do_GET(self):
if self.path.split("?")[0] != "/feed":
self.send_response(404)
self.end_headers()
return
with bot.lock:
body = json.dumps(bot.feed_dict()).encode()
self.send_response(200)
self.send_header("Content-Type", "application/json")
self.send_header("Access-Control-Allow-Origin", "*")
self.send_header("Cache-Control", "no-store")
self.send_header("Content-Length", str(len(body)))
self.end_headers()
self.wfile.write(body)
ThreadingHTTPServer(("0.0.0.0", port), H).serve_forever()
def selftest():
"""Boot gate — hard exit on failure."""
bot = LagBot.__new__(LagBot)
bot.state = fresh_state()
bot.groups, bot.cond_tok, bot.asset_info = {}, {}, {}
bot.last_px, bot.lead_hist, bot.last_burst = {}, {}, {}
bot.boot = time.time() - WARMUP_S - 1
bot.lock = threading.Lock()
bot.book_q = queue.Queue(2)
jobs = []
bot._queue_sibling = lambda *a: jobs.append(a) # capture
mk = lambda a, o, px, ev="ev1", cond=None: { # noqa: E731
"eventSlug": ev, "outcome": o, "asset": a,
"conditionId": cond or f"c-{a}", "price": px, "title": f"T{a}"}
# build a group: leader la (3 prints), sibling sb (1 print)
bot.on_trade(mk("la", "TeamX", 0.50))
bot.on_trade(mk("sb", "TeamX", 0.48))
bot.on_trade(mk("la", "TeamX", 0.52))
bot.on_trade(mk("la", "TeamX", 0.63)) # +13c within window
assert len(jobs) == 1 and jobs[0][3] == "sb", jobs # burst -> sibling
assert bot.state["counters"]["bursts"] == 1
bot.on_trade(mk("la", "TeamX", 0.75)) # cooldown blocks
assert bot.state["counters"]["bursts"] == 1
r = walk_asks([{"price": "0.52", "size": "300"}], cap=0.52)
assert r["filled"] and not r["partial"] and r["best_ask"] == 0.52
r = walk_asks([{"price": "0.56", "size": "300"}], cap=0.52)
assert not r["filled"] and r["best_ask"] == 0.56 # premium skip case
assert walk_asks([], 0.5)["filled"] is False
# complement resolution on down-moves
bot2 = LagBot.__new__(LagBot)
bot2.state = fresh_state()
bot2.cond_tok = {"cX": {"teamy": "tk_yes", "teamz": "tk_no"}}
bot2.asset_info = {"tk_yes": ("ev", "teamy", "cX", "t")}
bot2.last_px = {"tk_no": (0, 0.40)}
out = []
bot2.book_q = type("Q", (), {"put_nowait": lambda s, j: out.append(j)})()
bot2._skip = lambda *a: None
bot2.log_attempt = lambda *a: None
bot2._queue_sibling("ev", "teamy", "lead", "tk_yes", -0.12, -1, 0.3,
time.time())
assert out and out[0]["asset"] == "tk_no" and out[0]["stale"] == 0.40, out
log("selftest OK")
def main():
selftest()
bot = LagBot()
threading.Thread(target=serve_feed, args=(bot,), daemon=True).start()
for _ in range(BOOK_WORKERS):
threading.Thread(target=bot.book_worker, daemon=True).start()
for tag in ("a", "b"):
threading.Thread(target=run_conn, args=(tag, bot), daemon=True).start()
n = 0
while True:
time.sleep(60)
n += 1
with bot.lock:
bot.trim_maps()
if n % 2 == 0:
bot.settle_clob_pass()
if n % 5 == 0:
with bot.lock:
bot.persist()
with bot.lock:
c = bot.state["counters"]
msg = (f"paper pnl {bot.state['pnl_realized']:+.2f} · open "
f"{len(bot.state['open'])} · bursts {c['bursts']} att "
f"{c['attempts']} fill {c['fills']} prem-skip "
f"{c['premium_skips']} crat {c['craters']} · groups "
f"{len(bot.groups)} · settled {c['settled_total']}")
log(msg)
if __name__ == "__main__":
main()