{
  "approach": "Uses the supplied base's sub-tick belief estimator with a confidence pause: if estimated p changes by more than four previous diffusion-plus-posterior standard deviations, with a minimum trigger of one price tick, it cancels economic orders and sets main quotes and planned YES purchases to zero for that step. The original 0.01-share probes continue so information can recover before economic quoting resumes. Threshold selection used only seeds 1000-1599; code was frozen before holdout scoring. Tight posterior-width gates were rejected because their retail loss outweighed arbitrage saved. The holdout did not confirm the tuning gain.",
  "blocked": null,
  "python_version": "3.13.5",
  "base_mean_edge": 102.33197908899267,
  "score": {
    "mean_edge": 107.45298269594832,
    "retail_edge": 135.54694126664617,
    "arb_edge": -28.0939585707135,
    "simulations": 600,
    "seed_start": 1000,
    "failures": 0
  },
  "holdout": {
    "mean_edge": 95.74596803103867,
    "simulations": 600,
    "seed_start": 30000,
    "failures": 0
  },
  "code": "\"\"\"Stale-quote defence with tick probes, a sub-tick belief and cheap YES stock.\n\nEvery step the arbitrageur removes any order on the wrong side of p, so tiny\ntwo-sided probes around the estimate locate p to a tick; a grid filter narrows\nthat further.  Quotes sit one tick inside the competitor and are sized by the\nmarginal rule: one more share is worth quoting while the chance a retail order\nreaches it, times its edge, exceeds the expected arbitrage loss.  Near p=0 a\nretail buy pays almost its whole notional as edge, so YES is held to cover low\nasks.  That YES is bought when it is cheap: a large bid rests at tick 1 while p\nis above it, so the arbitrageur fills it on the crossing for about 0.1 tick a\nshare instead of nearly a full tick once p is near zero, and smaller bids sit on\nwhichever tick p is within 0.15 of.  Below tick 1 the estimate of p decays with\nthe time spent there, since no probe can see it.\n\"\"\"\nimport math\nfrom orderbook_pm_challenge.strategy import BaseStrategy\nfrom orderbook_pm_challenge.types import CancelOrder, PlaceOrder, Side\n\nP = dict(lam_r=0.25, m=5.0, sig_n=1.2, lam_j=0.0019, sig_j=0.35, sig=0.02,\n         qmax=3000.0, rmin=2, kA=0.7, kL=0.15, pL=0.05,\n         ammo=1.0, ydem=0.4, W=200.0, fy=0.6, ammo_q=60.0, ammo_pmax=0.05, ammo_maxpay=0.02,\n         fb=0.9, fu=0.7, xu=0.1, sticky=1.0, min_ahead=2.0, keep_hi=1.5, keep_lo=0.7, Hu=300.0, bins=8.0, filt=1.0, uni=0.01, jumpinfo=1.0, dmin=0.0, kLb=0.08, kLf=0.08, pinf=1.0, pre=0.0, H1=100.0,\n         ydf=0.6, Wf=2000.0, g_hi=0.15, g_lo=0.15, qf=400.0, pmaxf=0.05, pre_gap=1.0,\n         ydf1=0.6, g1=5.0, f1=0.6, a1=100.0, nb1=0.0, gs=1.0, p1t=1.0, m1=1.0, zd=1.0, pfl=0.0005)\nSQ2 = math.sqrt(2.0)\nSQ2PI = math.sqrt(2.0 * math.pi)\ndef cdf(x): return 0.5 * (1.0 + math.erf(x / SQ2))\ndef pdf(x): return math.exp(-0.5 * x * x) / SQ2PI\ndef inv(p):\n    p = min(max(p, 1e-9), 1 - 1e-9)\n    t = math.sqrt(-2.0 * math.log(min(p, 1 - p)))\n    x = t - (2.515517 + 0.802853 * t + 0.010328 * t * t) / (1 + 1.432788 * t + 0.189269 * t * t + 0.001308 * t ** 3)\n    x = -x if p < 0.5 else x\n    for _ in range(2):\n        d = pdf(x)\n        if d < 1e-12:\n            break\n        x -= (cdf(x) - p) / d\n    return x\ndef lpm(k):\n    return pdf(k) + k * cdf(k)\ndef fl(q):\n    return math.floor(q * 100 + 1e-9) / 100\n\nNODES = [(-2.2, .06), (-1.3, .17), (-0.6, .27), (0.6, .27), (1.3, .17), (2.2, .06)]\nPROBE = 0.01\n\n\nclass Strategy(BaseStrategy):\n    def __init__(self):\n        self.n = 0\n        self.ahead = set()     # main orders placed strictly ahead of any competitor order at their tick\n        self.res = {}          # order id -> uncovered quantity at placement\n        self.bc = None         # belief: sub-tick grid centres (ticks)\n        self.bw = None         # belief: weights\n        self.st = 0.5          # last one-step sd in ticks\n        self.pcb = self.pca = None\n        self.pb = set()\n        self.pa = set()\n        self.tb = 0\n        self.s = None          # competitor spread parameter\n        self.last_estimate = None\n        self.confidence_pause = 0\n        self.previous_predictive_sd = 1.0\n\n    def _oid(self, tag):\n        self.n += 1\n        return tag + str(self.n)\n\n    def on_step(self, state):\n        H = max(1, state.steps_remaining)\n        cb = state.competitor_best_bid_ticks or 0\n        ca = state.competitor_best_ask_ticks or 100\n        if self.s is None:\n            self.s = max(1, (ca - cb + 1) // 2)\n        lo, hi = cb, ca\n        for o in state.own_orders:\n            if o.side is Side.BUY:\n                lo = max(lo, o.price_ticks)\n            else:\n                hi = min(hi, o.price_ticks)\n        if P['pinf'] > 0:\n            sb = {o.price_ticks for o in state.own_orders if o.side is Side.BUY and o.remaining_quantity <= PROBE + 1e-9}\n            sa = {o.price_ticks for o in state.own_orders if o.side is Side.SELL and o.remaining_quantity <= PROBE + 1e-9}\n            ea = self.pa - sa\n            eb = self.pb - sb\n            if sa and ea:\n                lo = max(lo, min(max(ea), hi))\n            if eb and (sb or min(eb) <= P['p1t']):\n                hi = min(hi, max(min(eb), lo))\n        if P['jumpinfo'] > 0 and self.pcb is not None:\n            if ca > self.pca + 1:\n                lo = max(lo, ca - 1)\n            if cb < self.pcb - 1:\n                hi = min(hi, cb + 1)\n        self.pcb, self.pca = cb, ca\n        if lo > hi:\n            lo, hi = hi, lo\n        # sub-tick belief over p_(t-1): predict with one-step noise, truncate to [lo, hi]\n        wid = max(hi - lo, 1e-6)\n        nb = max(2, min(48, int(round(wid * P['bins']))))\n        cs = [lo + (i + 0.5) * wid / nb for i in range(nb)]\n        if self.bc is None or P['filt'] <= 0:\n            ws = [1.0 / nb] * nb\n        else:\n            sg = max(self.st, 0.6 * wid / nb)\n            k = -0.5 / (sg * sg)\n            ws = []\n            uni = P['uni'] * sg\n            for c in cs:\n                t = uni\n                for ci, wi in zip(self.bc, self.bw):\n                    d = c - ci\n                    d2 = d * d * k\n                    if d2 > -12.5:\n                        t += wi * math.exp(d2)\n                ws.append(t)\n            tot = sum(ws)\n            ws = [w / tot for w in ws]\n        self.bc, self.bw = cs, ws\n        mean_t = sum(c * w for c, w in zip(cs, ws))\n        var_t = max(sum(c * c * w for c, w in zip(cs, ws)) - mean_t * mean_t, 0.0)\n        phat = min(max(mean_t / 100.0, 0.002), 0.998)\n        var = P['sig'] ** 2 + P['lam_j'] * P['sig_j'] ** 2\n        sd = math.sqrt(H * var)\n        self.tb = self.tb + 1 if hi <= 1 else 0\n        if self.tb > 0 and P['zd'] > 0:\n            # below the last tick there is no probe: the score keeps walking away from it\n            phat = min(phat, max(P['pfl'], cdf(inv(0.01) - P['zd'] * 0.02 * math.sqrt(self.tb) / sd)))\n        z = inv(phat)\n        Z = z * sd\n        sp = pdf(z) / sd * P['sig']\n        self.st = sp * 100.0\n        s = math.sqrt(sp * sp + var_t / 1e4 + 1e-8)\n        sd1 = math.sqrt(max(1.0, H - 1.0) * var)\n        jp = [(cdf((Z + n * P['sig_j']) / sd1), w) for n, w in NODES]\n        lj = P['lam_j']\n\n        def AB(x):\n            k = (x - phat) / s\n            a = s * lpm(k)\n            b = a - (x - phat)\n            ja = sum(w * max(0.0, x - p) for p, w in jp)\n            jb = sum(w * max(0.0, p - x) for p, w in jp)\n            return (1 - lj) * a + lj * ja, (1 - lj) * b + lj * jb\n\n        if H == 1:\n            def AB(x):\n                return (1 - phat) * x, phat * (1 - x)\n        rho = P['lam_r'] * 0.5\n        mu = math.log(P['m']) - 0.5 * P['sig_n'] ** 2\n        kA = P['kL'] if phat < P['pL'] else P['kA']\n\n        def qsize(loss, gain, px):\n            loss *= kA\n            if gain <= 0:\n                return 0.0\n            tail = loss / (rho * gain)\n            if tail >= 1.0:\n                return 0.0\n            return min(P['qmax'], math.exp(mu + P['sig_n'] * inv(1 - max(tail, 1e-6))) / px)\n\n        ptk = phat * 100\n        def qbid(t):\n            if not 1 <= t <= 99:\n                return 0.0\n            a, b = AB(t / 100.0)\n            if phat * 100 - t < P['dmin']:\n                return 0.0\n            if P['nb1'] > 0 and self.s == 1 and t == 1 and ptk < 1.0:\n                return 0.0\n            if phat < P['pL']:\n                ytg = P['ydem'] * min(H, P['W']) / max(ca - 1, ptk + 1) * 100.0\n                a *= (P['kLb'] if state.yes_inventory < ytg else P['kLf']) / P['kL']\n            return qsize(a, b, max(phat, 0.05))\n\n        def qask(t):\n            if not 1 <= t <= 99:\n                return 0.0\n            a, b = AB(t / 100.0)\n            if t - phat * 100 < P['dmin']:\n                return 0.0\n            return qsize(b, a, t / 100.0)\n\n        bids = [o for o in state.own_orders if o.side is Side.BUY and o.remaining_quantity > PROBE + 1e-9]\n        asks = [o for o in state.own_orders if o.side is Side.SELL and o.remaining_quantity > PROBE + 1e-9]\n        sticky = P['sticky'] > 0\n        def top(orders, comp, step):\n            if sticky and sum(o.remaining_quantity for o in orders if o.price_ticks == comp and o.order_id in self.ahead) >= P['min_ahead']:\n                return comp\n            return comp + step\n        bid_t = top(bids, cb, 1)\n        ask_t = top(asks, ca, -1)\n        qb = qbid(bid_t)\n        qa = qask(ask_t)\n\n        Y = state.yes_inventory\n        if ask_t == 1 and H > P['H1'] and P['pre'] > 0:\n            qa = qa if Y >= P['a1'] > 0 else 0.0          # let the competitor's 2-tick ask erode instead of fronting it for nothing\n        # ammunition: hold YES so that low asks are covered\n        ptick = phat * 100\n        amt = 0\n        qam = 0.0\n        if P['ammo'] > 0 and ask_t >= 2 and phat < max(P['ammo_pmax'], P['pmaxf']):\n            wealth = state.cash + Y * phat\n            x = max(ask_t, ptick + 1) / 100.0\n            capw = P['fy'] * wealth / max(phat, 0.005)\n            target = min(P['ydem'] * min(H, P['W']) / x, capw)\n            full = min(P['ydf'] * min(H, P['Wf']) / x, capw) if phat < P['pmaxf'] else 0.0\n            t = max(1, min(99, int(math.ceil(ptick + 1e-9))))\n            g = t - ptick\n            if Y < full and g <= P['g_hi'] and bid_t <= t < ask_t:\n                amt, qam = t, min(P['qf'], full - Y)     # a tick p has just crossed: nearly free\n            elif Y < target and phat < P['ammo_pmax'] and bid_t <= t < ask_t and t / 100.0 - phat < P['ammo_maxpay']:\n                amt, qam = t, min(P['ammo_q'], target - Y)\n            elif Y < full and t >= 2 and 1.0 - g <= P['g_lo']:\n                amt, qam = t - 1, min(P['qf'], full - Y)  # rest just under p: filled only on the crossing, or by retail\n\n        if P['pre'] > 0 and ca == 2 and phat < 0.01 and H > P['H1'] and Y < P['pre'] and not qam and 1.0 - ptick < P['pre_gap']:\n            amt, qam = 1, min(P['ammo_q'], P['pre'] - Y)\n        q1 = 0.0\n        if P['ydf1'] > 0 and P['m1'] < ptick < P['g1'] + 1.0 and amt != 1:\n            q1 = max(0.0, P['ydf1'] * min(H, P['Wf']) / (self.s / 100.0) - Y)\n        # Reacquire confidence following an innovation outside the usual move range.\n        innovation = 0.0 if self.last_estimate is None else abs(ptick - self.last_estimate)\n        if innovation > max(1.0, 4.0 * self.previous_predictive_sd):\n            self.confidence_pause = 1\n        self.last_estimate = ptick\n        self.previous_predictive_sd = s * 100.0\n        if self.confidence_pause > 0:\n            self.confidence_pause -= 1\n            qb = qa = qam = q1 = 0.0\n        # order management\n        actions = []\n        free = state.free_cash - 2.0\n        keep_ask = 0.0\n        new_b = []\n        new_a = []\n        def cancel(o):\n            nonlocal free\n            actions.append(CancelOrder(o.order_id))\n            x = o.price_ticks / 100.0\n            if o.side is Side.BUY:\n                free += x * o.remaining_quantity\n            else:\n                free += (1 - x) * min(self.res.get(o.order_id, 0.0), o.remaining_quantity)\n            self.ahead.discard(o.order_id)\n            self.res.pop(o.order_id, None)\n        for o in state.own_orders:\n            if o.remaining_quantity <= PROBE + 1e-9:\n                cancel(o)\n        kept_ids = set()\n        def manage(orders, t, q, comp):\n            kept = 0.0\n            for o in orders:\n                if o.price_ticks == t and q > 0 and kept + o.remaining_quantity <= max(q * P['keep_hi'], 1.0):\n                    kept += o.remaining_quantity\n                    kept_ids.add(o.order_id)\n                else:\n                    cancel(o)\n            return kept\n        kb = manage(bids, bid_t, qb, cb)\n        ka = manage(asks, ask_t, qa, ca)\n        add_b = (qb - kb) if (kb < qb * P['keep_lo'] and bid_t != cb) else 0.0\n        add_a = (qa - ka) if (ka < qa * P['keep_lo'] and ask_t != ca) else 0.0\n\n        budget = P['fb'] * max(free, 0.0)\n        cost = add_b * bid_t / 100.0 + qam * amt / 100.0\n        if cost > budget:\n            f = budget / cost\n            add_b *= f\n            qam *= f\n            cost = budget\n        out = []\n        q1 = min(q1, P['f1'] * max(budget - cost, 0.0) * 100.0)\n        if q1 >= 0.01:\n            out.append(PlaceOrder(side=Side.BUY, price_ticks=1, quantity=fl(q1)))\n            free -= fl(q1) / 100.0 + 1e-6\n        if amt and qam >= 0.01:\n            q = fl(qam)\n            out.append(PlaceOrder(side=Side.BUY, price_ticks=amt, quantity=q))\n            free -= q * amt / 100.0 + 1e-6\n        if bid_t >= 1 and add_b >= 0.01:\n            q = fl(add_b)\n            oid = self._oid('b')\n            out.append(PlaceOrder(side=Side.BUY, price_ticks=bid_t, quantity=q, client_order_id=oid))\n            free -= q * bid_t / 100.0 + 1e-6\n            if bid_t > cb:\n                self.ahead.add(oid)\n        if ask_t <= 99 and add_a >= 0.01:\n            x = ask_t / 100.0\n            covk = sum(o.remaining_quantity - min(self.res.get(o.order_id, 0.0), o.remaining_quantity)\n                       for o in asks if o.order_id in kept_ids)\n            cov = max(0.0, Y - covk - 0.1)\n            unc = max(0.0, add_a - cov)\n            if unc > 0:\n                ub = max(0.0, free - (1 - P['fu']) * max(state.cash, 0.0)) if x >= P['xu'] else 0.0\n                if H < P['Hu'] and phat < 0.05:\n                    ub = max(0.0, free * (1 - H / P['Hu']))\n                unc = min(unc, ub / (1 - x))\n            q = fl(min(add_a, cov + unc))\n            if q >= 0.01:\n                oid = self._oid('a')\n                out.append(PlaceOrder(side=Side.SELL, price_ticks=ask_t, quantity=q, client_order_id=oid))\n                r = max(0.0, q - cov)\n                self.res[oid] = r\n                free -= r * (1 - x) + 1e-6\n                if ask_t < ca:\n                    self.ahead.add(oid)\n\n        # probes: tiny orders that the arbitrageur removes when p crosses them\n        f = int(math.floor(ptick + 1e-9))\n        r = max(int(P['rmin']), int(math.ceil(2.0 * s * 100)))\n        self.pb = set()\n        self.pa = set()\n        if free > 1.0:\n            for t in range(f, f + r + 1):\n                if cb < t < ca and 1 <= t <= 99:\n                    out.append(PlaceOrder(side=Side.BUY, price_ticks=t, quantity=PROBE))\n                    self.pb.add(t)\n            for t in range(f + 1 - r, f + 2):\n                if cb < t < ca and 1 <= t <= 99:\n                    out.append(PlaceOrder(side=Side.SELL, price_ticks=t, quantity=PROBE))\n                    self.pa.add(t)\n        return actions + out\n",
  "code_sha256": "3fd37a75874fd4b345a4248ca87d7077e5e5c679cbe6f14f61fd84feeef06229",
  "tried": [
    "Scorer revision b244530ee2a6247b5f930830eeca9dd950da0e7e, Python 3.13.5; python3 -m orderbook_pm_challenge run ../base.py --simulations 600 --seed-start 0 --json: mean_edge=102.33197908899267, retail_edge=127.3199510028995, arb_edge=-24.987971913912165, failures=0.",
    "Unchanged base, --simulations 600 --seed-start 1000 --workers 4 --json: mean_edge=107.2442092237515, retail_edge=135.80165652288784, arb_edge=-28.557447299150834, failures=0.",
    "Posterior SD gate >0.12 ticks (post_0.12_100.json): seeds 1000-1099, n=100, mean_edge=-2.881625466563, retail_edge=3.931531355288, arb_edge=-6.813156821837, retail_lost_vs_base=157.527675847295, arb_saved_vs_base=22.79491108837, paired_delta=-134.732764759009, paired_SE=19.590151910664314, failures=0.",
    "Posterior SD gate >0.18 ticks (post_0.18_100.json): seeds 1000-1099, n=100, mean_edge=-0.7756505250490001, retail_edge=10.064401494611, arb_edge=-10.840052019648999, retail_lost_vs_base=151.39480570797198, arb_saved_vs_base=18.768015890558, paired_delta=-132.62678981749502, paired_SE=18.999444272324276, failures=0.",
    "Observed p-bracket width gate >0.6 ticks (width_0.6_100.json): seeds 1000-1099, n=100, mean_edge=0.027744167256999987, retail_edge=0.519218954308, arb_edge=-0.491474787028, retail_lost_vs_base=160.939988248275, arb_saved_vs_base=29.116593123179, paired_delta=-131.823395125189, paired_SE=19.074631977253794, failures=0.",
    "Posterior SD gate >0.24 ticks (post_0.24_100.json): seeds 1000-1099, n=100, mean_edge=23.528914551948002, retail_edge=35.634019014319, arb_edge=-12.105104462295001, retail_lost_vs_base=125.82518818826401, arb_saved_vs_base=17.502963447912, paired_delta=-108.322224740498, paired_SE=14.687509588825316, failures=0.",
    "Posterior SD rescaled by min(1, base phat_ticks / grid mean_ticks), gate >0.18 ticks (boundary_0.18_1000_100.json): seeds 1000-1099, n=100, mean_edge=48.717726381178, retail_edge=63.25844815273, arb_edge=-14.540721771543001, retail_lost_vs_base=98.200759049853, arb_saved_vs_base=15.067346138664, paired_delta=-83.13341291126801, paired_SE=12.345088262308941, failures=0.",
    "Diffusion-plus-posterior predictive SD gate >0.25 ticks (pred_0.25_100.json): seeds 1000-1099, n=100, mean_edge=60.558769780194, retail_edge=78.243657227469, arb_edge=-17.684887447256, retail_lost_vs_base=83.21554997511402, arb_saved_vs_base=11.923180462951, paired_delta=-71.292369512252, paired_SE=10.903467183273952, failures=0.",
    "Posterior SD rescaled by min(1, base phat_ticks / grid mean_ticks), gate >0.24 ticks (boundary_0.24_1000_100.json): seeds 1000-1099, n=100, mean_edge=69.719496984974, retail_edge=89.70538422669601, arb_edge=-19.985887241681, retail_lost_vs_base=71.753822975887, arb_saved_vs_base=9.622180668526001, paired_delta=-62.131642307472, paired_SE=9.440720105906024, failures=0.",
    "Diffusion-plus-posterior predictive SD gate >0.4 ticks (pred_0.4_100.json): seeds 1000-1099, n=100, mean_edge=97.63478089706899, retail_edge=112.43504901738899, arb_edge=-14.800268120252, retail_lost_vs_base=49.024158185194004, arb_saved_vs_base=14.807799789955, paired_delta=-34.216358395377, paired_SE=6.234838448870084, failures=0.",
    "Main quote edge must clear 0.5 times predictive SD; same SD capped at 0.4 ticks (marginpred_0.5_1000_100.json): seeds 1000-1099, n=100, mean_edge=97.67743608992599, retail_edge=112.497236863704, arb_edge=-14.819800773711, retail_lost_vs_base=48.961970338879, arb_saved_vs_base=14.788267136496, paired_delta=-34.17370320252, paired_SE=6.228068137581022, failures=0.",
    "Predictive SD including base jump mixture, gate >0.9 ticks (jumpsd_0.9_1000_100.json): seeds 1000-1099, n=100, mean_edge=98.25299832034, retail_edge=117.63521975860199, arb_edge=-19.382221438266, retail_lost_vs_base=43.823987443981004, arb_saved_vs_base=10.225846471941, paired_delta=-33.598140972105995, paired_SE=5.042584764184199, failures=0.",
    "Diffusion-plus-posterior predictive SD gate >0.6 ticks (pred_0.6_100.json): seeds 1000-1099, n=100, mean_edge=105.48154368310001, retail_edge=126.41833530014101, arb_edge=-20.936791616963, retail_lost_vs_base=35.040871902442, arb_saved_vs_base=8.671276293244, paired_delta=-26.369595609346003, paired_SE=5.134821124721285, failures=0.",
    "Posterior SD rescaled by min(1, base phat_ticks / grid mean_ticks), gate >0.28 ticks (boundary_0.28_1000_100.json): seeds 1000-1099, n=100, mean_edge=114.436165672195, retail_edge=140.592431021803, arb_edge=-26.156265349567, retail_lost_vs_base=20.866776180780004, arb_saved_vs_base=3.45180256064, paired_delta=-17.414973620251, paired_SE=3.345861543180354, failures=0.",
    "Observed p-bracket width gate >1.5 ticks (width_1.5_100.json): seeds 1000-1099, n=100, mean_edge=114.58159832268501, retail_edge=143.015501926701, arb_edge=-28.433903603973, retail_lost_vs_base=18.443705275882, arb_saved_vs_base=1.1741643062340001, paired_delta=-17.269540969761, paired_SE=3.906755213818802, failures=0.",
    "Observed p-bracket width gate >1 ticks (width_1_100.json): seeds 1000-1099, n=100, mean_edge=114.58159832268501, retail_edge=143.015501926701, arb_edge=-28.433903603973, retail_lost_vs_base=18.443705275882, arb_saved_vs_base=1.1741643062340001, paired_delta=-17.269540969761, paired_SE=3.906755213818802, failures=0.",
    "Posterior SD / competitor spread gate >0.18 (spread_0.18_1000_100.json): seeds 1000-1099, n=100, mean_edge=118.75708196711001, retail_edge=142.663594997969, arb_edge=-23.906513030844, retail_lost_vs_base=18.795612204614002, arb_saved_vs_base=5.701554879363, paired_delta=-13.094057325336001, paired_SE=6.449682889761216, failures=0.",
    "Posterior SD / competitor spread gate >0.24 (spread_0.24_1000_100.json): seeds 1000-1099, n=100, mean_edge=118.89128895424899, retail_edge=143.149261135368, arb_edge=-24.257972181169, retail_lost_vs_base=18.309946067215, arb_saved_vs_base=5.350095729038, paired_delta=-12.959850338197, paired_SE=5.8650188041943245, failures=0.",
    "Posterior SD gate >0.3 ticks (post_0.3_100.json): seeds 1000-1099, n=100, mean_edge=119.771755514825, retail_edge=146.63145296496202, arb_edge=-26.859697450099002, retail_lost_vs_base=14.827754237621, arb_saved_vs_base=2.748370460108, paired_delta=-12.079383777621, paired_SE=3.0642480891061714, failures=0.",
    "Predictive SD including base jump mixture, gate >1.3 ticks (jumpsd_1.3_1000_100.json): seeds 1000-1099, n=100, mean_edge=121.018940278892, retail_edge=147.28461947202, arb_edge=-26.265679193175, retail_lost_vs_base=14.174587730563001, arb_saved_vs_base=3.342388717032, paired_delta=-10.832199013554002, paired_SE=1.856359195147877, failures=0.",
    "Main quote Gaussian next-step non-toxic probability >=0.95; predictive SD capped at 2 ticks (tailpred_0.95_1000_100.json): seeds 1000-1099, n=100, mean_edge=121.17971001835299, retail_edge=142.726940815858, arb_edge=-21.547230797491, retail_lost_vs_base=18.732266386725, arb_saved_vs_base=8.060837112716, paired_delta=-10.671429274093, paired_SE=4.987901265356185, failures=0.",
    "Main quote Gaussian next-step non-toxic probability >=0.9; predictive SD capped at 2 ticks (tailpred_0.9_1000_100.json): seeds 1000-1099, n=100, mean_edge=122.76863532962899, retail_edge=146.068430890136, arb_edge=-23.299795560510002, retail_lost_vs_base=15.390776312447, arb_saved_vs_base=6.308272349697001, paired_delta=-9.082503962817, paired_SE=5.6165601607059985, failures=0.",
    "Main quote posterior non-toxic probability >=0.6; posterior SD capped at 0.4 ticks (credible_0.6_1000_100.json): seeds 1000-1099, n=100, mean_edge=123.49740098318, retail_edge=150.781645035832, arb_edge=-27.284244052629003, retail_lost_vs_base=10.677562166751002, arb_saved_vs_base=2.323823857578, paired_delta=-8.353738309266, paired_SE=1.8885739741529595, failures=0.",
    "Main quote edge must clear 0.5 times posterior SD; same SD capped at 0.4 ticks (marginpost_0.5_1000_100.json): seeds 1000-1099, n=100, mean_edge=123.518089780041, retail_edge=150.781645035832, arb_edge=-27.263555255768, retail_lost_vs_base=10.677562166751002, arb_saved_vs_base=2.3445126544390003, paired_delta=-8.333049512405001, paired_SE=1.8883206391868714, failures=0.",
    "Main quote edge must clear 1 times posterior SD; same SD capped at 0.4 ticks (marginpost_1_1000_100.json): seeds 1000-1099, n=100, mean_edge=123.85552801583701, retail_edge=150.800032453453, arb_edge=-26.94450443759, retail_lost_vs_base=10.659174749130003, arb_saved_vs_base=2.663563472617, paired_delta=-7.995611276609001, paired_SE=1.90225504773929, failures=0.",
    "Diffusion-plus-posterior predictive SD gate >0.9 ticks (pred_0.9_100.json): seeds 1000-1099, n=100, mean_edge=123.974737483634, retail_edge=150.414743248177, arb_edge=-26.440005764501997, retail_lost_vs_base=11.044463954406002, arb_saved_vs_base=3.168062145705, paired_delta=-7.876401808812002, paired_SE=1.75147396070412, failures=0.",
    "Observed p-bracket width gate >2 ticks (width_2_100.json): seeds 1000-1099, n=100, mean_edge=124.645169230685, retail_edge=153.562269457201, arb_edge=-28.917100226549, retail_lost_vs_base=7.896937745382003, arb_saved_vs_base=0.6909676836580005, paired_delta=-7.205970061761003, paired_SE=1.657983992701333, failures=0.",
    "Predictive SD including base jump mixture, gate >2 ticks (jumpsd_2_1000_100.json): seeds 1000-1099, n=100, mean_edge=127.962323165835, retail_edge=157.02885870173202, arb_edge=-29.066535535954, retail_lost_vs_base=4.430348500851002, arb_saved_vs_base=0.5415323742530004, paired_delta=-3.888816126611001, paired_SE=0.7385404626374111, failures=0.",
    "Predictive SD / competitor spread gate >0.35 (predspread_0.35_1000_100.json): seeds 1000-1099, n=100, mean_edge=128.907226739161, retail_edge=153.148303519416, arb_edge=-24.241076780291, retail_lost_vs_base=8.310903683167, arb_saved_vs_base=5.366991129916, paired_delta=-2.943912553285, paired_SE=2.1312659848179014, failures=0.",
    "For H<=200, posterior SD gate >0.4 ticks (latepost_0.4_1000_100.json): seeds 1000-1099, n=100, mean_edge=130.223422819233, retail_edge=160.393254967911, arb_edge=-30.169832148756, retail_lost_vs_base=1.065952234672, arb_saved_vs_base=-0.561764238549, paired_delta=-1.627716473213, paired_SE=0.7553298308431821, failures=0.",
    "Posterior SD gate >0.8 ticks (post_0.8_100.json): seeds 1000-1099, n=100, mean_edge=130.37188078196402, retail_edge=158.599603470887, arb_edge=-28.227722688976, retail_lost_vs_base=2.8596037316959997, arb_saved_vs_base=1.3803452212310001, paired_delta=-1.4792585104819997, paired_SE=0.827554060723282, failures=0.",
    "For H<=200, bracket width gate >2 ticks (latewidth_2_1000_100.json): seeds 1000-1099, n=100, mean_edge=130.493987807786, retail_edge=161.053027054129, arb_edge=-30.559039246415, retail_lost_vs_base=0.40618014845399986, arb_saved_vs_base=-0.9509713362080002, paired_delta=-1.3571514846599997, paired_SE=0.8226180190502456, failures=0.",
    "Diffusion-plus-posterior predictive SD gate >1.3 ticks (pred_1.3_100.json): seeds 1000-1099, n=100, mean_edge=131.082156885671, retail_edge=159.90167262056, arb_edge=-28.819515734950002, retail_lost_vs_base=1.5575345820229998, arb_saved_vs_base=0.7885521752569999, paired_delta=-0.7689824067749993, paired_SE=0.29953415978687226, failures=0.",
    "Predictive SD including base jump mixture, gate >3 ticks (jumpsd_3_1000_100.json): seeds 1000-1099, n=100, mean_edge=131.16550307119, retail_edge=160.61440713648, arb_edge=-29.448904065354, retail_lost_vs_base=0.8448000661030016, arb_saved_vs_base=0.15916384485300006, paired_delta=-0.6856362212560002, paired_SE=0.28519212630685226, failures=0.",
    "Predictive SD cap 1.3 ticks for H<=100 or H>=1000, otherwise 2 (edges_1400_100.json): seeds 1400-1499, n=100, mean_edge=84.807021978873, retail_edge=115.37812095217599, arb_edge=-30.571098973345002, retail_lost_vs_base=0.7278289842090002, arb_saved_vs_base=0.37737099665600016, paired_delta=-0.3504579875419997, paired_SE=0.11831709679271288, failures=0.",
    "Predictive SD cap 1.3 ticks for H<=100, otherwise 2 (maturity_1400_100.json): seeds 1400-1499, n=100, mean_edge=84.91521747537199, retail_edge=115.499994518155, arb_edge=-30.584777042823998, retail_lost_vs_base=0.6059554182299995, arb_saved_vs_base=0.36369292717700014, paired_delta=-0.24226249104300007, paired_SE=0.10457049182714433, failures=0.",
    "Predictive SD cap 1.3 ticks for H>=1000, otherwise 2 (ageadaptive_1400_100.json): seeds 1400-1499, n=100, mean_edge=84.93768325211, retail_edge=115.616937830245, arb_edge=-30.679254578175, retail_lost_vs_base=0.4890121061400005, arb_saved_vs_base=0.26921539182599985, paired_delta=-0.2197967143049998, paired_SE=0.0730059434036753, failures=0.",
    "Belief innovation >max(1 tick, 1 prior predictive SD), pause 1 step(s) (shock_1_1_1000_100.json): seeds 1000-1099, n=100, mean_edge=131.869514738171, retail_edge=158.749732977482, arb_edge=-26.880218239353, retail_lost_vs_base=2.7094742251010024, arb_saved_vs_base=2.7278496708540003, paired_delta=0.018375445724999096, paired_SE=1.3004925704800456, failures=0.",
    "Belief innovation >max(1 tick, 8 prior predictive SD), pause 1 step(s) (shock_8_1_1000_100.json): seeds 1000-1099, n=100, mean_edge=132.015921471077, retail_edge=161.42890551852898, arb_edge=-29.412984047512, retail_lost_vs_base=0.030301684054002208, arb_saved_vs_base=0.1950838626950001, paired_delta=0.1647821786309997, paired_SE=0.14805672599640615, failures=0.",
    "Shock at 3 prior predictive SD (minimum 1 tick), stand aside until bracket width <=2 ticks (reacquire_3_2_1000_100.json): seeds 1000-1099, n=100, mean_edge=132.580496903403, retail_edge=161.081131765242, arb_edge=-28.500634861898003, retail_lost_vs_base=0.3780754373410038, arb_saved_vs_base=1.1074330483090002, paired_delta=0.7293576109570004, paired_SE=0.7886228219586485, failures=0.",
    "Shock pause at 3 prior predictive SD (minimum 1 tick), plus predictive SD gate >2 ticks (combo_3_2_1000_100.json): seeds 1000-1099, n=100, mean_edge=132.595588858786, retail_edge=160.98429695123298, arb_edge=-28.388708092503997, retail_lost_vs_base=0.47491025135000264, arb_saved_vs_base=1.2193598177029998, paired_delta=0.7444495663399995, paired_SE=0.782202439180801, failures=0.",
    "Belief innovation >max(1 tick, 1.5 prior predictive SD), pause 1 step(s) (shock_1.5_1_1000_100.json): seeds 1000-1099, n=100, mean_edge=132.60005457526802, retail_edge=159.543196384308, arb_edge=-26.943141809086, retail_lost_vs_base=1.9160108182750029, arb_saved_vs_base=2.664926101121, paired_delta=0.7489152828220008, paired_SE=1.23495183080697, failures=0.",
    "Belief innovation >max(1 tick, 3 prior predictive SD), pause 3 step(s) (shock_3_3_1000_100.json): seeds 1000-1099, n=100, mean_edge=132.61042252200002, retail_edge=160.511314753908, arb_edge=-27.900892231978002, retail_lost_vs_base=0.9478924486750014, arb_saved_vs_base=1.7071756782290002, paired_delta=0.759283229554001, paired_SE=1.1788038951995448, failures=0.",
    "Belief innovation >max(1 tick, 5 prior predictive SD), pause 3 step(s) (shock_5_3_1000_100.json): seeds 1000-1099, n=100, mean_edge=132.652618030762, retail_edge=161.409867216713, arb_edge=-28.757249186007, retail_lost_vs_base=0.04933998587000168, arb_saved_vs_base=0.8508187241999999, paired_delta=0.8014787383159991, paired_SE=0.8978279568856922, failures=0.",
    "Shock at 2 prior predictive SD (minimum 1 tick), stand aside until bracket width <=1 ticks (reacquire_2_1_1000_100.json): seeds 1000-1099, n=100, mean_edge=132.841333882469, retail_edge=160.236883059089, arb_edge=-27.395549176642998, retail_lost_vs_base=1.2223241434940035, arb_saved_vs_base=2.212518733564, paired_delta=0.9901945900230008, paired_SE=1.1994087207328372, failures=0.",
    "Shock at 2 prior predictive SD (minimum 1 tick), stand aside until bracket width <=2 ticks (reacquire_2_2_1000_100.json): seeds 1000-1099, n=100, mean_edge=132.906205063258, retail_edge=160.29740083037498, arb_edge=-27.391195767153, retail_lost_vs_base=1.1618063722080043, arb_saved_vs_base=2.216872143054, paired_delta=1.055065770812, paired_SE=1.2183210725062092, failures=0.",
    "Belief innovation >max(1 tick, 2.5 prior predictive SD), pause 1 step(s) (shock_2.5_1_1000_100.json): seeds 1000-1099, n=100, mean_edge=133.031639891783, retail_edge=160.88390641882, arb_edge=-27.852266527075997, retail_lost_vs_base=0.5753007837630024, arb_saved_vs_base=1.755801383131, paired_delta=1.180500599337, paired_SE=1.175898533743974, failures=0.",
    "Posterior SD gate >0.4 ticks (post_0.4_1000_600.json): seeds 1000-1599, n=600, mean_edge=96.78794101530033, retail_edge=124.11606359895734, arb_edge=-27.328122583650668, retail_lost_vs_base=11.685592923930498, arb_saved_vs_base=1.2293247155001668, paired_delta=-10.456268208451167, paired_SE=1.7138949529390366, failures=0.",
    "Posterior SD gate >0.55 ticks (post_0.55_1000_600.json): seeds 1000-1599, n=600, mean_edge=103.09441236817717, retail_edge=130.61433153300666, arb_edge=-27.519919164839667, retail_lost_vs_base=5.187324989881167, arb_saved_vs_base=1.0375281343111666, paired_delta=-4.149796855574334, paired_SE=1.0751311492021027, failures=0.",
    "For H<=200, posterior SD gate >0.3 ticks (latepost_0.3_1000_600.json): seeds 1000-1599, n=600, mean_edge=104.70103683657017, retail_edge=134.73701966659252, arb_edge=-30.035982830042165, retail_lost_vs_base=1.0646368562953332, arb_saved_vs_base=-1.478535530891333, paired_delta=-2.543172387181333, paired_SE=1.2699199196070738, failures=0.",
    "Main quote Gaussian next-step non-toxic probability >=0.75; predictive SD capped at 2 ticks (tailpred_0.75_1000_600.json): seeds 1000-1599, n=600, mean_edge=106.44962936712, retail_edge=135.30988063868452, arb_edge=-28.86025127158, retail_lost_vs_base=0.4917758842033328, arb_saved_vs_base=-0.30280397242916673, paired_delta=-0.7945798566315007, paired_SE=1.0949655761614183, failures=0.",
    "Belief innovation >max(1 tick, 2 prior predictive SD), pause 1 step(s) (shock_2_1_1000_600.json): seeds 1000-1599, n=600, mean_edge=106.5936551518015, retail_edge=134.37124881121167, arb_edge=-27.777593659426337, retail_lost_vs_base=1.4304077116761673, arb_saved_vs_base=0.7798536397244998, paired_delta=-0.6505540719499998, paired_SE=0.3137999356833309, failures=0.",
    "Shock pause at 2 prior predictive SD (minimum 1 tick), plus predictive SD gate >2 ticks (combo_2_2_1000_600.json): seeds 1000-1599, n=600, mean_edge=106.60561775904834, retail_edge=134.16209072278733, arb_edge=-27.556472963754835, retail_lost_vs_base=1.6395658001005005, arb_saved_vs_base=1.000974335396, paired_delta=-0.6385914647031664, paired_SE=0.3147007386376728, failures=0.",
    "Uniform-fallback weight >0.25, pause 1 step(s) (fallback_0.25_1_1000_600.json): seeds 1000-1599, n=600, mean_edge=106.79139496096467, retail_edge=134.96502031056468, arb_edge=-28.173625349612667, retail_lost_vs_base=0.8366362123231668, arb_saved_vs_base=0.38382194953816645, paired_delta=-0.45281426278683323, paired_SE=0.3698538630583181, failures=0.",
    "Shock at 3 prior predictive SD (minimum 1 tick), stand aside until bracket width <=1 ticks (reacquire_3_1_1000_600.json): seeds 1000-1599, n=600, mean_edge=107.098948175828, retail_edge=135.27582764194835, arb_edge=-28.176879466138168, retail_lost_vs_base=0.5258288809395002, arb_saved_vs_base=0.38056783301266667, paired_delta=-0.1452610479235006, paired_SE=0.2634832445146035, failures=0.",
    "Belief innovation >max(1 tick, 3 prior predictive SD), pause 1 step(s) (shock_3_1_1000_600.json): seeds 1000-1599, n=600, mean_edge=107.1371012915655, retail_edge=135.3293611840855, arb_edge=-28.192259892539, retail_lost_vs_base=0.4722953388023337, arb_saved_vs_base=0.36518740661183313, paired_delta=-0.10710793218600041, paired_SE=0.26306781527382017, failures=0.",
    "Belief innovation >max(1 tick, 4 prior predictive SD), pause 3 step(s) (shock_4_3_1000_600.json): seeds 1000-1599, n=600, mean_edge=107.18840604323499, retail_edge=135.25096347898517, arb_edge=-28.06255743576617, retail_lost_vs_base=0.550693043902667, arb_saved_vs_base=0.49488986338466656, paired_delta=-0.05580318051650038, paired_SE=0.3881711299867736, failures=0.",
    "Diffusion-plus-posterior predictive SD gate >2 ticks (pred_2_1000_600.json): seeds 1000-1599, n=600, mean_edge=107.22323754864483, retail_edge=135.54692227714315, arb_edge=-28.32368472851217, retail_lost_vs_base=0.25473424574466624, arb_saved_vs_base=0.2337625706386666, paired_delta=-0.02097167510666691, paired_SE=0.036092697536528044, failures=0.",
    "Diffusion-plus-posterior predictive SD gate >2.5 ticks (latepred_2.5_1000_600.json): seeds 1000-1599, n=600, mean_edge=107.24034506638083, retail_edge=135.70720895675368, arb_edge=-28.4668638903885, retail_lost_vs_base=0.09444756613416652, arb_saved_vs_base=0.09058340876233331, paired_delta=-0.0038641573706668926, paired_SE=0.02818123116880513, failures=0.",
    "Diffusion-plus-posterior predictive SD gate >6 ticks (latepred_6_1000_600.json): seeds 1000-1599, n=600, mean_edge=107.24451902381183, retail_edge=135.80140685622118, arb_edge=-28.556887832423836, retail_lost_vs_base=0.0002496666666666651, arb_saved_vs_base=0.000559466726999999, paired_delta=0.00030980006033333544, paired_SE=0.0006130276021549973, failures=0.",
    "Diffusion-plus-posterior predictive SD gate >4 ticks (latepred_4_1000_600.json): seeds 1000-1599, n=600, mean_edge=107.25665991691483, retail_edge=135.79840551621217, arb_edge=-28.541745599311, retail_lost_vs_base=0.0032510066756666588, arb_saved_vs_base=0.015701699839833334, paired_delta=0.012450693163333195, paired_SE=0.006638875704091303, failures=0.",
    "Diffusion-plus-posterior predictive SD gate >3 ticks (latepred_3_1000_600.json): seeds 1000-1599, n=600, mean_edge=107.26224916439634, retail_edge=135.77626680837298, arb_edge=-28.514017643990666, retail_lost_vs_base=0.02538971451483326, arb_saved_vs_base=0.043429655160166655, paired_delta=0.018039940644833198, paired_SE=0.008189239116837856, failures=0.",
    "One-step pause after either 4-SD innovation (minimum 1 tick) or uniform-fallback weight >0.5 (joint_4_0.5_1000_600.json): seeds 1000-1599, n=600, mean_edge=107.26745303300316, retail_edge=135.513489986142, arb_edge=-28.246036953159, retail_lost_vs_base=0.28816653674583353, arb_saved_vs_base=0.3114103459918331, paired_delta=0.02324380925166625, paired_SE=0.3062540895642861, failures=0.",
    "Main quote Gaussian next-step non-toxic probability >=0.6; predictive SD capped at 2 ticks (tailpred_0.6_1000_600.json): seeds 1000-1599, n=600, mean_edge=107.28478758831116, retail_edge=135.523330285312, arb_edge=-28.238542697014335, retail_lost_vs_base=0.27832623757583286, arb_saved_vs_base=0.3189046021364999, paired_delta=0.04057836455966652, paired_SE=0.03899538253602682, failures=0.",
    "Uniform-fallback weight >0.5, pause 1 step(s) (fallback_0.5_1_1000_600.json): seeds 1000-1599, n=600, mean_edge=107.35410113034133, retail_edge=135.575757464751, arb_edge=-28.221656334429166, retail_lost_vs_base=0.22589905813683367, arb_saved_vs_base=0.3357909647216664, paired_delta=0.10989190658983328, paired_SE=0.3012961916290223, failures=0.",
    "Uniform-fallback weight >0.5 pause, plus predictive SD gate >2 ticks (fallback_combo_1000_600.json): seeds 1000-1599, n=600, mean_edge=107.3652229909315, retail_edge=135.355532794641, arb_edge=-27.990309803728337, retail_lost_vs_base=0.44612372824683355, arb_saved_vs_base=0.5671374954224998, paired_delta=0.1210137671799999, paired_SE=0.3019718812259613, failures=0.",
    "Shock pause at 5 prior predictive SD (minimum 1 tick), plus predictive SD gate >2 ticks (combo_5_2_1000_600.json): seeds 1000-1599, n=600, mean_edge=107.37494001138484, retail_edge=135.39856712631635, arb_edge=-28.023627114946503, retail_lost_vs_base=0.40308939657149934, arb_saved_vs_base=0.5338201842043332, paired_delta=0.1307307876333332, paired_SE=0.26659024372395995, failures=0.",
    "Belief innovation >max(1 tick, 5 prior predictive SD), pause 1 step(s) (shock_5_1_1000_600.json): seeds 1000-1599, n=600, mean_edge=107.39521439179534, retail_edge=135.6529763592575, arb_edge=-28.257761967477837, retail_lost_vs_base=0.14868016363033307, arb_saved_vs_base=0.2996853316729999, paired_delta=0.15100516804383293, paired_SE=0.26423542359587715, failures=0.",
    "Shock pause at 4 prior predictive SD (minimum 1 tick), plus predictive SD gate >2 ticks (combo_4_2_1000_600.json): seeds 1000-1599, n=600, mean_edge=107.430909792436, retail_edge=135.29288588560382, arb_edge=-27.861976093183834, retail_lost_vs_base=0.5087706372839995, arb_saved_vs_base=0.6954712059669996, paired_delta=0.18670056868449975, paired_SE=0.2991846554940538, failures=0.",
    "Belief innovation >max(1 tick, 4 prior predictive SD), pause 2 step(s) (shock_4_2_1000_600.json): seeds 1000-1599, n=600, mean_edge=107.43110621021967, retail_edge=135.41364742409218, arb_edge=-27.982541213885998, retail_lost_vs_base=0.3880090987956669, arb_saved_vs_base=0.5749060852648331, paired_delta=0.18689698646816671, paired_SE=0.304738548567465, failures=0.",
    "Belief innovation >max(1 tick, 4 prior predictive SD), pause 1 step(s) (shock_4_1_1000_600.json): seeds 1000-1599, n=600, mean_edge=107.45298269594832, retail_edge=135.54694126664617, arb_edge=-28.0939585707135, retail_lost_vs_base=0.25471525624166647, arb_saved_vs_base=0.4634887284373331, paired_delta=0.20877347219683295, paired_SE=0.2971355312366736, failures=0.",
    "Threshold experiments used python3 -m orderbook_pm_challenge run ../NAME.py --simulations 100 (or 600) --seed-start 1000 --json (adaptive screens used --seed-start 1400). All search and diagnostic simulations used only seeds 1000-1599; holdout was excluded from selection. Width thresholds 1 and 1.5 are equivalent because observed endpoints are integer ticks.",
    "Focused checks: python3 .cache/check_gate.py (2 tests passed), python3 .cache/check_shock.py (2 tests passed); cancellation of existing economic exposure, continued probes, equality to base when confident, and recovery after a shock.",
    "--sandbox --simulations 20 --seed-start 1000: final_sandbox.json, 20 loader failures before strategy execution; Python 3.13 pathlib imports ntpath, which the simulator sandbox rejects. The same failure occurs for examples/starter_strategy.py. nsjail unavailable; no sandbox configuration was changed.",
    "--sandbox --simulations 20 --seed-start 1000: starter_sandbox.json, 20 loader failures before strategy execution; Python 3.13 pathlib imports ntpath, which the simulator sandbox rejects. The same failure occurs for examples/starter_strategy.py. nsjail unavailable; no sandbox configuration was changed.",
    "Frozen holdout control, scored only after selection: unchanged base mean_edge=96.0405203261775, retail_edge=124.3725985363135, arb_edge=-28.332078210128167; candidate retail_edge=123.982770604222, arb_edge=-28.23680257317633, paired_delta=-0.29455229513883385, paired_SE=0.4184176557036628, retail_lost_vs_base=0.38982793209150657, arb_saved_vs_base=0.09527563695183616, failures=0. No holdout value was used to change the frozen policy.",
    "Frozen from shock_4_1 with sha256=3fd37a75874fd4b345a4248ca87d7077e5e5c679cbe6f14f61fd84feeef06229. Final commands from clone: python3 -m orderbook_pm_challenge run ../frozen.py --simulations 600 --seed-start 1000 --workers 4 --json; same with --seed-start 30000; unchanged-base control ../base.py --simulations 600 --seed-start 30000 --workers 4 --json.",
    "Counter-only diagnostic shock_5_1: 1732 gated steps of 1200000, n=600, seed_start=1000, mean_edge=107.39521439179534, failures=0. Counts threshold activations, including steps where economic quantities would already be zero; source strategy actions were not changed by the counters.",
    "Counter-only diagnostic latepred_3: 5161 gated steps of 1200000, n=600, seed_start=1000, mean_edge=107.26224916439634, failures=0. Counts threshold activations, including steps where economic quantities would already be zero; source strategy actions were not changed by the counters.",
    "Counter-only diagnostic shock_4_1: 2406 gated steps of 1200000, n=600, seed_start=1000, mean_edge=107.45298269594832, failures=0. Counts threshold activations, including steps where economic quantities would already be zero; source strategy actions were not changed by the counters.",
    "Tuning stability for shock_4_1: median paired delta=0.0, positive paths=198, negative paths=226, symmetric 5%-trimmed paired mean=0.11001999223518535, six 100-seed block deltas=[1.7220124733749993, 0.40482582616900004, 0.12509042758699868, 0.4327118637470022, -0.5236935941750004, -0.908306163522002]. This is tuning-only analysis; subgroup diagnostics were not used to add conditional trading rules.",
    "Shadow shock_4_1 on unchanged-base paths: mean_edge=107.2442092237515, n=600, seed_start=1000, failures=0; trigger activated on 2393 steps, whose retail edge averaged 0.1309865701576954 per run and arb edge -0.47230052902779807. This diagnostic reproduces the base score exactly. Conditional fill totals are associations on base paths, not causal edge saved/lost by the delivered strategy.",
    "Selected frozen before holdout: full 600 tuning mean_edge=107.45298269594832, retail_edge=135.54694126664617, arb_edge=-28.0939585707135, retail_lost_vs_base=0.254715256241667, arb_saved_vs_base=0.46348872843733346, paired_delta=0.20877347219683295, paired_SE=0.2971355312366736, failures=0. Do not interpret this small, selected-sample difference as a guaranteed improvement."
  ],
  "next": "On fresh seed ranges, separate confidence gates for inventory acquisition and main quotes: suppressing cheap YES purchases can remove later retail edge. Calibrate belief-update reliability jointly with probe coverage, and recheck paired tail contributions when combining this gate with other improvements. Keep 30000-30599 out of future tuning."
}