⚡ Kalshi Fleet Dashboard

Updated 2026-09-18 15:17:01 UTC (11:17 EDT) — server-side regen every 15 min, 60s browser auto-refresh

By Arm

Arm A

+$558.04
Lifetime Bankroll
Today (since 00:00 EDT): +$0.00 (0 cl) • Fees -$108.78
Session (last 15m): +$0.00 (0 cl)
576 closes • 369W/207L • WR 64%

Arm B

+$41.49
Lifetime Bankroll
Today (since 00:00 EDT): +$0.00 (0 cl) • Fees -$55.80
Session (last 15m): +$0.00 (0 cl)
309 closes • 167W/142L • WR 54%

Arm C-HPF

+$949.25
Lifetime Bankroll
Today (since 00:00 EDT): +$0.00 (0 cl) • Fees -$32.43
Session (last 15m): +$0.00 (0 cl)
719 closes • 633W/86L • WR 88%

Arm Lottery

+$34.33
Lifetime Bankroll
Today (since 00:00 EDT): +$0.00 (0 cl) • Fees -$0.00
Session (last 15m): +$0.00 (0 cl)
0 closes • 0W/0L • WR —

By Brain (Bankroll = lifetime since reset)

Filter: 41 brains
Brain ⇅ Arm ⇅ Bankroll ⇅ Today PnL ⇅ Closes ⇅ Wins ⇅ Open ⇅ WR ⇅
v5_hpfC-HPF+$642.08+$0.00221199/22090%
v5_hpf_hybridC-HPF+$293.25+$0.00238213/25089%
v5_mA+$221.34+$0.0011991/28076%
v8_tfB+$209.07+$0.008842/46048%
v6A+$158.29+$0.0074/3057%
v8A+$143.47+$0.0014076/64054%
v5_mrA+$142.43+$0.0011186/25077%
v5_mxyA+$115.49+$0.0031/2033%
v5_pc_tfB+$85.90+$0.0000/00—
v5_2_rLottery+$59.83+$0.0000/00—
v5_mx_tfB+$56.00+$0.008452/32062%
v5_mxA+$55.75+$0.0014393/50065%
v5_mxy_tfB+$34.35+$0.0000/00—
v5_mr_tfB+$27.48+$0.004533/12073%
v5_hpf_pyrC-HPF+$12.89+$0.00214179/35084%
v6_tfB+$9.89+$0.0072/5029%
v5_m_tfB+$5.64+$0.004627/19059%
v5_bmA+$2.93+$0.004014/26035%
v5_bm_tfB+$1.17+$0.003111/20035%
v5_hpf_bC-HPF+$0.81+$0.004642/4091%
v5_rLottery+$0.33+$0.0000/00—
v5_hpf_realpaperC-HPF+$0.22+$0.0000/00—
v5_2A+$0.00+$0.0000/00—
v5_3A+$0.00+$0.0000/00—
v5_4A+$0.00+$0.0000/00—
v5_allA+$0.00+$0.0000/00—
v5_hourlyA+$0.00+$0.0000/00—
v5_hpf_strike_paperC-HPF+$0.00+$0.0000/00—
v5_momA+$0.00+$0.0000/00—
v5_pcpA+$0.00+$0.0000/00—
v5_pcA$-7.17+$0.0010/100%
v7A$-12.74+$0.0083/5038%
v119aLottery$-25.83+$0.0000/00—
v5_mxn_tfB$-31.11+$0.0020/200%
v5_mxnA$-33.48+$0.0041/3025%
v5_mxy_paperA$-40.21+$0.0000/00—
v5_pc_tf_paperB$-40.21+$0.0000/00—
v5_etpA$-68.96+$0.0000/00—
v5_tfB$-78.56+$0.0000/00—
v5A$-119.09+$0.0000/00—
v7_tfB$-238.13+$0.0060/600%

By Coin

Filter: 14 coins
Coin ⇅ Lifetime PnL ⇅ Lifetime W/L ⇅ Lifetime WR ⇅ Today PnL ⇅ Today Fees ⇅ Today W/L ⇅ Today WR ⇅ Session PnL ⇅
SOL———+$198.53-$40.46200/7274%—
sol+$183.52208/10068%—————
ETH———+$166.48-$37.64171/7570%—
eth+$157.63183/9765%—————
HYPE———+$126.34-$4.80109/1588%—
hype+$121.58114/3079%—————
DOGE———+$61.45-$45.78264/10971%—
doge+$57.11275/12768%—————
BNB———+$15.85-$3.46105/1488%—
BTC———+$15.00-$23.81110/5766%—
bnb+$8.69110/3079%—————
btc+$5.32121/8160%—————
XRP———$-85.19-$41.06210/9369%—
xrp$-104.13223/12664%—————

📋 Open Positions (Currently Held)

No open positions in this view.

Arms Details — Strategy, Bankroll, Daily, Last Session

⚠ Bankroll drift detected on 22 brain(s): v5_bm (BK $+2.93 vs daily $-28.53 = drift $+31.47), v5_bm_tf (BK $+1.17 vs daily $-8.67 = drift $+9.83), v5_hpf (BK $+642.08 vs daily $+81.89 = drift $+560.20), v5_hpf_b (BK $+0.81 vs daily $+0.21 = drift $+0.59), v5_hpf_hybrid (BK $+293.25 vs daily $+209.97 = drift $+83.28). Daily PnL is computed from exit_log (authoritative). Drift means kalshi:{brain}:bankroll is not being updated by exit_monitor for these brains.
💰 Fees today (since reset): $197.01 paid in fees | Gross PnL: +$694.56 | Net PnL: +$498.47 | Fee drag: 28.4% of gross | Every PnL number on the dashboard is NET of fees.
Filter: 41 brains Last session = 11:00-11:15 EDT • Win = pnl>0 (net of fees); Tie = pnl=0 (rare, excluded from W/L numerator)
Brain ⇅ Arm ⇅ Strategy ⇅ Lifetime BK ⇅ Daily PnL (net) ⇅ Daily Fees ⇅ Daily W/L/T ⇅ Daily WR ⇅ Session PnL (net) ⇅ Session Fees ⇅ Session W/L/T ⇅ Session WR ⇅
v5_hpfC-HPFHPF (High-Prob Farming)+$642.08+$81.89-$24.52199/2290%+$0.00—0/0—
v5_hpf_hybridC-HPF?+$293.25+$209.97-$0.75213/2589%+$0.00—0/0—
v5_mAMomentum late+$221.34+$88.18-$21.0091/2876%+$0.00—0/0—
v8_tfBSize-adaptive (+trend filter)+$209.07+$116.23-$12.1442/4648%+$0.00—0/0—
v6AMicrostructure+$158.29+$31.85-$4.404/357%+$0.00—0/0—
v8ASize-adaptive+$143.47+$47.82-$21.5476/6454%+$0.00—0/0—
v5_mrAMean reversion early+$142.43+$58.88-$17.8486/2577%+$0.00—0/0—
v5_mxyAYES @ 30-40c+$115.49$-26.48-$2.051/233%+$0.00—0/0—
v5_pc_tfBPolymarket copycat (+trend filter)+$85.90+$0.00—0/0—+$0.00—0/0—
v5_2_rLotteryLottery v1.2+$59.83+$0.00—0/0—+$0.00—0/0—
v5_mx_tfBMomentum full-cycle (+trend filter)+$56.00+$1.74-$9.7252/3262%+$0.00—0/0—
v5_mxAMomentum full-cycle+$55.75+$29.53-$21.5993/5065%+$0.00—0/0—
v5_mxy_tfBYES @ 30-40c (+trend filter)+$34.35+$0.00—0/0—+$0.00—0/0—
v5_mr_tfBMean reversion early (+trend filter)+$27.48+$22.98-$7.7533/1273%+$0.00—0/0—
v5_hpf_pyrC-HPF?+$12.89+$8.91-$5.58179/3584%+$0.00—0/0—
v6_tfBMicrostructure (+trend filter)+$9.89$-28.81-$5.402/529%+$0.00—0/0—
v5_m_tfBMomentum late (+trend filter)+$5.64+$12.00-$9.1527/1959%+$0.00—0/0—
v5_bmA?+$2.93$-28.53-$15.9414/2635%+$0.00—0/0—
v5_bm_tfB? (+trend filter)+$1.17$-8.67-$8.9311/2035%+$0.00—0/0—
v5_hpf_bC-HPFHPF + safeties (stopped)+$0.81+$0.21-$1.5842/491%+$0.00—0/0—
v5_rLotteryLottery v1+$0.33+$0.00—0/0—+$0.00—0/0—
v5_hpf_realpaperC-HPF?+$0.22+$0.00—0/0—+$0.00—0/0—
v5_2ATWAP v2+$0.00+$0.00—0/0—+$0.00—0/0—
v5_3ATWAP v3+$0.00+$0.00—0/0—+$0.00—0/0—
v5_4ATWAP v4+$0.00+$0.00—0/0—+$0.00—0/0—
v5_allAAll-features+$0.00+$0.00—0/0—+$0.00—0/0—
v5_hourlyA?+$0.00+$0.00—0/0—+$0.00—0/0—
v5_hpf_strike_paperC-HPF?+$0.00+$0.00—0/0—+$0.00—0/0—
v5_momA?+$0.00+$0.00—0/0—+$0.00—0/0—
v5_pcpAPoly copycat pro+$0.00+$0.00—0/0—+$0.00—0/0—
v5_pcAPolymarket copycat$-7.17$-5.56-$0.230/10%+$0.00—0/0—
v7ARegression-fit$-12.74$-24.02-$1.573/538%+$0.00—0/0—
v119aLotteryLottery v2 (cheap tail)$-25.83+$0.00—0/0—+$0.00—0/0—
v5_mxn_tfBNO @ 25-70c (+trend filter)$-31.11$-40.86-$1.190/20%+$0.00—0/0—
v5_mxnANO @ 25-70c$-33.48$-44.70-$2.621/325%+$0.00—0/0—
v5_mxy_paperA?$-40.21+$0.00—0/0—+$0.00—0/0—
v5_pc_tf_paperB?$-40.21+$0.00—0/0—+$0.00—0/0—
v5_etpA?$-68.96+$0.00—0/0—+$0.00—0/0—
v5_tfBTWAP base (+trend filter)$-78.56+$0.00—0/0—+$0.00—0/0—
v5ATWAP base$-119.09+$0.00—0/0—+$0.00—0/0—
v7_tfBRegression-fit (+trend filter)$-238.13$-4.10-$1.520/60%+$0.00—0/0—

Last 30 Closes

Filter: 30 rows
Time ⇅ Brain ⇅ Arm ⇅ Strategy ⇅ Ticker ⇅ Coin ⇅ Side ⇅ Conf ⇅ V128 Tier ⇅ $Size ⇅ Fee ⇅ Entry ⇅ Exit ⇅ PnL ⇅ Reason ⇅
12:20:13v5_hpfC-HPFHPF (High-Prob Farming)KXHYPE15M-26MAY201200-00HYPEYES0.940HPF$24.75-$0.020.9901.000+$0.23resolved_win
12:20:02v5_hpfC-HPFHPF (High-Prob Farming)KXBNB15M-26MAY201200-00BNBYES0.970HPF$29.22-$0.060.9741.000+$0.72resolved_win
12:19:15v5_hpfC-HPFHPF (High-Prob Farming)KXSOL15M-26MAY201200-00SOLYES0.950HPF$3.70-$0.070.7401.000+$1.23resolved_win
12:17:56v5_hpf_hybridC-HPF?KXXRP15M-26MAY201200-00XRPYES0.900HPF$19.14—0.8701.000+$2.86resolved_win
12:17:45v5_hpf_hybridC-HPF?KXETH15M-26MAY201200-00ETHYES0.920HPF$19.92—0.8301.000+$4.08resolved_win
12:17:01v5_hpf_hybridC-HPF?KXBTC15M-26MAY201200-00BTCYES0.900HPF$19.75—0.7901.000+$5.25resolved_win
12:10:20v6_tfBMicrostructure (+trend filter)KXDOGE15M-26MAY201200-00DOGENO0.8052.0x$19.74-$0.740.4701.000+$21.52resolved_win
12:10:08v5_hpf_hybridC-HPF?KXHYPE15M-26MAY201200-00HYPEYES0.940HPF$24.70—0.9501.000+$1.30resolved_win
12:10:02v5_hpf_hybridC-HPF?KXBNB15M-26MAY201200-00BNBYES0.970HPF$29.22—0.9741.000+$0.78resolved_win
12:09:14v5_hpf_hybridC-HPF?KXSOL15M-26MAY201200-00SOLYES0.950HPF$29.24—0.8601.000+$4.76resolved_win
12:07:56v5_hpfC-HPFHPF (High-Prob Farming)KXXRP15M-26MAY201200-00XRPYES0.900HPF$19.14-$0.180.8701.000+$2.68resolved_win
12:07:43v5_hpfC-HPFHPF (High-Prob Farming)KXETH15M-26MAY201200-00ETHYES0.920HPF$19.92-$0.240.8301.000+$3.84resolved_win
12:06:59v5_hpfC-HPFHPF (High-Prob Farming)KXBTC15M-26MAY201200-00BTCYES0.900HPF$19.75-$0.300.7901.000+$4.95resolved_win
12:05:14v5_hpfC-HPFHPF (High-Prob Farming)KXHYPE15M-26MAY201145-45HYPENO0.850HPF$9.24-$0.110.8401.000+$1.65resolved_win
12:00:18v6AMicrostructureKXDOGE15M-26MAY201200-00DOGENO0.8052.0x$19.74-$0.740.4701.000+$21.52resolved_win
12:00:06v5_hpf_pyrC-HPF?KXHYPE15M-26MAY201200-00HYPEYES0.850HPF$1.98-$0.010.9901.000+$0.01resolved_win
12:00:00v5_hpf_pyrC-HPF?KXBNB15M-26MAY201200-00BNBYES0.850HPF$1.95-$0.020.9740.997+$0.03time_stop
11:59:11v5_hpf_pyrC-HPF?KXSOL15M-26MAY201200-00SOLYES0.850HPF$1.48-$0.060.7400.977+$0.41take_profit
11:57:51v5_hpf_pyrC-HPF?KXXRP15M-26MAY201200-00XRPYES0.850HPF$1.72-$0.040.8600.565$-0.63hard_stop_dd
11:57:38v5_hpf_pyrC-HPF?KXETH15M-26MAY201200-00ETHYES0.850HPF$1.62-$0.060.8100.610$-0.46hard_stop_dd
11:56:55v5_hpf_pyrC-HPF?KXBTC15M-26MAY201200-00BTCYES0.850HPF$1.48-$0.060.7400.970+$0.40take_profit
11:55:34v5_bmA?KXSOL15M-26MAY201200-00SOLNO0.9102.5x$4.93-$0.580.1700.130$-1.74hard_stop_dd
11:55:12v5_hpf_pyrC-HPF?KXBNB15M-26MAY201145-45BNBNO0.850HPF$1.78-$0.020.8901.000+$0.20resolved_win
11:55:11v5_hpf_hybridC-HPF?KXHYPE15M-26MAY201145-45HYPENO0.850HPF$9.24—0.8401.000+$1.76resolved_win
11:50:22v5_hpf_pyrC-HPF?KXSOL15M-26MAY201130-30SOLYES0.850HPF$1.98-$0.010.9891.000+$0.01resolved_win
11:50:16v5_hpfC-HPFHPF (High-Prob Farming)KXXRP15M-26MAY201130-30XRPYES0.980HPF$29.61-$0.030.9871.000+$0.36resolved_win
11:50:15v5_hpfC-HPFHPF (High-Prob Farming)KXETH15M-26MAY201130-30ETHYES0.990HPF$29.64-$0.030.9881.000+$0.33resolved_win
11:50:15v5_hpfC-HPFHPF (High-Prob Farming)KXBNB15M-26MAY201130-30BNBYES0.970HPF$29.28-$0.050.9761.000+$0.67resolved_win
11:50:03v5_hpfC-HPFHPF (High-Prob Farming)KXBTC15M-26MAY201130-30BTCYES0.980HPF$29.67-$0.030.9891.000+$0.30resolved_win
11:45:07v5_hpf_pyrC-HPF?KXHYPE15M-26MAY201145-45HYPENO0.850HPF$1.68-$0.020.8401.000+$0.30resolved_win
⏸ LIVE HPF BRAIN STOPPED — no new live entries. Reason: manual. Open positions will resolve naturally.

By Arm (LIVE)

Filter: live brains
Brain ⇅ Arm ⇅ Truth BK ⇅ Today PnL ⇅ W/L ⇅ WR ⇅ Open ⇅ Mode Size $ Current
v5_hpf_liveTAKER$-0.64+$0.00 (0cl)3/175%0mode=default→live, size=$1.00
v5_hpf_hybrid_liveMAKER$-3.71+$0.00 (0cl)18/195%0mode=default→live, size=$1.00
bypassBYPASS—192752 hits——0ENABLED$2end-of-cycle squeeze, layered on HPF brains

By Brain (LIVE — lifetime since reset)

Filter: 7 brains
Brain ⇅ Arm ⇅ Bankroll ⇅ Today PnL ⇅ Closes ⇅ Wins ⇅ Open ⇅ WR ⇅
v5_etp_liveA+$43.98+$43.987656/20074%
v5_hpf_pyr_liveC-HPF+$0.00+$0.001514/1093%
v5_hpf_liveC-HPF$-0.64$-0.6443/1075%
v5_hpf_hybrid_liveC-HPF$-3.71$-3.711918/1095%
v5_hpf_strike_liveC-HPF$-364.51$-364.5100/00—
v5_mxy_liveA$-364.51$-364.5100/00—
v5_pc_tf_liveB$-364.51$-364.5100/00—

By Coin (LIVE — today, all coins)

Filter: 7 coins
Coin ⇅ Lifetime PnL ⇅ Lifetime W/L ⇅ Lifetime WR ⇅ Today PnL ⇅ Today Fees ⇅ Today W/L ⇅ Today WR ⇅ Session PnL ⇅
xrp+$14.7620/291%—————
doge+$10.4720/387%—————
sol+$7.0316/384%—————
hype+$6.539/282%—————
bnb+$1.085/271%—————
btc+$0.9210/759%—————
eth$-1.8411/473%—————

📋 Open Positions (Currently Held — LIVE)

No open positions in this view.

Arms Details — Strategy, Bankroll, Daily, Last Session

✓ Integrity check passed: all brain bankrolls match sum(realized_pnl) since fresh-start (within $0.01)
💰 Fees today (since reset): $0.65 paid in fees | Gross PnL: $-3.96 | Net PnL: $-4.61 | Fee drag: -16.4% of gross | Every PnL number on the dashboard is NET of fees.
Filter: 7 brains Last session = 11:00-11:15 EDT • Win = pnl>0 (net of fees); Tie = pnl=0 (rare, excluded from W/L numerator)
Brain ⇅ Arm ⇅ Strategy ⇅ Lifetime BK ⇅ Daily PnL (net) ⇅ Daily Fees ⇅ Daily W/L/T ⇅ Daily WR ⇅ Session PnL (net) ⇅ Session Fees ⇅ Session W/L/T ⇅ Session WR ⇅
v5_etp_liveA?+$43.98+$0.00—0/0—+$0.00—0/0—
v5_hpf_pyr_liveC-HPF?+$0.00$-0.26-$0.1814/193%+$0.00—0/0—
v5_hpf_liveC-HPF?$-0.64$-0.64-$0.023/175%+$0.00—0/0—
v5_hpf_hybrid_liveC-HPF?$-3.71$-3.71-$0.4418/195%+$0.00—0/0—
v5_hpf_strike_liveC-HPF?$-364.51+$0.00—0/0—+$0.00—0/0—
v5_mxy_liveA?$-364.51+$0.00—0/0—+$0.00—0/0—
v5_pc_tf_liveB?$-364.51+$0.00—0/0—+$0.00—0/0—

Last 30 Closes (LIVE)

Filter: 30 rows
Time ⇅ Brain ⇅ Arm ⇅ Strategy ⇅ Ticker ⇅ Coin ⇅ Side ⇅ Conf ⇅ V128 Tier ⇅ $Size ⇅ Fee ⇅ Entry ⇅ Exit ⇅ PnL ⇅ Reason ⇅
09:45:15v5_hpf_liveC-HPF?KXXRP15M-26JUN130945-45XRPNO0.880HPF$0.91-$0.010.9101.000+$0.08resolved_win
02:45:12v5_hpf_liveC-HPF?KXXRP15M-26JUN130245-45XRPYES0.910HPF$0.92-$0.010.9171.000+$0.08resolved_win
20:30:08v5_hpf_liveC-HPF?KXSOL15M-26JUN122030-30SOLYES0.880HPF$0.90-$0.010.9000.000$-0.91resolved_loss
15:00:03v5_hpf_liveC-HPF?KXSOL15M-26JUN121500-00SOLNO0.880HPF$0.89-$0.010.8901.000+$0.10resolved_win
06:20:41v5_hpf_pyr_liveC-HPF?KXETH15M-26MAY190615-15ETHNO0.850HPF$1.18-$0.040.5900.000$-1.22resolved_loss
02:49:40v5_hpf_pyr_liveC-HPF?KXXRP15M-26MAY190245-45XRPYES0.850HPF$1.98-$0.010.9891.000+$0.01resolved_win
02:45:12v5_hpf_pyr_liveC-HPF?KXBNB15M-26MAY190245-45BNBYES0.850HPF$1.93-$0.010.9671.000+$0.06resolved_win
02:45:05v5_hpf_hybrid_liveC-HPF?KXBTC15M-26MAY190245-45BTCYES0.910HPF$5.72-$0.030.9531.000+$0.25resolved_win
02:35:35v5_hpf_pyr_liveC-HPF?KXSOL15M-26MAY190230-30SOLYES0.850HPF$1.94-$0.010.9681.000+$0.05resolved_win
02:30:05v5_hpf_hybrid_liveC-HPF?KXBTC15M-26MAY190230-30BTCYES0.900HPF$3.61-$0.030.9021.000+$0.36resolved_win
02:15:12v5_hpf_hybrid_liveC-HPF?KXDOGE15M-26MAY190215-15DOGEYES0.970HPF$3.90-$0.020.9751.000+$0.08resolved_win
02:15:11v5_hpf_pyr_liveC-HPF?KXSOL15M-26MAY190215-15SOLYES0.850HPF$1.87-$0.010.9331.000+$0.12resolved_win
02:03:34v5_hpf_pyr_liveC-HPF?KXETH15M-26MAY190200-00ETHYES0.850HPF$1.98-$0.010.9901.000+$0.01resolved_win
02:00:13v5_hpf_pyr_liveC-HPF?KXBNB15M-26MAY190200-00BNBYES0.850HPF$1.97-$0.010.9841.000+$0.02resolved_win
01:58:27v5_hpf_pyr_liveC-HPF?KXDOGE15M-26MAY190145-45DOGEYES0.850HPF$0.91-$0.010.9101.000+$0.08resolved_win
01:15:06v5_hpf_hybrid_liveC-HPF?KXBTC15M-26MAY190115-15BTCNO0.930HPF$5.49-$0.050.9161.000+$0.46resolved_win
01:15:05v5_hpf_pyr_liveC-HPF?KXETH15M-26MAY190115-15ETHNO0.850HPF$1.92-$0.010.9601.000+$0.07resolved_win
01:00:12v5_hpf_hybrid_liveC-HPF?KXSOL15M-26MAY190100-00SOLYES0.940HPF$3.84-$0.010.9591.000+$0.15resolved_win
00:58:11v5_hpf_hybrid_liveC-HPF?KXXRP15M-26MAY190015-15XRPYES0.860HPF$5.52—0.9201.000+$0.48resolved_win
00:45:15v5_hpf_hybrid_liveC-HPF?KXHYPE15M-26MAY190045-45HYPENO0.910HPF$3.89-$0.020.9731.000+$0.09resolved_win
00:34:11v5_hpf_hybrid_liveC-HPF?KXXRP15M-26MAY190030-30XRPNO0.920HPF$5.85-$0.030.9751.000+$0.12resolved_win
00:00:19v5_hpf_hybrid_liveC-HPF?KXBTC15M-26MAY190000-00BTCNO0.980HPF$5.92-$0.020.9871.000+$0.06resolved_win
00:00:06v5_hpf_hybrid_liveC-HPF?KXETH15M-26MAY190000-00ETHNO0.920HPF$1.84-$0.010.9181.000+$0.15resolved_win
00:00:04v5_hpf_pyr_liveC-HPF?KXSOL15M-26MAY190000-00SOLNO0.850HPF$1.94-$0.010.9701.000+$0.05resolved_win
23:30:14v5_hpf_pyr_liveC-HPF?KXHYPE15M-26MAY182330-30HYPENO0.850HPF$1.94-$0.010.9701.000+$0.05resolved_win
23:30:08v5_hpf_pyr_liveC-HPF?KXDOGE15M-26MAY182330-30DOGEYES0.850HPF$0.87-$0.010.8701.000+$0.12resolved_win
23:15:11v5_hpf_hybrid_liveC-HPF?KXDOGE15M-26MAY182315-15DOGEYES0.890HPF$4.51-$0.040.9021.000+$0.45resolved_win
23:15:09v5_hpf_hybrid_liveC-HPF?KXXRP15M-26MAY182315-15XRPYES0.940HPF$5.84-$0.040.9731.000+$0.12resolved_win
22:48:35v5_hpf_hybrid_liveC-HPF?KXSOL15M-26MAY182245-45SOLNO0.950HPF$5.66-$0.010.9431.000+$0.33resolved_win
22:48:29v5_hpf_hybrid_liveC-HPF?KXXRP15M-26MAY182245-45XRPNO0.930HPF$7.53-$0.030.9411.000+$0.44resolved_win
ARCHITECTURE — Last updated 2026-06-12 (🧪 P1 VALIDATION EXPERIMENT LIVE — post-audit restart: v5_hpf_live ONLY, $1 taker, fills hard-capped to the 0.88–0.92 zone (V312 executor floor + 0.92 final-price cap, NO path exempt), hold-to-settle with no stops (V313: any *_live source = resolution-only), 21:45–23:00 EDT entry curfew (V316), max 2 live positions per 15-min cycle (V315 correlated-basket cap), fail-CLOSED fill confirmation + settle phantom guard (V310/V313), per-source truth reconciler (V314 ordermap), p1_tracker enforces the pre-registered −$15 abort → kalshi:halt=1. V319 kalshi_p1_analyzer: REAL-TIME per-trade Telegram (coin/side/entry/PnL + running coin WR, EDT-hour WR, (coin,side) loss streak) + throttled flags (coin WR<70%% n>=5, 3 consec losses, EDT hour bleeding) + queryable kalshi:p1:analyzer:* snapshots — ANALYSIS-ONLY, writes NO trading key, pre-registration intact. SUCCESS = WR>90.6%% with z≥1.65 on ≥280 settles (~18 days). V318 shadow quoter measures maker adverse selection (NO orders). Protocol: kalshi_v1/P1_PREREGISTRATION.md. Basis: 18-agent forensic audit — market calibrated (WR 91.1%% vs price 91.4%%), books overstated +$306 in 4 days, 0.88–0.92 the only candidate-positive zone. | Prior: LOCK-IN 2026-05-20: live = $1 HPF taker + bypass, BOTH $1-capped via V211 kalshi:settings:size; ALL paper brains PAUSED + autostart=false; hpf_live autostart=true so REBOOT-PROOF; realized-only -$10 kill + $225 floor; box overload = Polymarket rebuild_v6 swarm NOT Kalshi. Prior V306 — REAL-CASH truth tracker + ETP live KILLED + position reconciler + bypass re-enabled $1), V230 "whale in kalshi" paper brain (NEW: reads poly:whale_mirror:* read-only, fetches Polymarket trade details, mirrors $250+ crypto-market whale buys to Kalshi 15M paper — first offensive edge addition, runs as kalshi_whale_paper_brain isolated from all existing brains, $2 paper sizing), V228b pyramid 1c slip (V212A_SLIP_CENTS=0.01 — LEG 2/3 fire when current_bid >= LEG_1_entry - 1c, lifts pyramid completion 2-7% → 25-40% est while still blocking HYPE-pattern 3c drops), V228 SECS_MAX 600→420 (all V170/V127/V208/V5_BM windows tightened to 7 min after orderbook-reality investigation confirmed Kalshi 15M cycles have EMPTY real orderbooks for first 10-12 min — only seeded phantom data; real orders appear in final ~5 min), V227 KALSHI_BALANCE_FLOOR=265 (auto-halt all live entries if Kalshi cash drops below $265 via existing V112.Z mechanism — current cash $276.68, $11.68 cushion), V226 timing rollback (V220+V220b+V221+V224 reverted — paper attribution analysis revealed those "wins" were against phantom cached prices; V228 then re-tightened based on orderbook reality), V225 env watchdog (kalshi_env_watchdog.py polls /proc/PID/environ every 60s for 11 critical brains, Telegram alerts on mismatch — designed to catch the V220 zombie-env failure pattern within 60s), V223 V209C_PYR_DD_PCT 0.08→0.15 (pyramid LEG 2/3 hard_stop loosened — stops killing winners that briefly drawdown after HYPE 08:26 disaster), V222 HYPE strict pause (kalshi:coin_pause_strict:hype overrides V187 exemption — blocks HYPE on live HPF only, BNB still allowed; HYPE was -$16.54/24h biggest single bleed). | Prior (V197-V204): V204 dashboard by-coin lifetime columns (LIVE+PAPER coin tables auto-iterate d[closes] each regen — adds Lifetime PnL, W/L, WR columns next to today columns; coin rows sorted by lifetime PnL desc so best EV surfaces first; computes per-coin from realized_pnl_true with realized_pnl fallback), V203 LIVE cards auto-iterate live_sources (no more hardcoded source list — future brain added to KALSHI_LIVE_SOURCES env auto-appears as card with no gen.py edit; LIVE_CARD_META dict for custom branding fallback) + PAPER fleet banner adds "Fleet PnL today" item, V202 V126 floor 0.73→0.70 (loosens A/B emit gate that was filtering 663K signals/day to ~zero trades; arch tab brain lists now show "5 RUNNING / 9 conf" with explicit RUNNING + STOPPED breakdown), V201 PAPER cards Today PnL fix (V184 had silently set Daily=Lifetime — now computes true today PnL by filtering d[closes] to today_ts), V200 LIVE cards add Today PnL line + 3rd PYR card (pyr_live promoted to first-class live brain via V193+V197), V199 dashboard cleanup + bk drift prevention (removed "Backend LIVE Program Status" dead block; truth_reconciler now auto-reconciles kalshi:{src}:bankroll = truth every 60s — drift cannot accumulate beyond 1 min; V199c one-time correction zeroed $110.09 of accumulated phantom drift across 3 live brains), V198a TRACK_SOURCES adds pyr_live (truth reconciler now backfills pyr_live exit_log entries — was missing realized_pnl_true on 100% of pyr_live closes), V198b maker_order taker cross verify-fill (place_maker_then_cross taker fallback was marking filled=True on place_order ack alone — caused 3 phantom hybrid "wins" today with truth=$0; now polls /portfolio/orders + /portfolio/fills 3× before marking filled), V197 PYR_LIVE full integration (crypto_15min_v5_hpf_pyr_live was missing from _V5_PER_SOURCE_POS → V189b position-check ALWAYS failed → pyramid LEG 2/3 NEVER fired since V179 deploy; also missing from _per_source_dedup and V167 cross-source peer set — caused BNB 03:30 double-fire $1.38 phantom loss. V197 fixes: 3-way V167 with same-source bypass, TAKER LIVE path now tags per-source + V167 cross-side, per-source dedup bypass for pyr LEG 2/3). Previously: V195 combined HPF live boost (HPF taker + hybrid live brains auto-bump dollar2→dollar3 at COMBINED 75 wins AND 85% WR — backfilled 54W/6L 90% WR, 21 wins from boost), V194 boost gate WR floor (V189d now requires BOTH 75 wins AND 85% WR before bumping pyr live dollar1→dollar3 — adds V194a losses_set tracking + V194b WR check), V193 REAL MONEY PYRAMID enabled (crypto_15min_v5_hpf_pyr_live ADDED to KALSHI_LIVE_SOURCES env — all V179/V183/V186/V189/V190 had been paper-only due to missing whitelist entry. Now real Kalshi orders flow. V193b adds bk mirrors so kalshi:v5_hpf_pyr_live:bankroll actually updates on close. V193c propagates v179_pyramid_leg from thesis through position hash to exit_log for leg-level audit. V191 watchdog now sees real PnL deltas), V190 pyr LEG2/3 checkpoint relaxation (CKPT_2 300→350, CKPT_3 120→180 — fires earlier before books widen, fewer V44 slippage rejects), V189 pyramid robustness + adaptive sizing (V189a Redis tracker persistence survives brain restarts, V189b position-check before LEG 2/3 emit (no more phantom adds), V189c v179_pyramid_leg flows to exit_log for leg-level audit, V189d LIVE auto-bumps dollar1→dollar3 after 75 lifetime wins via SCARD of kalshi:v179:wins_set), V188 V5.BM drift-aware Brownian brain (NEW: Black-Scholes binary fair_p with realized sigma + drift estimation. Standalone brain on Arm A + Arm B paper. v1 (driftless) showed classic no-drift failure - wanted NO on all trending markets. v3 (drift): flipped ETH/SOL/XRP to weak YES agreement with market. Only fires when gap >= 15c. Universe: 5 spot-validated coins.), V187 HPF coin_pause bypass (HPF live brains exempted from BNB+HYPE V168 pause - HPF doesnt need spot WS validation that V5/V117 brains need, so it can trade these coins. First BNB HPF live fill at 20:50 EDT 2026-05-16: YES @ 0.939, $1.88), V186 V170 LEG 1 only (pyramid LEG 2/3 bypass V170 MID_BID_FLOOR — they add to already-winning positions, different statistical population than V170 historical bleed cell. Also fixed V183 LEG 2 omission in live brain.), V184 dashboard arm daily uses bankroll (reconciles with Lifetime BK after phantom cleanups), V183 V179 signal-based MTM (pyramid pyramids on any 0.85+ bid, not strict bid>thesis - catches DOGE-style high-entry winners), V179 LIVE PYRAMID (v5_hpf_pyr_live: $1/leg LEG1+LEG2+LEG3 at secs=300/120 checkpoints, max $3 exposure, asymmetric design preserves baseline), V175 v3 SURGICAL crossed-market guard (rejects only when entryV182 V170 floor 0.88->0.86 (DOGE-style 0.86-0.87 bids no longer blocked - was over-aggressive), V181 PAPER fleet math reconciles (fleet=sum of arms=sum of bankrolls), V180 LIVE dashboard dynamic (per-source counts not account-aggregated), V178 hybrid pause/resume, V177 phantom reconciler rewrite (insufficient data: 6 trades, truth -$7.57; resume via supervisorctl), V174 V128 time-gate (last 5min only) (V128 boost now only fires when secs_left<300 — additive to V152 WR gate. Matches user instinct: more time to expiry = bid less reliable, so dont amplify), V173 consolidated BNB/HYPE blocking to V168 only (removed redundant V142 executor check + V165 brain check + 4 KALSHI_SKIP_COINS env vars — V168 coin_pause is sole source of truth), V172 HPF win-ladder removed (live brains now FLAT $2 — simpler, predictable risk profile), V170 HPF entry tightening (secs cap 600 + bid floor 0.88 in 2-4m window — surgical cuts to historical bleed cells, expected +$30/day at current volume), V168 universal coin-pause (BNB+HYPE indefinitely paused — no spot WS validation), V167 TAKER/MAKER cross-source dedup (live HPF brains no longer both bet on same ticker+side - prevents wasted capital), V166 maker meta in exit_log (maker_mode + maker_saved + maker_v1 now flow from position to exit_log for full maker auditability), V165 brain-level SKIP_COINS (all 4 HPF brains - paper + live - now skip BNB/HYPE at scan time, was executor-only), V164 phantom-position reconciler (every 60s pulls /portfolio/positions from Kalshi, deletes Redis records >90s old with no matching Kalshi contract - fixes dashboard showing 7-8 open while Kalshi has 0), V156 hybrid maker/taker LIVE (v5_hpf_hybrid_live runs parallel to v5_hpf_live — same Kalshi account, V25-C posts limit @ bid+1c with taker fallback), V160 dashboard adds C-LIVE TAKER + C-LIVE MAKER cards (replaces Lottery), V157 multi-live source support, CF DROPPED entirely (was misleading 30-130s stale, replaced by Coinbase+Kraken WS @ sub-second), V128 ladder gated by 80% WR (no more $19 losses from sub-80% brains), V142 executor records actual Kalshi fill prices (kills phantom PnL at source), V138 truth reconciler pulls Kalshi /portfolio/settlements every 60s (authoritative bookkeeping), HPF LIVE back to FLAT $1 with win-ladder, BNB+HYPE skipped (no WS support).
AUDIT 2026-05-16 12:30 EDT (full-fleet sweep)
• Paper today: Arm A 335W/194L (63%) +$583 | Arm B 92W/54L (63%) +$160 | C-HPF paper 75W/13L (85%) +$51 | Lottery 10W/44L (19%) -$26
• LIVE today: Kalshi balance $289.15 (+$1.24 session delta, 0.9h) | exit_log reports 32W/7L +$93.76 | truth reconciler says 6W/2L -$2.70
• Phantom PnL still present: exit_log XRP entry recorded at $0.042 (107 sh) but Kalshi actually filled at ~$0.91 - V142 fill-poll patch in place but not always overwriting thesis_entry. Truth = -$2.70, not +$94.
• Per coin (paper): ETH +$313, SOL +$234, DOGE +$192, BTC +$74, BNB +$18, HYPE +$3, XRP -$65 (only loser)
• Top brains (paper): v7 +$106, v8 +$104, v5_pc +$101, v5_mx +$93, v5_m +$90, v8_tf +$82, v5_mr +$74 | Bottom: v5 -$82, v5_tf -$45, v119a -$26
• Infrastructure: 32 brain PIDs · truth reconciler 17s ago · multi_ws @ sub-sec (CB+Kraken) · 5 traded coins fresh · V130 0 pauses · 25 open positions
• Hybrid LIVE (V25-C maker): 2 closes today (1W resolved_win + 1 even resolved_win) - maker still building sample. Recent emits showed mode=maker fill=0.86 saved=$0.02

Fleet at a Glance

Arm A (Parents)

Brains0 RUNNING (paper PAUSED 2026-05-20) / 9 conf
Base size$10/trade (V128 boosted)
ConfidenceBrain-computed (0.65 - 0.85)
FiltersV117 + V126 (conf>=0.70, V202) + V127 (secs<=600s) + V128 sizing
ExitsV5-family resolution-only + V5 trail @ HWM>=0.80, give-back 45%
Brain listRUNNING: v5_mr, v5_mxn, v5_mxy, v6, v5_bm (V188 Brownian). STOPPED: v5, v5_m, v5_mx, v5_pc, v7, v8 (killed 2026-05-16/17 for drift)

Arm B (_tf Twins)

Brains0 RUNNING (paper PAUSED 2026-05-20) / 9 conf
Base size$10/trade (V128 boosted)
ConfidenceBrain-computed (same as A)
FiltersSame as A + V114 trend filter (ONLY diff). V116 entry gate now OFF.
ExitsV124 resolution-only (no early TP/trail/dd-stop)
Brain listRUNNING: v5_mxn_tf, v5_mxy_tf, v5_pc_tf, v6_tf, v5_bm_tf (V188 Brownian twin). STOPPED: v5_tf, v5_m_tf, v5_mr_tf, v5_mx_tf, v7_tf, v8_tf (killed 2026-05-16/17 for drift)

Arm C - HPF (paper + LIVE)

Brains2
Base sizePaper: $10 x bid tier. LIVE: $1/trade (V211 cap 2026-05-20; only v5_hpf_live runs, hybrid+pyr STOPPED, paper PAUSED) (was V172 — ladder removed 2026-05-16)
ConfidenceMarket bid (0.85 - 1.00)
FiltersNo V117/V126/V127/V128. Trigger = bid>=0.85 + secs in [60,600] (V170 tightened from 840). V170: bid>=0.88 required in 2-4m window (120-240s). BNB+HYPE+SOL paused via V168.
ExitsHold to resolution + 40c stop-loss. V142 executor polls Kalshi /portfolio/orders for actual fill price BEFORE writing position (kills phantom PnL at source).
Brain listv5_hpf (paper), v5_hpf_hybrid (paper, V134 maker), v5_hpf_live (LIVE TAKER, FLAT $2 via V172), v5_hpf_hybrid_live (LIVE MAKER V156, resumed via V178), v5_hpf_pyr_live (V179 PYRAMID — V193 REAL MONEY enabled 2026-05-17 02:11 EDT, $1/leg max $3, V183 signal-based MTM, V189d auto-bumps $1→$3 after 75 wins). Plus paper variants: v5_hpf_paper, v5_hpf_b_paper, v5_hpf_hybrid_paper, v5_hpf_pyr_paper. Truth reconciler V138 pulls Kalshi API every 60s.

Lottery

Brains1
Base size$1 flat
Confidencen/a (V126-exempt)
Filtersentry<0.30 only (cheap-tail capture)
ExitsHold to resolution
Brain listv119a

Filter Pipeline (every emit flows top-down)

Brain candidate emit
Each brain checks markets per strategy. Reads kalshi:rti:current:{coin} (sub-second WS aggregate)
▼
Price source: Coinbase+Kraken WS aggregator
V140: kalshi_multi_ws subscribes to CB Advanced Trade + Kraken WS. Median of available exchanges. 0.1-0.4s freshness. Binance geo-blocked (HTTP 451).
▼
V168 universal coin-pause (V142+V165 retired via V173)
BNB+HYPE indefinite via V168 Redis key. SOL paused 3hr. Single source of truth. V142 SKIP_COINS env and V165 brain check were REMOVED in V173 (redundant).
▼
V126 conf floor >= 0.70
Block emits below 0.70 confidence (V202 lowered from 0.73 to unblock A/B which had near-zero emit volume)
▼
V127 secs cap <= 600s
Block if >10 min to close
▼
V170 HPF SECS_MAX=600 (HPF-only)
Drops HPF emits with secs_left > 600 (was 840). Eliminates the 10-14m EARLIEST entry bucket — historically the only losing HPF window (-$1.52 over 23 trades).
▼
V170 HPF MID_BID_FLOOR (HPF-only, 2-4m window)
When secs_left in [120,240), requires bid>=0.86 (V182 relaxed from 0.88). Cuts the worst single HPF cell. Counterfactual: 100 trades (63% retained), +$30.38 net improvement, WR 85%->92%.
▼
V175 v3 — Crossed-market guard (HPF-only)
Reject HPF emit if trigger_entry < trigger_bid - 0.30 (ask suspiciously low relative to bid). Catches XRP-shape phantom emits (bid=89c ask=3c) without blocking normal wide spreads. Replaced V164 phantom reconciler protection at brain level.
▼
V188 — V5.BM drift-aware Brownian motion brain (Arm A + Arm B paper)
Standalone brain that uses Black-Scholes binary pricing with realized sigma AND recent log-return drift. Compares model fair_p to Kalshi market mid. Fires when gap >= 0.15 (15c+ mispricing). Universe: BTC, ETH, SOL, DOGE, XRP. Gates: V174 (last 5min only), V117 consensus (block counter-trend), V175 v3 (skip crossed books), V128-style edge-tier sizing 1.0/1.5/2.0/2.5x. v1 (driftless) showed model wanted to bet NO on all trending markets — fundamental no-drift failure. v3 adds drift_15m estimator (rolling log-return capped at +/- 2*sigma); flipped 3 of 5 from disagreement to weak agreement with market direction.
▼
V179 — Asymmetric pyramid sizing (PYR brain)
v5_hpf_pyr_live/paper. Emits LEG1 $1 on first signal, LEG2 $1 more at secs<=300 if winning, LEG3 $1 more at secs<=120. Max $3 exposure. V183 uses signal-based MTM (bid>=0.85) — pyramids on high-entry trades. V186: V170 MID_BID_FLOOR applies to LEG 1 only — pyramid LEG 2/3 bypass it (adds to winners, not cold entries). Counterfactual showed +62% better than flat $2 ($173->$282 net per 220 trades).
▼
V114 trend filter (Arm B only)
Block counter-trend when regime=TREND
▼
V117 consensus (V133 cached)
Skip counter; boost 2x aligned. V133: consensus cached 10s in Redis so A+B see identical value (eliminates race)
▼
V128 conf-tier sizing
1.0/1.5/2.0/2.5x — but V152: blocked unless brain last-20 WR >= 80% AND V174: blocked unless secs_left<300 (last 5min). Saves us from $19 losses on losing brains AND early-contract uncertainty.
▼
Executor receives thesis
kalshi:thesis_queue
▼
Executor PER-SOURCE dedup
kalshi:placed:{source}:{ticker} — separate paper/live
▼
V167 HPF live CROSS-SOURCE dedup
TAKER and MAKER live brains share the underlying Kalshi account — V167 checks kalshi:placed:hpf_live_xside:{ticker}:{side} to prevent both from betting on same ticker+side. Wasted-capital prevention.
▼
V130 streak-pause drop (paper only)
If 3 consecutive losses, drops emit for 1hr
▼
V44 slippage check (15% cap)
Rejects orders if live price diverges from thesis
▼
Position opens
PAPER: simulated fill. LIVE: V142 polls Kalshi /portfolio/orders for ACTUAL fill price after place_order, updates position record with REAL entry. No more phantom PnL.
▼
V124 resolution-only (Arm B + V5-family + HPF)
Hold to settle; no early TP/trail/dd-stop
▼
V5 trail (Arm A + B)
HWM>=0.80 + 45% give-back
▼
HPF 40c stop (Arm C only)
Force-close if price < 0.40
▼
V138 Truth Reconciler
Polls Kalshi /portfolio/settlements + /portfolio/fills every 60s. Updates exit_log with realized_pnl_true. Maintains kalshi:v5_hpf_live:bankroll:truth (authoritative for live PnL).
▼
V164 Phantom Reconciler
Polls Kalshi /portfolio/positions every 60s. Deletes Redis position records older than 90s that don't match any Kalshi open contract. Backs up every deletion to kalshi:phantom_cleanup:auto:{ts}. Fixes dashboard showing stale open positions when maker orders never filled.
▼
Bankroll mirror updates
Per-source: kalshi:{brain}:bankroll incrbyfloat

A vs B - The Single Designed Difference

TREND_FILTER_ENABLED

Arm A = 0 (off)   |   Arm B = 1 (on)
When this is the only diff, A vs B PnL delta = pure V114 trend-filter EV. Everything else is identical: same brain code, same V117/V126/V127/V128 stack, same exits.

Sizing Tier Tables

V128 (Arm A & B)

ConfidenceTierSize
< 0.70BLOCKED0 (V126 floor — applies to all V117 brains paper + live)
0.73 - 0.751.0x$10
0.75 - 0.801.5x$15
0.80 - 0.852.0x$20
>= 0.852.5x$25
V117 aligned + tierx V117 boost, cap 3.0xmax $30 (observed exceed: investigate)

HPF bid-tier (Arm C)

Triggering bidTierSize
V172 (current)FLAT $2 — ladder removedEnv: V172_FLAT_SIZE_USDC default 2.0

LIVE Deployment Config

SettingWhat it controlsCurrent value
V242 hybrid bypass V167 dedupExtends V232 asymmetric pattern to hybrid_live. Now all 3 HPF live brains (taker + hybrid + pyr) can fill same ticker+side concurrently. Only taker SETS the peer lock. Hybrid + pyr both bypass entirely.ACTIVE — verified V242 BYPASS firing in executor.err at 21:12; expected +$30-60/day on hybrid from 5x fill rate jump
V126 PAPER LOOSEN (REVERTED)Experiment 21:07-21:21 EDT: lowered V126 paper floor from 0.70 to 0.65. REVERTED after first cycle showed 3 losses totaling -$60 (all in newly-unlocked band, V117 boost amplified to 2x, V5 trail did not arm because entries below 0.80). Now back at 0.70 on all 20 paper brains.REVERTED 21:21 EDT — commit c1b389d
V241 V5_HPF_EMIT_TTLBrain re-emits a (ticker,side) every 10s within a cycle (was 600s blocking retry, V240 60s, V241 10s)10 (ACTIVE on all 7 HPF brains — captures intra-cycle price flow, enables V232 3-way fills)
V239 V5_HPF_WIN_MINBrain emit window LOWER bound — was 60s (skipped final minute), now 10s (full 7-min window)10 (ACTIVE on all 7 HPF brains — captures previously-skipped final-minute price evolution)
V238 manual_reconciler fixes3 bugs fixed: API query=dict (was inline), bot classifier matches kalshi-/cross-/maker-, count_fp + *_price_dollars field namesACTIVE — reconciler heartbeat fresh, 0.17s/loop. Catches new manual opens within 60s. Backfill of pre-deploy manuals not supported.
V237 halt+staleness alertskalshi_env_watchdog now sends Telegram KALSHI LIVE HALT ACTIVE/CLEARED + 30+ min staleness warningsACTIVE — replaces 95-min undetected halt window (18:08-19:43 incident)
V236 halt LIVE-onlykalshi:halt check moved from process() top to AFTER thesis pop in executor; gated by paper/live flag — paper bypasses, live blocksACTIVE — preserves A/B parity during live halts (paper kept trading during 18:08-19:43 halt)
V235 manual reconcilerkalshi_manual_reconciler.py polls /portfolio/positions every 60s, records user manual trades under crypto_15min_manual_live sourceRUNNING — heartbeat 50s, classifier matches bot prefixes kalshi-/cross-/maker-; manual_live bk None (no new manuals since V238 deploy)
V234 V112.Z buffer + auto-clearKALSHI_BALANCE_FLOOR_BUFFER=30 (effective floor=$235 vs nominal $265) + KALSHI_BALANCE_RECOVERY_MARGIN=5 (auto-clear when balance > $270). Prevents transient pending-cash dips from triggering haltACTIVE — env vars now EXPLICIT in supervisord.conf executor block (was code-default before audit)
kalshi:live:balanceReal Kalshi account (from reconcile API)Currently authoritative source
KALSHI_LIVE_SOURCESSources routed to real Kalshi ordersv5_hpf_live, v5_hpf_hybrid_live, v5_hpf_pyr_live, v5_hpf_strike_live, v5_mxy_live, v5_pc_tf_live (6 brains)
V222 HYPE strict pausekalshi:coin_pause_strict:hype overrides V187 exemption (executor enforces)ACTIVE — blocks HYPE on live (HYPE was -$16/24h); BNB still allowed
V223 pyr hard_stop_ddV209C_PYR_DD_PCT on kalshi_exit_monitor — pyramid LEG 2/3 drawdown floor before hard_stop0.15 (was 0.08 — less aggressive winner kills)
V225 env watchdogkalshi_env_watchdog process polls /proc/PID/environ for 11 brains every 60s, Telegram on mismatchRUNNING — catches env-drift failures in <60s (designed after V220 zombie-env incident)
V227 KALSHI_BALANCE_FLOOR + V234 bufferV227 nominal floor $265 + V234 buffer $30 = effective floor $235. Avoids 95-min halt incident from transient pending-cash dips. Auto-clears when balance > $270.$265 floor + $30 buffer + $5 recovery margin ACTIVE
V228 SECS_MAX 600→420 + V239 WIN_MIN 60→10Brain HPF emit window is now secs [10, 420] = 6m50s (was [60,600]=9min). V228 tightened upper bound for orderbook reality; V239 lowered lower bound to capture final-minute price evolution.420s+10s ACTIVE on 7 HPF brains — full 7-min window utilized
V228b V212A_SLIP_CENTSPyramid LEG 2/3 allows current_bid >= LEG_1_entry - 1c (was strict >)0.01 (lifts pyramid completion 2-7% → 25-40% est while blocking HYPE-pattern 3c+ drops)
V230 whale paper brainkalshi_whale_paper_brain.py mirrors Polymarket crypto-whale trades to Kalshi 15M paper (reads poly:whale_mirror:* read-only)RUNNING — 88 wallets tracked, $250 min notional, $2 paper sizing, 0 crypto matches in first 35min (whales appear macro/political)
V208 strike brainWas scanning KXBTCD daily strikesKILLED V231 — STOPPED + autostart=false (0 emits in lifetime, markets near-zero liquidity, code bug)
V232 + V242 asymmetric dedupBoth PYR and HYBRID bypass V167 cross-dedup; all 3 HPF live brains can fill same signal concurrently. Only taker sets peer lock.ACTIVE — V242 deployed 21:12 EDT, immediate firing verified in executor.err
V230b whale all-trades logkalshi_whale_paper_brain logs EVERY detected trade (incl non-crypto) to kalshi:whale:all_trades_logACTIVE — gives visibility into what 88 tracked whales actually trade
KALSHI_LIVE_SIZEPer-trade cap on LIVE positions$5 (V144 — allows win-ladder full range)
KALSHI_SKIP_COINS envREMOVED in V173 (was redundant with V168 coin_pause)REMOVED 2026-05-16
V128 ladder gateWR threshold for V128 sizing (V152)0.80 (only brains with last-20 WR >= 80% can scale beyond 1.0x)
V130_LOSS_STREAKConsecutive losses -> 1hr pause3 (paper only; live HPF exempt)
V142 fill recordingActual Kalshi fill price on LIVE opensENABLED (polls /portfolio/orders post-place_order)
V138 truth reconcilerPulls Kalshi /portfolio/settlements every 60sENABLED (authoritative for live bankroll)
V164 phantom reconcilerPulls Kalshi /portfolio/positions every 60s, deletes Redis records with no matching Kalshi contract (90s grace for maker fills)ENABLED (kills stale dashboard counts)
V165 brain skip-coinsREMOVED in V173 — consolidated to V168 coin_pause (single source of truth)REMOVED 2026-05-16
V166 maker meta in exit_logexit_monitor forwards maker_mode/maker_saved/maker_v1 from position to exit_log for full auditENABLED
V167 HPF live cross-source dedupTAKER and MAKER cannot both place on same ticker+side - prevents wasted capitalENABLED via V167_HPF_LIVE_CROSS_DEDUP=1
V168 universal coin-pausekalshi:coin_pause:{coin} blocks sources by default. V187 (2026-05-17): HPF live brains exempted via V187_COIN_PAUSE_EXEMPT_SOURCES env. HPF reads Kalshi book only, doesnt need spot WS validation that V5/V117 paper brains needACTIVE: BNB+HYPE blocked for V5/V117 paper, ALLOWED for HPF live; SOL temp pause
V175 v3 crossed-market guardHPF brain rejects emit if entryACTIVE on all 7 HPF brains
V179 LIVE PYRAMIDv5_hpf_pyr_live - asymmetric $1+$1+$1 pyramid on HPF winnersRUNNING (paper + live)
V183 V179 signal MTMV179 LEG 2/3 fires when bid >= 0.85 (signal-based) not strict bid>thesisACTIVE
V186 V170 LEG 1 onlyV170 MID_BID_FLOOR moved into LEG 1 block - pyramid LEG 2/3 bypass V170 (adds to already-winning positions)ACTIVE
V189 pyramid robustness + adaptive sizingV189a tracker persists to Redis (survives brain restarts), V189b verifies position-open in Redis before LEG 2/3 emit (no more phantom adds when LEG 1 was rejected), V189c v179_pyramid_leg now flows to exit_log for leg-level audit, V189d auto-bump LIVE size from dollar1 to dollar3 after 75 lifetime wins (paper stays dollar2)ACTIVE
V188 V5.BM drift-aware Brownian brainv5_bm_paper + v5_bm_tf_paper — new standalone brain on both arms. Black-Scholes binary fair_p with realized sigma + recent log-return drift. MIN_GAP=0.15, V117+V174+V175 stack. Paper-only, isolated $0 bankrolls.ACTIVE (paper)
V187 HPF coin_pause bypassV168 coin_pause checks V187_COIN_PAUSE_EXEMPT_SOURCES env - HPF taker/hybrid/pyr live sources allowed to trade BNB+HYPE even when paused for V5/V117 brainsACTIVE
V182 V170 floor relaxedMID_BID_FLOOR 0.88->0.86, allows DOGE-style bid=0.86-0.87 tradesACTIVE
V170 HPF entry tighteningsecs_left max 600 (was 840) + bid floor 0.88 in 2-4m window. Surgical cuts to bleed cells.ENABLED on all 5 HPF brains via V170_SECS_MAX=600, V170_MID_BID_FLOOR=0.88, V170_MID_LO=120, V170_MID_HI=240
CF BenchmarksSettlement source for Kalshi crypto marketsDROPPED (V142) — public scraper 30-130s stale, paid API too expensive. WS sub-second is fresher anyway.
Kalshi Fleet Dashboard • srv1507271.hstgr.cloud • Fresh-start ts: 1778904000 • midnight EDT scheduled reset (lifetime PnL from now)