SPHINX EDGE
"Sphinx Edge doesn't predict the market. It reads the room and only acts when the odds are in your favor."

Performance — paper trading validation

Every result on this page is simulated (paper). No live capital is at risk yet — bots earn funding by surviving validation.
Net P&L
+$5,361.17
17 paper trades since May 18
Win rate
71%
12 wins / 5 losses
Profit factor
3.4x
Gross wins / gross losses
Largest win
+$1,732
SOXS — June 5 semis selloff
Week of June 3–6: the Nasdaq and semiconductors sold off hard — the most violent week of the validation window. The fleet closed 5 trades: 3 wins, 2 losses, net +$1,137.58. Every dollar of profit came from the bear-side entries (SOXS, FNGD). One loss came from an overnight gap on a 3x ETF blowing through its stop — a structural risk we've now documented and are watching. A trend-resume entry upgrade shipped to all three rotation bots on June 6.

Trade results

All 17 closed paper trades, in order. Wins and losses.
+$117BTC 1
+$79BTC 2
+$130BTC 3
+$109BTC 4
-$69BTC 5
-$142BTC 6
+$575TQQQ 1
+$149TQQQ 2
+$459TQQQ 3
+$1,165FNGU 1
-$109TQQQ 4
+$1,687SOXL 1
+$59TQQQ 5
-$678SOXL 2
-$1,226TQQQ 6
+$1,732SOXS 1
+$1,251FNGD 1

Active strategies

What's running right now, and what each bot's own ledger says.
Running

ETF rotation — SOXL / SOXS

3x semiconductors — bull AND bear — MACD trend following + trend-resume (v2)
Trades3
Win rate67%
Net P&L+$2,741
Running

ETF rotation — TQQQ / SQQQ

3x Nasdaq — bull AND bear — MACD trend following + trend-resume (v2)
Trades6
Win rate67%
Net P&L-$91
Running

ETF rotation — FNGU / FNGD

3x FANG+ — bull AND bear — MACD trend following + trend-resume (v2)
Trades2
Win rate100%
Net P&L+$2,417
Running · data collection

Alt scanner — 8 coins

Trend-structure conviction scoring (0–10) on 8 Coinbase altcoins. Trades nothing — logs every signal for outcome analysis.
DeployedJune 6 (v2 scorer)
First checkpoint~June 11–13
Funding barSignals must beat fees
Rebuilt · redeploying

ETH short — v2

Coinbase perps, dry run. v1 retired at 19 trades / 21% win rate — fixed stops were narrower than ETH's normal noise. v2: ATR-sized stops, breakeven + trailing locks, trend-resume re-entry.
v1 final record19 trades · 21% WR
v2 statusPending relaunch

Retired strategies

Killing bad strategies is the strategy. These failed validation and will never trade real money.
Retired June 6

LSTM BTC bot

Failed its June 1 validation gate: thin margins, weak entries, spread losses. The gate existed for exactly this — no funding.
Retired June 2

BTC short v1

A scan-aware retest exposed look-ahead bias in the original backtest: claimed +72.8% ROI re-tested at -9.2%. Retired before risking a dollar. Full story on the Stress Test page.
Parked June 6

BTC long (Binance.US)

Logic behaved correctly (refused to buy a downtrend) but the venue's liquidity is too thin post-2023 and the stop sizing shared v1's flaws. May return on Coinbase using the v2 chassis.
Retired May

HMM regime bot

Market regimes flipped every 3–5 bars — too unstable to trade on. Concept shelved permanently.

Complete trade log

Paper account. Entry to exit, every position.
#BotEnteredEntry $ExitedExit $HoldP&L
1BTCMay 18, 7:42 AM$76,075May 18, 8:44 AM$76,3961h 2m+$117.73
2BTCMay 18, 2:18 PM$76,530May 18, 3:37 PM$76,7491h 19m+$79.94
3BTCMay 20, 9:14 AM$77,241May 20, 11:04 AM$77,6031h 50m+$130.82
4BTCMay 20, 1:22 PM$77,174May 20, 3:10 PM$77,4761h 48m+$109.28
5BTCMay 21, 10:35 AM$77,239May 21, 12:20 PM$77,0461h 45m-$69.98
6BTCMay 22, 11:08 AM$76,188May 22, 12:20 PM$75,8351h 12m-$142.36
7TQQQMay 21, 1:15 PM$77.01May 22, 9:30 AM$78.6020h 15m+$575.58
8TQQQMay 26, 3:25 PM$82.14May 27, 9:35 AM$82.5818h 10m+$149.16
9TQQQMay 28, 11:15 AM$83.04May 29, 9:41 AM$85.0022h 26m+$459.81
10FNGUMay 28, 11:15 AM$32.39May 29, 9:31 AM$33.7522h 16m+$1,165.30
11TQQQJun 1, 11:55 AM$86.07Jun 2, 9:31 AM$85.6121h 36m-$108.57
12SOXLJun 1, 12:23 PM$226.50Jun 2, 9:31 AM$240.0021h 8m+$1,686.88
13TQQQJun 2$86.87Jun 3, 9:37 AM$87.05overnight+$58.62
14TQQQJun 4, 12:55 PM$85.28Jun 5, 9:31 AM$81.6920h 36m-$1,225.89
15SOXLJun 4, 2:16 PM$272.18Jun 4, 3:56 PM$263.131h 40m-$678.38
16SOXSJun 5, 9:31 AM$5.83Jun 5, 12:17 PM$6.332h 46m+$1,732.00
17FNGDJun 5, 9:31 AM$34.07Jun 5, 12:31 PM$36.183h 0m+$1,251.23
Trade 14 is the lesson of the week. A TQQQ long held overnight gapped down past its -3% stop and exited at -4.2%. Stops cannot protect a position across a market close — on 3x leveraged ETFs, overnight gap risk is structural. We're tracking whether it repeats before deciding on an end-of-day flattening rule.

Dry run log — crypto short bots

Full transparency: these signals fired without placing orders, on simulated positions, and are not in the main P&L above. Shown because hiding losses would defeat the point.
#BotEnteredEntry $ExitedExit $HoldReasonP&L %
1BTC DRYMay 27, 1:00 PM$75,073.94May 27, 5:01 PM$75,123.974.0hprofit_protection-0.07%
2BTC DRYMay 28, 10:06 AM$72,577.83May 28, 10:14 AM$73,137.000.1hhard_stop-0.77%
3BTC DRYMay 28, 11:20 AM$72,572.52May 28, 12:14 PM$72,968.670.9hhard_stop-0.55%
4ETH DRYMay 29, 9:01 AM$1,995.37May 29, 11:06 AM$2,004.802.1hprofit_protection-0.47%
5ETH DRYMay 29, 6:05 PM$2,005.72May 29, 9:22 PM$2,015.993.3hhard_stop-0.51%
6ETH DRYMay 30, 12:09 AM$2,009.11May 30, 2:35 AM$2,016.962.4hloss_cut-0.39%
7ETH DRYMay 30, 3:35 AM$2,013.41May 30, 6:30 AM$2,016.272.9hloss_cut-0.14%
8ETH DRYMay 30, 7:15 PM$2,018.18May 30, 9:17 PM$2,028.882.0hhard_stop-0.53%
9BTC DRYMay 30, 7:39 PM$73,676.18May 30, 9:16 PM$74,051.711.6hhard_stop-0.51%
10BTC DRYMay 31, 3:33 AM$73,739.78May 31, 12:25 PM$73,457.168.9hprofit_protection+0.38%
What happened next: the ETH short ran on through June 5 and finished v1 at 19 trades, 21% win rate — against a backtest that claimed 63%. The autopsy found the backtest graded stops on hourly closes while the live bot checks every 2 minutes, so the backtest never saw the intrabar wiggles that killed real trades. v1 is retired. v2 ships with ATR-sized stops, breakeven and trailing locks, and trend-resume re-entry — and gets re-validated at 2-minute granularity before any funding talk. The BTC short was retired outright after the same class of retest. Details on the Stress Test page.

Validation window

May 18 → June 15 (extended). Day 19.
Closed trades
17
Across all stock bots
Window net
+$5,361
Paper, all results included
June 15 gate
Sample first
Rotation bots need 10–15 more trades before any funding decision — if the sample is thin, the date moves, not the standard.