Top Picks Now — Live Tracker

🎯 The 4 Promotion Gates — ELI5 ⬅ NOT applied to picks below — these are for strategy graduation

A strategy must pass ALL 4 gates before it can trade real money. Think of them as a bouncer, a statistician, a track record check, and a sample-size check.

Gate 1 — DSR (Deflated Sharpe Ratio)
What it is: The Sharpe Ratio measures how good your returns are relative to risk (volatility). But the regular Sharpe can be gamed — if you test 100 strategies and pick the best one, it’ll look great just by luck. DSR deflates (punishes) the Sharpe based on how many strategies were tested.
ELI5: 1,000 monkeys flip coins. One gets heads 20 times in a row. Regular Sharpe says “genius!” DSR says “you tested 1,000 monkeys — expected by chance.” st_fear_greed_contrarian: DSR=19.70 ✅

Gate 2 — FDR p-value (False Discovery Rate)
What it is: p-value is the probability the strategy’s results happened purely by luck. Lower = better. p=0.05 means “5% chance this is noise.” FDR adjusts for testing many strategies at once (if you test 56, ~3 will look significant by luck).
ELI5: You’re hiring 56 job candidates. Some ace the interview by luck. FDR ranks everyone and only hires down the list until the expected flukes stay below 5%. st_fear_greed_contrarian: p=0.0000 ✅

Gate 3 — Win Rate ≥ 40%
What it is: % of winning trades out of total closed forward (live) trades. Not backtest — real money or paper.
ELI5: You don’t need to be right more than half the time — a 40% WR with 2:1 reward-to-risk still makes money. The gate just filters out clearly broken strategies. st_fear_greed_contrarian: WR=53% ✅

Gate 4 — Trade Count ≥ 30
What it is: Requires enough real trades for the stats to be meaningful. With 5-10 trades, a 100% win rate means nothing — you just got lucky.
ELI5: Flipping a coin 3 times and getting 3 heads doesn’t prove it’s rigged. Flipping 300 times and getting 200 heads? Now you have evidence. st_fear_greed_contrarian: n=430 ✅

Note: Profit Factor (PF) is tracked separately for tier classification (T1/T2) but is NOT one of the 4 gates. A strategy can be promoted with PF<1.5 if it passes the 4 gates — the tier reflects how strong the edge is.

🧩 How Methodology + Gates Work Together: These are two separate layers. The multi-factor methodology above scores each pick RIGHT NOW (e.g., “AMZN scores 133/100 today”). These are research signals — logged to picks_now_tracker for tracking, not money-ready. The 4 promotion gates evaluate whether a full strategy (like the quant_multifactor_screener itself, or sub-strategies like st_fear_greed_contrarian) has proven itself over hundreds of trades and is ready for real capital. Individual picks on this page do NOT need to pass the gates — the gates are for strategy-level graduation. Currently 0/6 asset classes have a strategy that passes all 4 gates, which is why the disclaimer says “Zero asset classes are currently Money-Ready” at the top of this page.

Model / Family	AMZN	NVDA	AAPL	META	Key Stats Cited	Framework Used
Hybrid Ensemble (large) Local proxy · multi-model blend	✅ APPROVE RSI 38, 94 analysts, EPS+75%	⚠️ CAUTION Altman-Z 51.4 great; RVOL 46% too high	✅ APPROVE Piotroski 9/9; most resilient; low vol	❌ REJECT 3m=-8.3%; Piotroski 4/9; SL risk	Altman-Z, Piotroski, RVOL, 3m momentum, EPS	quality-resilience
Cloudflare LLaMA Cloudflare Workers AI	✅ APPROVE RSI 38, 94 analysts, EPS+75%	⚠️ CAUTION 79 analysts bull; RVOL 46% flagged	⚠️ CAUTION Piotroski 9/9 but only +5.4% upside	❌ REJECT Piotroski 4/9; 3m trend weak	Piotroski, analyst count, 3m trend, upside %	quality screening
DeepSeek Chat (proxy) Local proxy · DeepSeek-direct	✅ APPROVE RSI 38 + 94 analysts = safe entry	⚠️ CAUTION EPS+215% great; RVOL 46% risky today	✅ APPROVE Piotroski 9/9 + +0.3% today = safe haven	❌ REJECT Piotroski 4/9; 3m=-8.3%	RSI, Altman-Z, Piotroski, RVOL, 3m trend	mean-reversion + quality
LLaMA 3.3 70B (Groq) Meta AI · Groq inference	✅ APPROVE RSI 38; 94 analysts STRONG_BUY	⚠️ CAUTION EPS+215% compelling; RVOL 46% → wait	✅ APPROVE Piotroski 9/9; +0.3% defensive	❌ REJECT Piotroski 4/9; 3m trend negative	RSI, analyst consensus, Piotroski, 3m trend	quality-defensive
LLaMA 4 Scout (Groq) Meta MoE next-gen · Groq	⚠️ CAUTION Strong fundamentals; risk-off day risk	❌ REJECT RVOL 46% in risk-off = catching a knife	✅ APPROVE Piotroski 9/9; low vol; best risk-off pick	❌ REJECT Piotroski 4/9; high ask in selloff	RVOL, market regime, Piotroski, SL downside	risk-parity / regime-aware
Qwen3-32B (Alibaba/Groq) Alibaba DAMO Academy	✅ APPROVE Analyst sentiment + fundamentals align	⚠️ CAUTION EPS+215% strong; sector volatile in selloff	✅ APPROVE Piotroski + analyst support + low vol	❌ REJECT Weak Piotroski + high valuation ask	Analyst consensus, Piotroski, RVOL, valuation	fundamental screening
DeepSeek Chat (API direct) DeepSeek · direct API key	✅ APPROVE RSI 38 + 94 analysts = compelling entry	⚠️ HOLD EPS+215% offset by 46% vol; RSI neutral	✅ APPROVE Perfect quality + positive relative strength	❌ REJECT Weak fundamentals + negative momentum	RSI, Altman-Z, Piotroski, 3m trend, analyst count	quality-momentum

📋 Expand: AI verdicts for remaining 17 picks (5 models) click to show

Sym	Class	Score	Verdict	Key Stat	Methodology	Model reasons
AVGO	EQUITY	128	✅ 5/5	RSI 39.9 oversold	momentum+quality	5 approve: oversold + analyst target +33% ↓ risk-off
GBPUSD=X	FOREX	92	✅ 5/5	DB_n=114 WR=58.8% edge	DB-evidence	5 approve: DB-corroborated WR=58.8% n=114, only screener pick with real proof
AMD	EQUITY	100	⚠️ 2/5	DB_n=29 DB_WR=41.4%	momentum+quality	2 approve: best WR in semi universe; 3 reject: 130% 3m = extended
DIS	EQUITY	96	⚠️ 2/5	RSI 41 near oversold	quality	2 approve: pullback in downtrend; 3 reject: -1.9% 3m = no trend
MA	EQUITY	96	⚠️ 2/5	-5% 3m = cheap entry	quality	2 approve: defensive payments; 3 reject: -5% 3m downtrend avoiding
V	EQUITY	100	⚠️ 2/5	+0.9% 5d relatively flat	quality	2 approve: stable; 3 reject: unexciting in selloff, no catalyst
GOOGL	EQUITY	121	❌ 1/5	DB_n=7 WR=28.6%	momentum+quality	4 reject: 20% 3m gain = extended; analyst target already baked in
IWM	ETF	96	❌ 1/5	DB_n=0 no history	momentum	4 reject: small-cap etf gets hit hardest in selloffs
LRCX	EQUITY	98	❌ 0/5	DB_n=0 no history	momentum	0/5: 43% 3m gain extended, semi equipment in selloff = avoid
MSFT	EQUITY	90	❌ 1/5	-7.5% 5d sharp drop	quality	4 reject: -7.5% 5d in selloff, no DB proof; 1 approve: buy dip
MU	EQUITY	100	❌ 0/5	DB_n=0 no history	momentum	0/5: 122% 3m gain massively extended; no DB proof; semi selloff
QQQ	ETF	96	❌ 1/5	RSI 48 neutral	momentum	4 reject: tech ETF gets hammered in risk-off; RSI neutral
TXN	EQUITY	95	❌ 0/5	46% 3m extended	momentum	0/5: 46% 3m run = extended; semi sector; 0 DB proof
UBER	EQUITY	91	❌ 0/5	-4.2% 3m = downtrend	quality	0/5: -4.2% downtrend, unprofitable in risk-off, avoid
UNH	EQUITY	94	❌ 0/5	RSI 68 near overbought	momentum	0/5: RSI 68 near overbought; 40% 3m gain extended; no proof
VGT	ETF	96	❌ 1/5	26% 3m extended	momentum	4 reject: tech ETF + extended 3m + no DB proof
AMAT	EQUITY	96	❌ 0/5	33% 3m extended	momentum	0/5: semi equipment, extended, bad entry on risk-off

5 models: Hybrid Ensemble · Cloudflare LLaMA · DeepSeek Chat · LLaMA 3.3 (Groq) · Qwen (Groq). Only AVGO and GBPUSD=X had unanimous APPROVE across all models in this batch. Evaluated: June 6, 2026.

📊 Multi-Model AI Review Panel (informational — does not change scores)

Each pick was evaluated by 7 independent AI models. Verdicts are informational — pick scores unchanged. Evaluated: June 6, 2026 00:37–01:15 EST.

AMZN · $246.03 LONG

✅ 6/7 APPROVE · 1/7 CAUTION

Why strong buy: 3-month uptrend + 5-day flush = classic TREND+DIP. RSI 38 oversold. 62+ analysts STRONG_BUY. Piotroski 6/9, Altman-Z 5.1 safe. Average target $307 = 25% above today.

✅ Hybrid Ensemble: quality-momentum — "strong fundamental growth, oversold RSI suggests bounce"

✅ Cloudflare LLaMA: momentum+quality — "analyst consensus + EPS growth support bounce"

✅ DeepSeek Chat (proxy): mean-reversion+quality — "RSI 38 oversold + 94 analysts STRONG_BUY"

✅ LLaMA 3.3 70B: quality-momentum — "oversold + analyst consensus in risk-off"

✅ DeepSeek Chat (direct): quality-momentum — "key stats: RSI 38 oversold, 94 analysts STRONG_BUY"

⚠️ paid-mode-large (proxy): CAUTION/HOLD — "strong fundamentals but risk-off day makes entry risky; wait for clearer direction"

✅ Others (2 models): APPROVED — cited analyst consensus + oversold RSI

NVDA · $205.10 LONG

⚠️ DISPUTED · 5/7 CAUTION/HOLD · 1 REJECT · 1 APPROVE

Why split: EPS +215% YoY + Piotroski 7/9 + Altman-Z 51.4 (exceptional) vs 46% RVOL (highly volatile). In a risk-off market, high-vol names get sold first. The fundamentals are great but the entry timing is debated.

✅ LLaMA 3.3 70B: APPROVE — "EPS+215% + exceptional Altman-Z outweigh vol concern"

⚠️ Hybrid Ensemble: CAUTION — "exceptional financial health and growth, but high RVOL 46% warrants caution"

⚠️ Cloudflare LLaMA: CAUTION — "high analyst consensus + exceptional Altman-Z but high vol is a concern"

⚠️ DeepSeek Chat (proxy): CAUTION — "EPS+215% + Altman-Z 51.4 compelling but RVOL 46% on risk-off day"

⚠️ DeepSeek Chat (direct): HOLD — "exceptional earnings offset by elevated volatility"

❌ paid-mode-large (proxy): HOLD — "strong fundamentals overshadowed by risk-off sentiment; SL at $190 vulnerable"

⚠️ 1 additional model: CAUTION — volatility too high for current regime

AAPL · $307.34 LONG

✅ 6/7 APPROVE · 1/7 CAUTION (safety over upside)

Why near-unanimous: Perfect 9/9 Piotroski. Most resilient stock today (+0.3% vs market -3%). Lowest equity RVOL (18.5%). All models said: if you have to buy something today, AAPL is the one — it's not going to crash.

✅ Hybrid Ensemble: quality-resilience — "perfect Piotroski 9/9 + +0.3% today = ideal defensive play"

✅ DeepSeek Chat (proxy): risk-parity — "lowest vol + perfect quality score = safest haven in selloff"

✅ LLaMA 3.3 70B: quality-defensive — "perfect Piotroski + resilience in down market = safe-haven"

✅ DeepSeek Chat (direct): quality-momentum — "perfect quality score + positive relative strength in down market"

⚠️ Cloudflare LLaMA: CAUTION — "perfect score supports bounce, but +0.3% performance not impressive as a BUY signal"

✅ 2 additional models: APPROVED — cited 9/9 Piotroski + lowest vol

META · $593.00 LONG

❌ 7/7 REJECT · unanimous rejection

Why ALL 7 models rejected: Piotroski 4/9 is the weakest financial health score in the entire pick universe. The 3-month trend is DOWN (-8.3%) — this is NOT a dip in an uptrend, it's an actual downtrend. Paying $593 with shaky fundamentals into a market selloff was deemed the worst risk/reward of all 4 picks.

❌ Hybrid Ensemble: quality-momentum — "negative 3m trend + 4/9 Piotroski + SL -5.6% = unsuitable"

❌ Cloudflare LLaMA: quality — "weak Piotroski + poor recent performance outweigh strong analyst consensus"

❌ DeepSeek Chat (proxy): quality-momentum — "weak fundamentals + negative momentum outweigh high target"

❌ LLaMA 3.3 70B: trend-valuation — "weak Piotroski + negative trend = unattractive on risk-off day"

❌ DeepSeek Chat (direct): quality-momentum — "weak basics + negative trend unsuitable for risk-off"

❌ paid-mode-large (proxy): — "4/9 Piotroski + 3/3 AI rejections = asymmetric downside to $560"

❌ 1 additional model: REJECTED — same reasoning: Piotroski 4/9 + 3m negative momentum

Methodology: Each model was given the same 4 equity picks with full fundamental data (price, TP, SL, RSI, momentum, Piotroski, Altman-Z, analyst consensus, EPS growth, volatility). Models rated independently via local proxy (Hybrid Ensemble, Cloudflare LLaMA, DeepSeek Chat (proxy), paid-mode-large, LLaMA 4 Scout, Qwen3-32B) and direct APIs (LLaMA 3.3 70B via Groq, DeepSeek Chat via deepseek.com). The key debate was timing vs fundamentals — all models agreed the long-term case is strong for AMZN/NVDA/AAPL, but split on whether this risk-off day is the right entry.

💰 Money-Ready Criteria — how close each asset class is to full qualification

A class must pass ALL 6 gates to be money-ready. Data from money_ready_verdict.json (live DB: ejaguiar1_stocks.at_raw_picks). 0/8 pass. EQUITY needs n=29 more resolved picks; CRYPTO needs WR + PF improvement.

Class	n (≥100)	WR (≥50%)	PF (≥1.5)	DSR	MDD	Recency	Gates Passed	Verdict
EQUITY closest to T2	71 ❌ −29	54% ✅	1.84 ✅	✅	❌	✅	3/6	INSUFFICIENT_DATA
CRYPTO n OK but weak WR/PF	171 ✅	48% ❌ −2pp	0.95 ❌	❌	❌	✅	1/6	NOT_READY
COMMODITY new candidates emerging	15 ❌ −85	40% ❌	1.10 ❌	✅	✅	✅	3/6	INSUFFICIENT_DATA
ETF	18 ❌ −82	33% ❌	0.71 ❌	❌	❌	✅	1/6	INSUFFICIENT_DATA
FOREX	22 ❌ −78	18% ❌	0.04 ❌	❌	❌	✅	1/6	INSUFFICIENT_DATA
FUTURES	15 ❌ −85	13% ❌	0.41 ❌	❌	✅	✅	2/6	INSUFFICIENT_DATA
BOND no data	0 ❌	—	—	—	—	—	0/6	INSUFFICIENT_DATA

🎯 Closest Edge Candidates — detailed breakdown

How each candidate scores on the 6 money-readiness gates:

Candidate	n (≥100)	WR (≥50%)	PF (≥1.5)	DSR	MDD≤20%	Recency	Gap to money-ready
stocks_rsi2_pullback LONG EQUITY · strategy (multi-symbol)	894 ✅	58.8% ✅	2.68 ✅	✅	✅	❌	RECENCY FAIL 0 new picks in 8 days (last emission May 29). 5/6 gates passed except emissions. ⚠️ DARK — emission pipeline broken, 14d WR collapsed to 29.9%.
GBPUSD=X LONG FOREX · AlphaEngine (single-symbol)	114 ✅	58.8% ✅	—	—	✅	✅	PF/DSR UNKNOWN Meets n≥100 + WR≥50% + recency. PF estimate needs live PnL tracking. Current: entry 1.3336, TP 1.3536, SL 1.3203. ⚠️ Single-source (AlphaEngine only).
RENDERUSDT inverse_ml SHORT CRYPTO · actively emitting	15 ❌ need 85 more	80% ✅	7.7 ✅	—	—	✅	SMALL-n SAMPLE WR/PF excellent but n=15 only. Actively emitting (last: Jun 06 03:51 UTC). At current rate, needs ~3-4 weeks to reach n=100. Best T2 candidate for CRYPTO. 🔵 WATCH.

🏆 AI Tournament T1 — What the best models do differently

Forward-tested picks, n=43-89 resolved per model. Ranked by combined WR + PF score.

cursor_agent Rank #2 · 66.1% WR

PF 2.35 · n=59 resolved

Highest WR of all models — balanced across classes.
• Avg win about +2.3%, avg loss about +1.0% = positive expectancy across the board
• Best: IWM LONG (100% n=5), BTCUSDT SHORT (100% n=2), META SHORT (100% n=2)
• Worst: ETHUSDT LONG (0% n=2), GC=F LONG (0% n=2), IEF LONG (0% n=2)
• Only model with no single-symbol WR disaster (all above 20%)

deepseek_v4 Rank #1 · PF 3.72

55.8% WR · n=43 resolved

Highest PF — few big wins offset losses.
• Avg win about +4.8% (about 2x bigger than other models' avg wins)
• Best: QQQ LONG (100% n=2), OJ=F LONG (50% n=2), USDCAD SHORT (50% n=2)
• Worst: NEARUSDT SHORT (0% n=2), BND SHORT (0% n=2), IEF LONG (33% n=3)
• Risk: lower n (43) = wider confidence interval; PF may regress with more emissions

llama4_scout Rank #3

61.4% WR · PF 2.26 · n=57

Solid but unremarkable — similar profile to cursor_agent.
• Best: SPY LONG (100% n=3), APTUSDT LONG (100% n=2), NZDUSD SHORT (100% n=2)
• Worst: MSFT LONG (0% n=3), BTCUSDT LONG (0% n=3), QQQ SHORT (0% n=2)
• Slightly lower WR on top picks vs cursor_agent

grok3 Rank #4

58.4% WR · PF 2.02 · n=89

Most picks, lowest PF — spread too thin.
• Makes about 2x more picks than other models (89 resolved, 113 total)
• Best: GBPUSD LONG (100% n=2), EURUSD SHORT (100% n=2), BTCUSDT SHORT (66.7% n=3)
• Worst: ETHUSDT LONG (0% n=4), SPY SHORT (0% n=4), BND LONG (0% n=2)
• Many low-conviction bets drag down PF; 47% still OPEN

Takeaway: cursor_agent wins by being selective (n=59) with balanced WR/PF. deepseek_v4 gets highest PF by going for larger winners (avg win 4.8%) on fewer bets. grok3 dilutes its edge by casting a wider net (113 total picks). Source: ejaguiar1_stocks.tournament_picks. Full report: ai-tournament.html

⚠️ NOT FINANCIAL ADVICE. Research only. All picks tracked in ejaguiar1_stocks.picks_now_tracker. Data: yfinance + at_pick_outcomes (39,880 resolved trades). Auto-refresh every 10 min.

📊 Top Picks Right Now — Multi-Asset Quant Screener

🧠 How Each Pick Is Generated — Full Multi-Factor Methodology ⬅ THIS IS THE SCORING SYSTEM FOR PICKS BELOW

🎯 The 4 Promotion Gates — ELI5 ⬅ NOT applied to picks below — these are for strategy graduation

📉 What does "Risk-Off -3% to -6%" mean?

🎯 How TP and SL work (ELI5)

🎿 Safety Tiers — Like a Ski Hill