Premium Users
🚀 System Upgrade: Alpha V2 + RCS (Confidence Engine) - Now Live
Over the past few weeks, I have been rebuilding the core intelligence layer behind our signals. Today we rolled out two major upgrades designed to move us from rule-based filtering → probability-driven structure analysis.
🧠 Alpha V2 — Probability-Based Structural Ranking
Instead of simple pass/fail rules, Alpha V2 estimates how much stronger a structure is vs market baseline using:
• Historical outcome probabilities
• Caller structure quality (Edge)
• Validator / Role-chain behavior
• Early structure stability
• MC + timing context
This lets us rank opportunities instead of just labeling them.
📊 New RCS (Rolling Confidence Score)
RCS is now an evidence-based confidence model (not just a multiplier score).
It evaluates:
• Structural consistency
• Cohort outcome stability
• Time-based signal maturation
• Sample reliability
Goal: Help users understand how much to trust a signal, not just whether it looks good.
🎯 Why This Matters
Markets change. Groups change. Structures evolve.
We are moving toward:
→ Data-calibrated probabilities
→ Structure-first signal evaluation
→ Confidence-aware trade decisions
Not hype. Not noise. Not guesswork.
We have only 4 days left before launching the premium bot. Please contact me if you haven’t received the user guidebook yet.
As mentioned before, I will lock in the current price exclusively for first-month premium users.
Users who are not first-month members will be subject to a different pricing plan starting in the third month, when the dataset reaches 120 days.
All signals are data-driven.
The dApp learns from historical data, including clusters of groups from successful coins, and evaluates the quality of shilling groups and their roles (Source, Scout, Validator, Confirmer, Volume, and Noise).
The system continues to improve as more data is accumulated.
Data doesn’t lie.
What does Tier B + Alpha 100 mean?
It means two independent quality checks are both satisfied:
1️⃣ Tier B = Who is calling the coin
Tier B groups are:
- Consistently profitable over history
- Early enough to matter, but not reckless
- More reliable than Tier C/D
- Less selective than Tier A/S
2️⃣ Alpha 100 = How the coin is being built
Alpha score evaluates:
- Order of group participation (Source → Validator)
- Timing and clustering
- Early conviction vs spam
- Market structure (not price alone)
✳️Tier B alone:
Good coins
~63% X3 in 24h
But includes some noise
✳️Alpha 100 alone:
Perfect structure
But can include risky or volatile group mixes
✳️Tier B + Alpha 100 together:
✅ Credible callers
✅ Ideal buildup pattern
✅ Early enough
✅ Lower variance
✅ Repeatable performance
The real strength of this dApp comes from data accumulation.
Simply put, more data kills noise.
Early on, the market is filled with random calls, copycat groups, and short-lived hype. Those patterns may appear once or twice, but they don’t reproduce. As the dataset grows, these behaviors are naturally filtered out.
What remains are repeatable structures: the same high-quality callers, the same clusters of shilling groups, and the same role transitions (Source → Scout → Validator) that consistently precede strong price moves.
As the dApp accumulates more historical data, it doesn’t just get better at filtering noise — it becomes capable of detecting high-Tier coins and high Alpha scores earlier, at much lower market caps.
That’s because the system recognizes structure before price, volume, or late confirmation appears.
The dApp continuously learns:
- which groups repeatedly push coins to high market caps
- which clusters matter and which ones fade
- which early signals evolve into Tier A or Tier S outcomes
Weak signals lose influence. Strong structure compounds.
This isn’t about chasing momentum or reacting faster.
It’s about seeing structure earlier and more clearly.
As more data is collected, the system becomes more selective, more accurate, and more reliable which allowing high-Tier, high-Alpha opportunities to surface earlier in the lifecycle.
That’s why the dApp gets stronger over time.
Data doesn’t lie.
Day 51
Alpha score performance report (24h window)
Dataset: 25,036 coins with outcomes
Alpha : 1,204 coins
Overall X3 win rate:
• All coins: 11.33%
• Alpha coins: 32.06%
X3 win rate by Alpha score bucket:
• Score 60–69: ~25%
• Score 70–79: ~40%
• Score 80–89: ~60%
• Score ≥90: ~75%
Key insight:
Alpha score is not linear — it creates clear performance tiers.
• 60–69 = weak Alpha (structure ok, timing noisy)
• 70–79 = core Alpha (best balance)
• 80–89 = strong Alpha (Elite candidates)
• ≥90 = extreme tail (dominant)
Day 47.
Tier coins are showing better accuracy and performance. As you know, Tier C does not require cluster confirmation, while higher-tier coins (Tier B / A / S) do. Higher tiers demand a stronger history of cluster confirmation from high-quality coins.
The data is continuing to mature. Tier B coins are already showing good performance, and Tier A and Tier S are expected to emerge within 80 days.
The Validator plays a crucial role in helping users make informed decisions.
13 more days until the premium bot launch. Stay tuned. I will contact users who applied for the whitelist on Day 59, and premium users will receive a dApp guidebook, which I strongly recommend reading carefully.
1.
📊 Why This dApp Is Data-Driven (Not Hype-Driven)
Most crypto tools react to hype.
This dApp is built to measure behavior.
Every signal in the system is derived from observed data, not opinions, influencers, or paid promotions.
No signal is hand-picked. No coin is “featured.” Everything is computed.
🧠 Built on Mathematical & Statistical Models
The dApp is coded around quantitative models, including:
✅ Time-window analysis
(early / mid / late behavior instead of static labels)
✅Role classification
(Source, Scout, Validator, Volume) derived from repeatable group behavior
✅ Bayesian statistics
to avoid overfitting and reduce false positives
✅ Clustering & transition detection
to identify phase changes, not just momentary spikes
Outcome-based validation
where past behavior is continuously checked against real results
Duke retweeted
Been keeping an eye on @WIDITDOTFUN and I’m hyped fir their upcoming launch.
Widi is an AI powered copy trading app on solana built for serious traders.
– AI tracks trades in real time so you can keep up.
– Copy top traders without doxxing your wallet.
– Backed by solana foundation Grant + followed by Anatoly himself.
– partnerships with Nansen (Analytics) + GoPlusSecurity (rug protection)
iOS + Android apps are dropping soon, worth a follow.
Shoutout to the founder:
@Matty_Solana and team:
@AndyPollyAI, @Hannaah_Solana, @oldnicker , @Manoxcc, @tojidotsol
Big things coming.
Duke retweeted
Replying to @Duke_AW
I am going to burn everything IDC how much it is... 10K 100K 1M all will be burned idc
Duke retweeted
A short cut from our upcoming dApp launch clip
Your Rules. Your Pools. 🌊
Earn when your followers copies your trades.
Be WIDI’t or get buried in the trenches. 🔜
Duke retweeted
WIDI just hit the Solana Mobile Hackathon.📱
Ready to hack the duopoly WIDI't.
Beta dropping 🔜.

















