SplashWatch — The Boring Tide Chart Engine

One-Line Pitch

The market always goes up or down. We watch why, wait for the tide to commit, then bet with it.

The Philosophy

"We don't predict the future. We read a calendar that already exists."

Every tradeable crypto moves in cycles. BTC rises after halving events. SOL surges when Ethereum gasps. Gold moves when solar activity suppresses satellite-dependent markets. These are not predictions — they are patterns that repeat because the forces behind them repeat: human greed, celestial mechanics, institutional rebalancing schedules, quarterly liquidations.

SplashWatch is not a trading bot. It is a tide reader. It watches many incoming streams — Telegram analyst channels, solar weather data, rare metals prices, Fear & Greed indices, order book physics — and waits for them to agree. When they do, the tide is committed. That is when we bet.

The moon doesn't push the tides. It's simply present. Because of its presence, the tides happen anyway. SplashWatch aspires to be that kind of presence: calm, persistent, always watching, never panicking.

The Two Things You Need to Know

1. The market is always going up or down

There is no flat. Every symbol is always in one of three states: rising, falling, or compressing (coiling before the next move). A long is a bet on a rise. A short is a bet on a fall. Compression is a signal to watch closely. Our job is simply to know which state we're in before the majority does.

2. Tides have a cause

Observable forces precede price moves. We collect those forces. We correlate them to past moves. When a known force returns, we know what the tide tends to do. The unknown is when — and our forward simulation engine is how we answer that question.

What We Watch

Crypto Universe

Symbol	Why
BTC	Reserve asset. All other markets follow. Halving cycles, ETF flows, macro fear are primary drivers.
ETH	Protocol utility. Gas costs, L2 launch sentiment, staking yields.
SOL	Speed narrative. Institutional attention tends to cycle between ETH and SOL.
XRP	Legal clarity plays. Court-driven sentiment swings. Correlated to SEC news cycles.
DOGE	Cultural momentum. Meme-sentiment driven — high volume, high speed, shorter holding windows.
LINK	Oracle infrastructure. Tends to respond to smart contract adoption waves.
ADA	Academic cycles. Development roadmap releases drive delayed price action.
AVAX	Gaming + NFT volume. Correlated to gaming season and metaverse news cycles.

Adjacent Signals

The part most traders ignore entirely. These are Lane A — the observable forces that precede the price moves in Lane B (the tradeable).

Signal	Source	Why It Matters
Kp Index (Geomagnetic Activity)	NOAA Space Weather	Historically correlated to volatility spikes. High Kp → nervous markets
X-Ray Solar Flux	NOAA GOES satellite	Extreme solar events have preceded liquidation cascades
Sunspot Cycle	NOAA Solar Cycle 25	11-year cycles correlate loosely to decade-scale sentiment in commodities
Gold Price (XAU/USD)	TradingView / Metals API	Gold rising = fear trade activated = BTC sideways or lagging
Silver Price (XAG/USD)	TradingView / Metals API	Silver moves with industrial demand — tech sector confidence proxy
Platinum / Palladium	Metals Focus API	Rare earth pressure → supply chain fear → industrial crypto correlation
VIX (Fear Index)	CBOE	Stock market fear bleeds into crypto 4-6 hours later. Reliable lag
Fear & Greed Index	Alternative.me	Extreme fear = buyers incoming. Extreme greed = top is near
US GDP Releases	Federal Reserve calendar	Macro surprise = dollar volatility = crypto volatility same day
BTC Dominance %	Binance / CoinGecko	When dominance rises, alts bleed. When it falls, alts surge (alt season)
Exchange Inflows/Outflows	Glassnode (planned)	Whale movements. Coins moving to exchanges = sell pressure incoming
Moon Phase	Astronomical calculation	Three independent studies found mild correlation to BTC 48h price action

The Intelligence Pipeline

Stage 1 — Tip Ingestion (150 Telegram Channels)

We monitor 150 Telegram channels — analyst groups, whale trackers, on-chain signal bots, news aggregators. Each message is parsed by the ML-brain for:

Symbol being discussed
Direction (long / short / neutral)
Confidence expressed (explicit or inferred from language)
Timeframe (scalp / swing / position)
Entry price mentioned (if any)
Target price mentioned (if any)

Each channel has an accuracy score — how often were their past tips followed by the predicted move? A channel calling BTC long 20 times with 17 correct moves scores highly. A channel that posts 300 tips with random accuracy scores low and is eventually demoted.

Stage 2 — Consensus Aggregation

Tips from multiple channels for the same symbol are merged into a consensus signal. The consensus score is a weighted average:

consensus = Σ(tip.direction_weight × channel.accuracy × signal.recency_weight)

A consensus above 0.72 is a Validated signal. Above 0.78 is Acting grade — the system will consider opening a position.

Stage 3 — Physics Capture

Every 5 minutes, the physics engine captures a snapshot of each tracked symbol:

Velocity — rate of price change (first derivative)
Acceleration — rate of velocity change (second derivative)
Tension — distance from rolling mean normalised by amplitude. High tension = stretched band, likely to snap back
Amplitude — size of recent oscillations (how live is this market?)
Volume Ratio — current volume vs 24h average (is something happening?)
Boomerang Signal — when tension exceeds threshold AND velocity is decelerating, the system flags a likely reversal. This is the moment.

Stage 4 — Walker Dispatch

At each KHash boundary (a deterministic clock tick), walker agents are dispatched:

Walker Type	Frequency	What It Does
Micro Scouts	Every minute	Scan order book depth, spot sudden changes
Day Striders	Every 15 min	Check tip consensus, update signal scores
Cycle Riders	Every hour	Compare current price position to historical cycle patterns
Tide Readers	Every 4 hours	Run full tidal analysis across all adjacent signals

Each walker produces a Shard — a small piece of intelligence. Shards accumulate for a symbol until there are enough to form a Conviction (a qualified belief about what the symbol will do next).

Stage 5 — Forward Simulation (Building Now)

This is the missing piece. Once we have a Conviction, we need to answer: when?

The forward simulation takes:

Current physics state (tension, velocity, amplitude)
Historical pattern library (how many cycles matched this setup?)
Adjacent signal state (what are the Lane A indicators saying?)
Current consensus direction and confidence

And runs 100 simulated futures through a stochastic model, each future seeded with slightly different timing and magnitude assumptions. The output is:

Forward Simulation Result:
- 72% probability of upward move within 12-48 hours
- Median expected magnitude: +4.2%
- Worst case (5th percentile): -2.1%
- Best case (95th percentile): +11.4%
- Confidence in timing window: 0.68
- Recommended action: LONG BTC at next dip, 12h window

This is what's currently missing. We have the inputs. We have the execution. We don't yet have the model that connects them into a time-qualified plan.

How We Make Bets

The Tidal Model

Every symbol oscillates between extremes. The oscillation has:

A mean — the centre of gravity
An amplitude — how far it typically swings
A period — how long each swing takes

When price stretches one standard amplitude above the mean, it tends to come back. When it stretches below, it tends to come back. This is not a prediction. It is physics. The oscillation exists because of the constant opposing forces of over-enthusiasm and over-fear.

The tide reader's job is to:

Know where the mean is right now
Know how stretched we currently are (tension)
Know what caused the stretch (adjacent signals, tip consensus)
Know the typical period for this symbol (when does the snap usually come?)

Then wait for the stretch to decelerate — that's the boomerang signal — and enter just before the snap.

LONG vs SHORT Decision Framework

┌─────────────────────────────────────────────────────────────────┐
│                   BET DECISION FRAMEWORK                         │
├──────────────────────┬──────────────────────────────────────────┤
│ Condition            │ Action                                    │
├──────────────────────┼──────────────────────────────────────────┤
│ Price well below mean │ LONG — stretched down, tide turns up     │
│ Tension high + decel  │ LONG or SHORT — boomerang forming       │
│ Consensus ≥ 0.78     │ Execute (direction from consensus)        │
│ Adjacent signals fear │ SHORT or stay CASH on alts              │
│ Adjacent signals calm │ LONG — fear premium deflating           │
│ BTC dominance rising  │ SHORT alts, LONG BTC if momentum holds  │
│ Multiple chan. agree  │ Higher conviction, larger position       │
│ Single tip, no physics│ Watch only, no bet                      │
│ Opposing signals      │ Stay out — tide is confused             │
└──────────────────────┴──────────────────────────────────────────┘

Position Sizing

position_size = base_size × (1 + conviction_multiplier)

Where:
  base_size = 2% of portfolio per trade
  conviction_multiplier = (consensus_score - 0.72) × 5
  
At 0.72 consensus: 1× base size = 2% of portfolio
At 0.85 consensus: 1.65× base size = 3.3% of portfolio
At 0.95 consensus: 2.15× base size = 4.3% of portfolio

Maximum single position: 5% of portfolio (hard cap). No exceptions.

The Self-Correction Loop

Every closed trade feeds back into the model:

Trade closed →
  → Update channel accuracy scores (were the tips that supported this trade right?)
  → Update tidal calibration (was the amplitude estimate correct?)
  → Update lane correlation strengths (did the adjacent signal predict correctly?)
  → Update walker patterns (which walker type saw this coming first?)
  → Adjust forward simulation model weights accordingly

This is how the system improves without being retrained. It watches itself. Each trade is a lesson. After 100 trades, the calibration is noticeably better than after 10. After 1,000, the system has seen enough tide cycles to have genuine edge.

The Architecture

┌─────────────────────────────────────────────────────────────────────────┐
│                              SPLASHWATCH                                 │
└─────────────────────────────────────────────────────────────────────────┘
                                    │
         ┌──────────────────────────┼─────────────────────────────┐
         ▼                          ▼                              ▼
┌────────────────┐      ┌────────────────────┐      ┌─────────────────────┐
│  INGESTION     │      │  INTELLIGENCE       │      │  EXECUTION          │
│                │      │                     │      │                     │
│ 150 Telegram   │      │ Physics Engine      │      │ Binance API         │
│ channels       │──►   │ Correlation Engine  │──►   │ Position Manager    │
│ Solar data     │      │ Walker Fleet        │      │ Sentinel (guard)    │
│ Metals prices  │      │ Brain Loop (5-min)  │      │ Risk Manager        │
│ VIX, F&G Index │      │ Forward Simulation* │      │ Exit Engine         │
│ Exchange flows │      │ Shard Aggregator    │      │ PnL Tracker         │
└────────────────┘      └────────────────────┘      └─────────────────────┘
         │                          │                              │
         └──────────────────────────┼──────────────────────────────┘
                                    ▼
                       ┌────────────────────┐
                       │  KRONOSVECTOR DB   │
                       │  PostgreSQL        │
                       │                    │
                       │ Tips, Consensus    │
                       │ Physics Snapshots  │
                       │ Shards, Walkers    │
                       │ Trade Journal      │
                       │ Channel Scores     │
                       └────────────────────┘
                                    │
                                    ▼
                       ┌────────────────────┐
                       │  SURFACE (RN/Web)  │
                       │                    │
                       │ Portfolio Value ↑  │
                       │ Equity Curve       │
                       │ Active Tides       │
                       │ Trade Feed         │
                       │ Deep Dive Tabs     │
                       └────────────────────┘

* = Under active development

What Exists Today

Component	Status	Notes
Telegram ingestion (150 channels)	✅ Live	~2,448 parsed tips
Consensus engine	✅ Live	Direction + confidence per symbol
Physics engine	✅ Live	Velocity, tension, boomerang detection
Binance execution	✅ Live	Real trades, IP-restricted
Walker fleet	✅ Built	Micro/15min/hourly/4h dispatch
Shard system	✅ Built	Intelligence fragment accumulation
Brain autonomous loop	✅ Live	5-minute cycle
Channel accuracy scoring	✅ Live	Weighted consensus
Freddy LLM contracts	✅ Built	Structured AI inference calls
KHash time coordination	✅ Built	Deterministic tick clock
Adjacent signal ingestion	⚠️ Partial	Kp Index + VIX. Rare metals pending
Solar cycle data	⚠️ Partial	Kp Index live. Sunspot cycle pending
Forward simulation engine	❌ Missing	The single biggest gap
Moon phase correlation	❌ Missing	Planned
Exchange inflow/outflow	❌ Missing	Glassnode integration pending
Self-correction loop	⚠️ Partial	Outcome tracking exists, weight updates missing
Surface (calm dashboard)	⚠️ Partial	Functional but noisy. Redesign pending
Trade journal	✅ Live	Trade history accessible
Equity curve	❌ Missing	PnL chart not implemented

What We're Building Next

Sprint: The Missing Intelligence Layer

Goal: Connect our existing perception to time-qualified bets

Forward Simulation Engine — stochastic model seeded with physics + consensus + adjacent signals → probability distribution of move timing and magnitude
Rare Metals Ingestor — Silver, Gold, Platinum/Palladium prices on a 15-minute cadence
Moon Phase Calculator — simple astronomical function, daily tick
Solar Cycle Feed — NOAA sunspot monthly data (existing Kp Index covers short-term, need medium-term)
Self-Correction Weight Updates — close each trade → retrain channel weights, tidal calibration, lane correlations
Equity Curve Endpoint — daily PnL series → Surface graph
Surface Redesign — one big number, one equity curve, the active tide gauges

The Deliverable

When this sprint is done:

You open the app and see your portfolio value
Below it, a 30-day equity curve going up (or down — this is not a lie machine)
3-5 active tides showing what the system is watching
A trade opens roughly every 4-8 hours. You understand why. You trust it.
When something interesting is forming, you get a nudge. Not five nudges. One.

The Bigger Picture

SplashWatch is a personal sovereign fund. It runs on your server. It connects to your exchange account. No third party has custody of your money. No algorithm has control you can't override. The surf is always there. You choose whether to stake the board.

Over time, the self-correction loop means the system learns your market — the specific channels you've subscribed to, the specific symbols you trade, the specific patterns that reliably precede the moves in your corner of the ocean. It is tuned to your tide, not a generic tide.

This is the difference between renting a boat and learning to read the sea.

Part of the Kronos Architecture→ — a sovereign intelligence infrastructure. Live at crypto.krisadamstv.com