BTC 15M Predictor

BTC / USD15M TERMINAL

CONNECTING

Layout mode: drag cards by the dotted handle, use 1×/2×/3× to change width, and drag the bottom edge to change height. Cards snap into the dashboard grid automatically. Layout is saved on this device.

BTC LIVE PRICE — Waiting for spot

LOCKED TARGET — No target locked

ACTION FLAG WARMING UP Signal — • Confidence —

TARGET PROBABILITY

ABOVE—

BELOW—

WINDOW REMAINING — Close —

BITCOIN / USD

—

0 feeds • WAITING FOR SPOT —

PRICE LOCK — backup — REST check —

PRICE GRID • WHEEL OR BUTTONS TO ZOOM

15M VIEW

NEXT 15 MINUTES

WARMING UP

ACTION FLAG

CALCULATING

—SIGNAL

Waiting for enough live data.

UP —

DOWN —

FLAT —

LIKELY PRICE—

80% RANGE—

EXPECTED MOVE—

CONFIDENCE—

MARKET STATE—

UP vs DOWN EDGE—

DATA QUALITY—

15-MINUTE TARGET PRICE

Will BTC finish above or below this price?

SET TARGET

TARGET UNLOCKED

Set a target to hard-lock it.

ABOVE TARGET—

BELOW TARGET—

MOST LIKELY—

PROBABILITY EDGE—

TARGET—

MOVE NEEDED—

DISTANCE NOW—

CONFIDENCE—

Enter a BTC price to calculate the probability that the 15-minute forecast finishes above or below it.

MINI PRICE PATH • 15M WINDOW

Ballpark checkpoints toward the real close

LIVE

3 MIN——

5 MIN——

7 MIN——

9 MIN——

11 MIN——

CURRENT MINUTE—

PROJECTED 15M CLOSE—

PATH BIAS—

Future checkpoint prices are intentionally ballpark estimates, not exact ticks. Past checkpoints switch to observed prices when available.

15M TRAJECTORY MODEL

ELAPSED—

REMAINING—

REAL-WORLD 15M CLOSE —

START—

5 MIN—

10 MIN—

13 MIN—

REQUIRED VELOCITY—

RECENT VELOCITY—

WEIGHTED VELOCITY CLOSE—

LINEAR REGRESSION CLOSE—

ACCELERATION CLOSE—

COMBINED MATH CLOSE—

MATH WEIGHT IN FINAL TARGET MODEL —

TARGET OCCUPANCY & PERSISTENCE MODEL

BUILDING

0/100

TIME ABOVE—

TIME BELOW—

LAST 3 MIN—

CURRENT RUN—

CROSSINGS—

RECLAIMS—

MARKOV—

OCCUPANCY MODEL—

DISTANCE Z—

MODEL WEIGHT0%

Waiting for a locked target and trajectory samples.

CF BRTI PROXY

—

BRTI PROXY • NOT OFFICIAL CF BRTI

STARTING

Reconstructing an estimated CF BRTI from constituent-exchange order books.

VS COINBASE—

VS KALSHI TARGET—

QUALITY—

ACTIVE—

BIAS—

ANCHOR MAE—

CF SETTLEMENT PROXY • FINAL MINUTE —

ESTIMATING

Collecting canonical 1-second proxy samples.

BRTI PROXY—

KALSHI TARGET—

RUNNING 60s AVG—

SAMPLES—

REMAINING—

REQUIRED REMAINING AVG—

SETTLEMENT CUSHION—

PRESSURE—

CALIBRATION ERROR—

SAMPLE QUALITY—

ESTIMATE • NOT OFFICIAL BRTI • KALSHI market.result STILL GRADES THIS MARKET

CONSTITUENT ORDER-BOOK FEEDS

PROXY INTERNALS

CALIBRATION STATISTICS

ERROR BY FEED COMPOSITION

ENSEMBLE INTEGRATION

SHADOW MODE • model weight 0%

This is an independent reconstruction from publicly available constituent order books and the publicly documented BRTI methodology. It is not the official CF Benchmarks BRTI, no CF Benchmarks page is scraped, and it never grades a Kalshi market — the official Kalshi market.result remains the only settlement authority.

KALSHI BTC 15M SCALP ENGINE

Searching KXBTC15M…

—

API host —

AUTO ROLLOVER ON • Waiting for Kalshi.

CONNECTING

KALSHI TARGET—

REAL CLOSE—

ABOVE / UP

Kalshi bid—

Kalshi ask—

Our fair probability—

Net model edge—

BELOW / DOWN

Kalshi bid—

Kalshi ask—

Our fair probability—

Net model edge—

SCALP SIGNAL

SEARCHING

Waiting for the live Kalshi market and our target probability.

BEST SIDE—

NET EDGE—

ASK / ENTRY—

EST. FEE / CONTRACT—

MIN EDGE PTS STRONG EDGE PTS TAKE PROFIT % STOP LOSS % CONTRACTS

AUTO KALSHI ROLLOVER KALSHI RELAY ENDPOINT

Browser market requests use the bundled same-origin Cloudflare relay at /api/kalshi only.

Signal-only integration: this browser does not store a Kalshi private API key or submit orders. Track the price you actually paid so HOLD / TAKE PROFIT / EXIT uses a real entry basis.

10-MINUTE ENTRY DECISION

Final 5-minute ABOVE / BELOW call

BUILDING DATA

CALCULATING

The bot waits until roughly minute 10 so the trajectory, learned cases, velocity, order flow, and target distance have enough information.

DECISION IN—

SIDE PROBABILITY—

ENTRY STRENGTH—

LEARNED SUPPORT—

WAITING FOR 10:00

No entry call is issued before the data gate opens.

KALSHI MARKETSEARCHING

SCALP CALL—

NET EDGE—

BRTI PROXY—

KALSHI TARGET—

PROXY GAP—

COINBASE GAP—

CALIBRATION BIAS—

PROXY QUALITY—

ANCHOR MAE—

DYNAMIC RISK EXIT / STOP ZONES

SET TARGET

Risk levels will mirror the favored ABOVE/BELOW thesis.

WAITING

LEVEL 1 • CAUTION — —

LEVEL 2 • PROTECTIVE EXIT — —

LEVEL 3 • HARD EXIT — —

ADVERSE ANCHOR—

REMAINING σ—

NEAREST STRUCTURE—

THESIS PROBABILITY—

These model-generated levels tighten with remaining-time volatility and nearby market structure.

SELF-LEARNING TARGET MEMORY

Wins + losses become future evidence

0 CASES

RESOLVED0

WINS0

LOSSES0

ACCURACY—

SIMILAR-CASE ALGORITHM No learned adjustment yet.

The learner activates after resolved target cases accumulate.

10M ENTRY WINS0

10M ENTRY LOSSES0

10M ENTRY ACC.—

ENTRY SAMPLE0

SIMILAR CASES0

HIST. ABOVE—

AVG ERROR—

LEARN WEIGHT0%

RECENT RESOLVED TARGETS

WHY ABOVE / WHY BELOW

Live evidence for the target

CURRENT READ Set a target to generate evidence.

Confidence — Data quality —

WHY ABOVE IS POSSIBLE

WHY BELOW IS POSSIBLE

COPY-READY DATA BLOCK

Waiting for live data and a target price.

MEASURED PERFORMANCE

RESOLVED0

DIRECTION—

HIGH CONF.—

RETURN MAE—

1m candles loaded0

Model warm-up0%

Last update—

Live price source—

Price feed age—

Authority lock age—

REST consensus—

The local adaptive ensemble changes expert weights based on resolved 15-minute forecasts stored on this iPhone.

SIGNAL RULES

BUY ABOVEmodel favors finishing above the locked target

BUY BELOWmodel favors finishing below the locked target

The bot always outputs BUY ABOVE or BUY BELOW. Signal strength tells you whether the edge is weak, moderate, or strong; low-strength calls should be treated as low-confidence model output.

PC + iOS UNIVERSAL MODE

Runs directly in Safari on iPhone/iPad and modern desktop browsers on Windows/macOS. The live BTC display prioritizes fresh Coinbase/Kraken spot quotes and rejects stale/outlier exchange packets. No Python server, exchange account, API key, or trading connection is required. Forecast history stays in the current browser. iOS may suspend WebSockets when the app is closed or left in the background, so the engine refreshes and reconnects when you reopen it.