Active tested strategyLoading…$100 trading pot · $400 protected · next —
Dan’s manual controls · 10×Owner-controlled · TP switch available
DanLoading Dan’s account…
Shared across devices
Add text to the chart2. Click or tap where it should appear on the chart.
Saved drawings on all devicesSelect one to show it on the chart.
SMA 5SMA 20Volatility bandsSet the line’s endpoint yourself, even past the last candle
AI lines loading…
FIB · swing levelsWaiting for chart data
DAN SHORT · ENTRY —TP1 —TP2 —
Building the market viewLoading deeper history…
Running continuously on Cloudflare
Seven-day live Bitcoin paper trial
Starting the live one-minute feed…
7 daysremaining
Current account value$100.00Total cash plus remaining position
Current position P/LEVEN $0.00Open-position result only
Cash secured$100.00Cash already realised and available
Remaining position value$0.00Live value of the open position
Open positionFLATWaiting for a signal
Latest actionWAITStarting now
Completed trades00 wins
Fees$0.0010× notional fees
Funding$0.00Open-position carrying cost
Profit protectionNot armedTP can be switched on or off
10× paper leverageApplied to this fresh $100 paper pot when a position is open.
TP1 · take half—Next Dan trade · 0.6% BTC move
TP2 · close rest—Next Dan trade · 1.2% BTC move
Dan’s method:Loading current exit settings…
Time
Action
Bitcoin price
Position
Trading pot
Profit / loss
Fee this action
Total fees
Explanation
Waiting for the first completed one-minute candle.
Local program · completed candles · paper-only
Zac’s autonomous Bitcoin portfolio
Starting the independent seven-day paper experiment…
7 daysLocal reviews
Current trading pot$100.00$400 protected and unavailable
Unrealized P&L$0.00No open position
Total trading fees$0.00All opening, closing and reversal fees
Zac positionFLATWaiting for its first decision
Latest decision—One decision per completed candle
Closed trades00 wins
Local decisions0No external AI API usage
Bitcoin benchmark0.00%BTC movement since experiment start
3× paper leverageApplied to the trading pot when a position is open; the protected $400 is not used.
Local controller:Our deterministic program decides from completed candles, trend, momentum, zones, Fibonacci context and transaction costs. No Jeb/OpenRouter calls or real-money orders. TP1/TP2 and fixed 10% stops are off.
Time
Local decision
Applied
BTC price
Position
Pot
Profit / loss
Fee this action
Total fees
Evidence / explanation
Waiting for the local program’s first completed-candle decision.
Local program: independent $100 paper pot · completed one-minute candles with five-minute context · seven days · 3× leverage · long and short · fees, slippage and funding included · TP1/TP2 and fixed 10% stop disabled · no real orders. This is an experimental rule system, not a validated forecast. All amounts are USD.
Indicator rules · autonomous paper experiment
Bot experiment
A separate bot recognises combinations of price momentum, RSI and trend. It has its own $100 paper pot and cannot change Dan or Jev.
The pattern classifier was trained on the hindsight entry and exit labels for this exact week. A training pass does not establish future predictive skill. No timestamps or future prices are inputs when it decides.
Every target trade, including misses
Each target must match one separate trade near the benchmark entry, earn at least 75% of its benchmark net gain, and retain at least 75% of its own observed peak net profit. Extra trades and losses stay in the results.
Month backtest: every trade
How the bot activates and exits
Two learned decision trees inspect completed candles: one chooses a long, short or wait; the other recognises exit patterns. Orders fill at the next minute’s open with fees and slippage. The live paper account uses its available pot; the fixed-stake research comparison reuses $100 without compounding. Protection queues a close at a $3 net loss, or after a 25% give-back once peak net profit reaches $10. It also closes after 72 hours. Gaps can exceed those levels.
Search attempts and checks
12,000 rule configurations were searched on the week. A separate 1,600-configuration search selected parameters using the first 20 days only, then froze them for the final 10. The final period includes the already reviewed week, so it is not a wholly untouched test. Candle chronology and decision inputs are checked for future-data leakage.
Fresh $100 trial controlled by our local program. The buttons below remain available as your manual override. 3× leverage, fees, slippage and funding apply; fixed TP and 10% stop are off.
Waiting for your account…
Separate local $100 paper trial · 3× leverage
Local swing · experimental
Our local evidence policy manages this larger-swing experiment. Its existing loss cap, profit trail and 72-hour time limit remain; fixed TP levels and the 10% stop are disabled. Experimental, not a validated forecast.
Waiting for the first fresh candle…
What changed in Jev
$5 is a minimum opportunity, not an exit command.
Hold valid winners; trail net profit after $5 rather than taking a fixed target.
Review trend, actual price, volatility, full-day range and reversal evidence.
Reject weakly scored entries; early reversals require stronger evidence.
Use a $3 close-based loss budget, a 30-minute entry cooldown and a 72-hour holding limit.
Save the exact model inputs, answer and stated evidence for future audits.
Four causal rule tests with winners allowed to run
These are rule-based proxies, not Jev itself. They now use trailing exits rather than closing at $5. The final three days were inspected previously and are not an untouched holdout. Each trade uses a fixed $100 stake; costs are included.
Four actual Jev snapshot tests
Four completed-candle snapshots test the revised larger-swing prompt. Selected cases are diagnostic, not a reliable capture-rate measurement.
Best total-profit plan with hindsight
These are maximum-profit non-overlapping trades inside the observed week, each worth at least $5 net. Perfect future knowledge chooses their entry and exit; this is an upper bound for comparison, not a live prediction.
Why the old table showed about $5 — and the bigger outcomes
The old table stopped at the first $5 crossing. Below, each of those same entries has its best later outcome anywhere in the observed week. These later outcomes overlap; they must not be added together.
Recovered trade evidence
Original answers survive for some trades. Missing records and inferred explanations are labelled in the review.
Jev experiment lab
1,000 randomized strategy backtests
Preparing the fixed-stake experiment…
0 / 1,000experiments complete
Trading-pot average—Best strategy across periods
Profitable periods—Out-of-sample hit rate
Average fees—Paid only by the trading pot
Active strategy—Waiting for samples
Jev input—Tokens measured by OpenRouter
Jev cost—Output tokens are free
Optimization rounds will appear here.
#
Prompt strategy
Test period
Jev preset
Pot return
Buy & hold
Orders
Win rate
Fees
Drawdown
Experiments will appear here as Jev completes them.
Method: 100 randomized 30-day periods × 10 strategy prompts · the $100 trading pot compounds its own gains and losses · the other $400 is never available to the strategy · 60-day warm-up · 0.10% fee + 0.05% simulated slippage per order · one-hour signals · no real orders.
Original monthly discovery
Monthly Jev scout · 1×
Thirty-day exploratory runs using one-hour candles and the original unleveraged $100 pot.
Losing entry clue1.75 visitsOversold alone was insufficient
What Jev discovered:Trade count was not the deciding factor. Fee-aware selection produced 16 orders and +17.56%, while micro-swing trading used 20 orders and lost 10.95%. Repeated lower-zone visits plus wider volatility separated winners better than RSI alone.
#
Scout prompt
Phase
Test period
Pot return
Ending pot
Orders
Wins
Fees
Scout experiments are loading.
Monthly method: one-hour candles · thirty-day test · 60-day warm-up · 1× unleveraged pot · exploratory periods differed, so this page remains discovery evidence rather than equal-period validation.
Signal-only five-minute long + short research
3× Three-Month Jev lab
Jev trades only confirmed buy and sell signals.
0 / 25three-month validations complete
Best three-month pot—Started at $100
Leading signal strategy—Five shared periods
Orders—Signal-only long and short
Trading fees—Full 3× notional
Funding—0.01% per eight hours
Liquidations—5% maintenance margin
Validation is running:Five strategies are being compared on the same five weeks.
#
Weekly strategy
Shared week
Start date
Pot return
Ending pot
Orders
Wins
Fees
Five-minute weekly experiments are loading.
Three-month method: five-minute candles · trades only on confirmed buy-zone reclaim, sell-zone rejection, or momentum-failure signals · 90-day test with the previous week as context · SMA 5/20/50/200 · EMA 9/21 · volatility bands · buy/fair-value/sell zones · VWAP · MACD · ATR · stochastic · RSI · volume and multi-period returns · 3× paper leverage with full-notional fees, slippage, funding, maintenance margin and liquidation.