The math
Textbook definitions on closed candles only, and the risk rules and the accounting written out just as precisely. Nothing here predicts prices: it is how ForecastsPro measures the market and limits what can happen to your money.
01Indicators
Six textbook indicators, computed on closed candles only, on two timeframes for every coin. They describe the market; none of them predicts it.
n = 14Wilder smoothing
Average gain against average loss over 14 closed candles. The averages start as a simple mean and are then smoothed Wilder’s way (L̄ the same as Ḡ). A perfectly flat market reads 50; readings above 70 are conventionally called overbought, below 30 oversold.
12 / 26 / 9EMA seeded with an SMA
Two exponential averages of the close; the gap between them is momentum. The signal line smooths that gap, and the histogram shows whether momentum is building or fading. Each EMA starts from the simple average of its first n values.
20 candles2 σ (population)
A 20-candle average with a band two standard deviations either side. %B places the latest close inside the band — 0 at the lower edge, 1 at the upper, above 1 outside it — and bandwidth says how wide, that is how volatile, the market currently is.
n = 14Wilder smoothing
The typical size of one candle, gaps from the previous close c included. It starts as the mean of the first 14 true ranges and is then smoothed. ForecastsPro reports it as a percentage of the price P, so a $60,000 coin and a $0.10 coin compare directly.
n = 14Wilder smoothing
+DI and −DI (DI⁺, DI⁻) say which side has been pushing — DM⁻ mirrors DM⁺ with d and u swapped, and S is Wilder’s running sum. ADX says how strongly a trend is running, whichever its direction. The DIs need 15 candles and ADX needs 28, because its first value averages 14 DX readings.
SMA 50 / 200two timeframes
Each coin gets this label on its trading timeframe and on the next one up (1h with 4h, 4h with 1d, 1d with 1w), so a short-term signal is read against the bigger trend. Before 200 candles exist, price against SMA 50 must agree with EMA 12 against EMA 26. The label is context for the AI, not a gate rule.
02Risk rules
The risk gate’s sizing and kill switch, the protective orders and the accounting, then the three risk profiles, exactly as coded.
budget Bcap share p
The AI proposes a size; the gate clamps it to whichever cap binds first — the per-coin cap or the total budget cap — and names it. Both caps count only what the bot bought, each position at the larger of what it cost and what it is worth, so neither a rally nor a drawdown makes them look roomier than they are. Buys approved earlier in the same cycle count too.
rolling 24 hper venue
Realized profit and loss over the last 24 hours, average-cost basis, protective stop-loss and take-profit fills included. Past the limit, new buys pause until the losses roll out of the window. It blocks buys, never exits: sells and protective orders always go through, because they only reduce risk. A sell with no recorded cost basis counts as zero, never as profit.
per exchange accountnet of fees
Positions and profit and loss are replayed from your journal’s filled orders, one book per venue and exchange account, so paper results never mix with real ones. Quantities are what really filled, net of commission; proceeds are net of the fee. Coins you already held are never part of it.
one-cancels-the-otherafter every buy
A take-profit limit above and a stop below, from your own levels or the risk profile’s (s, τ in the table below); a level the market has already passed is moved just past the current price P. The stop sells at market where the pair supports it, otherwise as a stop-limit 0.5% under its trigger. A stop does not guarantee the exit price.
Percentages are of your budget B. Stop-loss and take-profit are defaults you can change.
| Risk level | Low | Medium | High |
|---|---|---|---|
| Per-coin cap p | 5% | 15% | 30% |
| Daily-loss limit ℓ | 2% | 5% | 10% |
| Min. confidence | 0.70 | 0.60 | 0.50 |
| Stop-loss s | 3% | 5% | 10% |
| Take-profit τ | 6% | 10% | 20% |
Every buy and sell also needs the coin on your watchlist, a confidence at or above the minimum, and no earlier order for that coin still waiting for the exchange to confirm it. These limits reduce risk; they do not remove it, and none of this predicts returns.
Create your account and pay on one page. Then watch these rules at work with play money, on the exchange testnet, before any of your trading capital is at risk.