Weekly Signal
A free weekly-chart signal. Below: what it fires on this week, how often it has worked, and how often the same target is met with no signal at all.
Firing on the last completed week
There is no 90% signal. No combination we tested reached 90% on data before 2022. The highest, with at least 300 signals, was 82.87%. We do not publish a 90% signal we could not find.
The request asked for a 90% version. An honest empty list is the result of that search.
| Asset | Signal week | Its own record | Period tested |
|---|
Weekly chart with the signal drawn
How the signal works, what counts as a success, and why there are no patterns
Signal: weekly RSI(14) below 40 AND weekly MACD(12,26,9) histogram rising versus the prior week. After it fires on an asset it cannot fire again on that asset for 13 weeks.
Success: a weekly close at least +3% above the signal week's close within the next 13 weekly closes. Anything else is a failure. A signal whose 13 weeks are not over yet is counted as pending, not as a success.
The target is an easy one: about 76 of every 100 weeks meet it with no signal at all. The signal lifts that to about 80 to 82. That is an edge of roughly 4 to 6 points, not 80 points.
Crypto alone is weaker: 78.6% from 2022 on 215 signals, below the 80% bar, and only 39 crypto signals exist before 2022. Stocks alone: 82.5% from 2022 on 1,204 signals.
What was searched: Indicators only: RSI, Bollinger Bands and MACD conditions on weekly closes. The best combination that used a Bollinger Band reached 82.5% before 2022, against 82.9% for the best without one, so the published version uses RSI and MACD.
What was not: Chart patterns. No head and shoulders, double top or bottom, cup and handle, trend line, support or resistance level or candlestick formation was tested or is detected. The request asked for patterns and indicators; this page delivers the indicator half. A pattern needs a written geometric definition (how deep, how many bars, how level) before it can be tested honestly, and we have not yet written and tested definitions we would publish. A pattern describer with published definitions is proposed at feature 791.
Check it yourself on TradingView
This script is not published in TradingView's library yet, so there is no link that opens a chart with it loaded. Until it is: copy it, open a weekly chart, open the Pine Editor, paste, and choose Add to chart. It marks the weeks the signal fires, with the same 13-week gap between signals.
Open a weekly TradingView chart
//@version=5
indicator("EventTrader weekly signal (80% version)", overlay=true)
// Use on the weekly (1W) timeframe.
r = ta.rsi(close, 14)
[m, s, h] = ta.macd(close, 12, 26, 9)
cond = r < 40 and h > h[1]
var int lastFire = na
fire = cond and (na(lastFire) or bar_index - lastFire >= 13)
if fire
lastFire := bar_index
plotshape(fire, style=shape.triangleup, location=location.belowbar, color=color.aqua, size=size.small)
// Success = any of the next 13 weekly closes is at least 3% above this week's close.
The same data as JSON, no key: /api/v1/weekly-signal?all=true lists every asset we analysed with its own record.
What this does not show
- Survivorship: only symbols listed and liquid today are tested; delisted losers are absent.
- No trading costs, spreads, slippage, borrow or taxes.
- Success = a weekly CLOSE reaching the target inside the window. It is not a trade result: no stop, no exit rule, and the fall on the way there is not measured.
- Signals cluster in the same market sell-offs, so the number of independent observations is far below the signal count.
- 672 combinations were compared on data before 2022; data from 2022 onward was checked once for the chosen one and never searched.
- Stock history is the vendor's split- and dividend-adjusted close as stored on the scan date.
- This is a historical count on weekly closes. It is not advice and it does not say what any asset will do next.