Any market
Pick a market. Get levels.
See whether the edge is real.
loading…
Every market
The whole system at once
| Market | Price | 24b | ATR% | ADX | RSI | Bias | Signals | P(up) | ±Move | Spread | Flow |
|---|---|---|---|---|---|---|---|---|---|---|---|
| loading | |||||||||||
Click any row to drop into that market — the chart, the levels, the
order flow, the estimate and the full breakdown of what ran all follow
from that one click. The scan refreshes every fifteen minutes on its own.
Long ladder
Short ladder
Patterns on this bar
—
Market structure
Positioning & sentiment
Live signal performance
Did our own signals actually hit?
| Issued | Market | Dir | Entry | Stop now | Stage | Status | P&L | Rule |
|---|---|---|---|---|---|---|---|---|
| loading | ||||||||
These are forward-tested: the signal was recorded first, then the market decided.
No hindsight is possible here, which makes this record more trustworthy than any backtest.
Holdout
Under the hood
What ran on this market, and where it came from
analyse a market to see this
Every indicator here is listed with who built it, when, and what it was
built for — and then with what it did on our own data. That last part is
the one nobody publishes. An indicator can be forty-seven years old, on
every platform on earth, and still rank outside our top twenty-five
features. Both facts are true and both are shown.
Live estimate
If you bought right now
two to four minutes
pick a market above and press Analyse
How far it could move
This half is a volatility forecast, and it is the better half — but
only modestly. Across many markets it explains roughly two thirds of
the variation; within one market over time, closer to a tenth.
Treat the range as a rough envelope, not a promise.
Which way
Read the edge, not the probability. A market that rises 51.7% of the
time will show a 51.7% probability with no skill at all — the only
number that means anything is how far above that base rate we sit.
Were we right when we said it?
| We said | It actually was | Gap | N | Avg move |
|---|---|---|---|---|
| no resolved estimates yet | ||||
Every estimate is stored the moment it is made, and its outcome is
recorded once the horizon passes. If we said 55% and it came out 48%,
this table says so. It fills in over two to three days and is the only
thing here that can prove the probabilities honest.
Method arena
Every method runs. None is switched off.
| Method | Family | Signals | Closed | Hit | Stopped | Win | Avg P&L | Judgeable |
|---|---|---|---|---|---|---|---|---|
| loading | ||||||||
Does anything beat random?
| Method | Trades | Difference | p-value | Verdict |
|---|---|---|---|---|
| loading | ||||
One of the six methods enters at random. Any method that cannot beat it
is not a method — it is a coincidence with a name. Almost no signal
service publishes this comparison.
Leader stability
loading
When methods agree
| Agreement | N | Avg P&L | Up |
|---|---|---|---|
| loading | |||
Agreement between different families of method carries more weight
than any single leaderboard position.
Admin — where signals fail
Failure analysis
Worst signals — study these
| Issued | Market | Dir | P&L | Outcome | Bars | ADX | RSI | ATR% | Patterns |
|---|---|---|---|---|---|---|---|---|---|
| loading | |||||||||
Audits
Is any of this actually clean?
not run yet
| Test | AUC | Expected | Verdict |
|---|---|---|---|
| loading | |||
Shuffling the labels should destroy any signal. If it does not, the
pipeline is leaking the future somewhere and every result above is
worthless. This is the test almost nobody runs on their own work.
Survivorship
| Selection | Markets | Trades | Win | PF | Expectancy |
|---|---|---|---|---|---|
| not run yet | |||||
Picking today's biggest markets means only the survivors get tested.
Point-in-time selection ranks them by volume at the start of the window
instead. The gap between the two is the bias, measured.
A different question: how much will it move?
not run yet
| Method | AUC | R² | Bars |
|---|---|---|---|
| loading | |||
Direction scored 0.545. Volatility is a different question and usually a
far easier one — but a model only earns its place if it beats persistence
and HAR-RV, which are almost free.
Bars & bet sizing
Does the shape of a bar carry information?
not run yet
| Bar type | Bars | AUC long | AUC short | Kurtosis | Jarque-Bera | Avg uniqueness |
|---|---|---|---|---|---|---|
| loading | ||||||
Time bars split the day evenly; the market does not. Dollar bars close
when a fixed amount has traded, so each bar carries roughly equal
information. Lower kurtosis and Jarque-Bera mean returns are closer to
normal — which is what most models assume.
Bet size from confidence
Equal risk on every signal throws away what the model knows. Size scales
with how far the probability sits above a coin flip; below 5% size,
no position is taken at all.
Information test
Could our data have seen it coming?
not run yet
Which features carry the signal
What the rules missed
These are the features that differed most between opportunities the
rules caught and ones they walked past. A large gap means the
information was there and the rule ignored it.
New listings
New coins — screened before they list
These checks tell you what is dangerous — honeypots, open mint,
hidden owners, concentrated supply. They do not tell you what will go up.
That second question is answered by the outcome table below, once enough
time has passed. Until then, no coin here is a recommendation.
| Token | Chain | Grade | Score | Liquidity | Vol 24h | Holders | Top10 | Age | Listed | Blockers | |
|---|---|---|---|---|---|---|---|---|---|---|---|
| loading | |||||||||||
Do our grades actually predict anything?
| Grade | Horizon | N | Avg | Median | Up | Down >50% |
|---|---|---|---|---|---|---|
| not enough time has passed yet | ||||||
Listings detected
| Detected | Exchange | Symbol | We scanned it? | Grade |
|---|---|---|---|---|
| baseline set — new listings appear from here on | ||||
Broker link
MT5 bridge
checking…
| Symbol | Bid | Ask | Spread | Age |
|---|---|---|---|---|
| no ticks | ||||
Rules backtest — the gap, closed
Do the live signal rules actually work?
loading…
All configurations tested
| ADX band | MTF | Trades | Win | PF | Expectancy | Return |
|---|---|---|---|---|---|---|
| loading | ||||||
Per-market breakdown — with real spreads
| Market | Trades | Win | PF | Expectancy | Spread | Cost | Source |
|---|---|---|---|---|---|---|---|
| loading | |||||||
Scoreboard
Markets that passed validation
| Market | TF | Trades | Win | PF | Expectancy | Return |
|---|---|---|---|---|---|---|
| loading | ||||||
A market appears above only if profit factor > 1.15 and expectancy is positive
after fees and slippage, measured out-of-sample. Everything else is listed below
as tested and rejected. An empty table is a real answer, not a bug.
Tested, no edge found
| Market | TF | Trades | Win | PF | Expectancy | Return |
|---|---|---|---|---|---|---|
| loading | ||||||
Validation queue
| # | Market | TF | Status | Requested |
|---|---|---|---|---|
| — | ||||