Leaning down over the next hour: 57% chance the close is lower

On this name's own 5-minute candles the close 12 candles later was higher 32% of the time (20 of 62 windows). The evidence in progress: Inside bar, then a breakout (recent, down): went its way 3 of 7 times on this name; RSI into overbought / oversold (recent, up): went its way 311 of 665 times across 94 names. Weighed together as 47 observations against the base rate's 20, that gives 43%.

The range: 860.30 to 868.37 — where four in five earlier windows of 12 candles ended

From the last close of 867.50, this name's own 12 candles moves ran from -0.83% (the 20th percentile) to +0.10% (the 80th); median -0.19%, typical size either way 0.30%. One in five windows ended outside that range — the worst was -1.28%, the best +0.71%. The two references are those percentiles on today's close, nothing more.

What the patterns leaning this way typically reached

After earlier hits of the same reading, the move within 12 candles typically reached +0.28% in the hit's favour and -0.33% against it before the window closed. Typically — half the hits did less.

The trend line: falling 0.03% per candle, R² 0.53

Slope -0.03% per candle (down), R² 0.53 (choppy), 13 candles from the window's high, -0.69% from it; the last close is with the line. If that line simply continued for 12 candles it would print -0.31% — arithmetic, not a call; lines this choppy say little about the next candle.

How these outlooks have fared: 53 of 100 leaned the right way (53%)

Of the last 100 outlooks with a lean at this candle size and horizon across every name, 53 saw the close go the way they leaned; 67 of 100 closed inside the range printed. Every outlook issued is counted, none dropped.

Our arithmetic from stored candles — no recommendation, no forecast of any price: a chance counted from this name's own record, with the times it was wrong counted beside it. The decision, and the risk, are yours.

Leaning down over the next two hours: 65% chance the close is lower

On this name's own 15-minute candles the close 8 candles later was higher 12% of the time (2 of 17 windows). The evidence in progress: Marubozu (live, up): went its way 180 of 423 times across 106 names; Volume spike (live, up): went its way 160 of 425 times across 59 names; Trend line: slope, length, strength (live, down): went its way 70 of 121 times across 45 names. Weighed together as 80 observations against the base rate's 20, that gives 35%.

The range: 856.48 to 864.90 — where four in five earlier windows of 8 candles ended

From the last close of 867.50, this name's own 8 candles moves ran from -1.27% (the 20th percentile) to -0.30% (the 80th); median -0.59%, typical size either way 0.59%. One in five windows ended outside that range — the worst was -1.82%, the best +0.13%. The two references are those percentiles on today's close, nothing more.

What the patterns leaning this way typically reached

After earlier hits of the same reading, the move within 8 candles typically reached +0.70% in the hit's favour and -0.74% against it before the window closed. Typically — half the hits did less.

The trend line: falling 0.12% per candle, R² 0.87

Slope -0.12% per candle (down), R² 0.87 (straight), 17 candles from the window's high, -2.28% from it; the last close is off the line. If that line simply continued for 8 candles it would print -0.96% — arithmetic, not a call; lines this straight have tended to continue more often than not.

How these outlooks have fared: 42 of 100 leaned the right way (42%)

Of the last 100 outlooks with a lean at this candle size and horizon across every name, 42 saw the close go the way they leaned; 55 of 100 closed inside the range printed. Every outlook issued is counted, none dropped.

Our arithmetic from stored candles — no recommendation, no forecast of any price: a chance counted from this name's own record, with the times it was wrong counted beside it. The decision, and the risk, are yours.

Leaning down over the next six trading hours: 65% chance the close is lower

On this name's own hourly candles the close 6 candles later was higher 0% of the time (0 of 1 windows). The evidence in progress: Marubozu (live, up): went its way 29 of 73 times across 36 names; Opening range breakout (recent, down): went its way 99 of 183 times across 52 names; Move of X% within M candles (recent, down): went its way 151 of 279 times across 54 names. Weighed together as 80 observations against the base rate's 20, that gives 35%.

The range: 852.84 to 852.84 — where four in five earlier windows of 6 candles ended

From the last close of 867.50, this name's own 6 candles moves ran from -1.69% (the 20th percentile) to -1.69% (the 80th); median -1.69%, typical size either way 1.69%. One in five windows ended outside that range — the worst was -1.69%, the best -1.69%. The two references are those percentiles on today's close, nothing more.

What the patterns leaning this way typically reached

After earlier hits of the same reading, the move within 6 candles typically reached +1.26% in the hit's favour and -1.27% against it before the window closed. Typically — half the hits did less.

How these outlooks have fared: 56 of 100 leaned the right way (56%)

Of the last 100 outlooks with a lean at this candle size and horizon across every name, 56 saw the close go the way they leaned; 53 of 100 closed inside the range printed. Every outlook issued is counted, none dropped.

Our arithmetic from stored candles — no recommendation, no forecast of any price: a chance counted from this name's own record, with the times it was wrong counted beside it. The decision, and the risk, are yours.

Leaning down over the next ten sessions: 55% chance the close is lower

On this name's own daily candles the close 10 sessions later was higher 45% of the time (112 of 250 windows). The evidence in progress: Move of X% within M candles (live, up): went its way 42 of 89 times on this name; Gap at the open (live, up): went its way 24 of 63 times on this name; Lower highs and lower lows (live, down): went its way 14 of 31 times on this name; Morning star (live, up): went its way 3 of 8 times on this name; Impulse wave (five swings, Elliott-style) (live, down): went its way 2 of 5 times on this name; RSI into overbought / oversold (recent, up): went its way 8 of 20 times on this name; N consecutive falling candles (recent, down): went its way 6 of 10 times on this name. Weighed together as 80 observations against the base rate's 20, that gives 45%.

The range: 826.69 to 896.93 — where four in five earlier windows of 10 sessions ended

From the last close of 872.50, this name's own 10 sessions moves ran from -5.25% (the 20th percentile) to +2.80% (the 80th); median -0.46%, typical size either way 3.12%. One in five windows ended outside that range — the worst was -17.05%, the best +12.29%. The two references are those percentiles on today's close, nothing more.

What the patterns leaning this way typically reached

After earlier hits of the same reading, the move within 10 sessions typically reached +3.66% in the hit's favour and -4.15% against it before the window closed. Typically — half the hits did less.

The trend line: falling 0.07% per candle, R² 0.08

Slope -0.07% per candle (down), R² 0.08 (choppy), 11 candles from the window's high, -0.85% from it; the last close is off the line. If that line simply continued for 10 sessions it would print -0.67% — arithmetic, not a call; lines this choppy say little about the next candle.

How these outlooks have fared: too few judged to say

Only 2 outlooks at this candle size and horizon has a closed window so far; the record is printed once there are 20.

Our arithmetic from stored candles — no recommendation, no forecast of any price: a chance counted from this name's own record, with the times it was wrong counted beside it. The decision, and the risk, are yours.