Pattern Intelligence ยท TAI

2025 quant read

Chien Shing Stainless Steel Co., Ltd. ยท Basic Materials

The pattern is not clean enough for a strong directional call. The chart needs confirmation.

1-week positive match rate 26.3%

Historically negative ยท avg analogue -2.04%

Moderate historical lean
4-week blended positive rate 76.6%

Historically positive ยท avg analogue +4.84%

Stronger historical lean
Completed-week chart evidence Chart evidence is mixed

The current setup has both supportive and challenging components.

Evidence agreement Timeframe conflict

A decisive weekly close that pulls the 1-week and 4-week evidence into agreement.

Current weekly tape

Price, trend line, and fair value

Compare the stored pattern with completed weekly prices. The Fair Value line is the latest reference within this chart, held constant across the display.

Close
13.10
Trend Line
12.22
Fair value
14.32
Hover chart for weekly context.

Signal stack

What changed underneath price?

Each signal is standardized against its own recent history, so the lines show whether the current reading is unusually strong or weak for this stock.

Hover for standardized signal values.

Historical memory

Selected 4-week example returns

Each bar is a stored exampleโ€™s subsequent return, ordered by similarity. This selected set differs from the weighted global, sector and industry blend above.

Hover for analogue return and similarity rank.
Outcome frame

How varied were the historical outcomes?

The lower quartile, median and upper quartile of the selected examples. Actual future outcomes can fall outside this historical range.

Lower quartile +1.88%

What weaker historical analogues tended to deliver.

Median example +3.62%

The middle outcome from the nearest resolved examples.

Upper quartile +8.69%

What stronger historical analogues tended to deliver.

Evidence stack

Why the model sees this setup

These are the current conditions feeding the analogue read. The point is not the number alone; it is whether the evidence agrees with the historical read.

Price vs Trend Line +7.16%

Price is comfortably above its medium-term trend.

Price vs Fair Value -8.52%

Price is close to Fair Value.

Market Dynamics -0.00

Market Dynamics are neutral.

Relative Strength -0.16

The stock is lagging relative strength.

Market Activity +0.10

This component is supportive.

Price Cycle -0.09

This component is a drag.

Historical rhymes

Closest resolved setups

The table keeps the model honest: these are the actual examples behind the selected-example statistics.

Ticker Week Similarity Outcome 4W return
KTC
SET
2026-06-05 0.521 higher +15.57%
MOSHI
SET
2026-06-05 0.494 higher +2.67%
SKY
SET
2026-06-05 0.4817 lower -13.29%
MICRO
SET
2026-06-05 0.48 lower -4.85%
JUBILE
SET
2026-06-05 0.4754 higher +4.17%
SC
SET
2026-06-05 0.4658 higher +6.63%
TQM
SET
2026-06-05 0.4638 higher +8.57%
BBIK
SET
2026-06-05 0.4632 lower -6.37%
JMT
SET
2026-06-05 0.4631 higher +6.31%
QH
SET
2026-06-05 0.463 higher +2.99%
BCP
SET
2026-06-12 0.4607 higher +12.88%
MICRO
SET
2026-06-12 0.4604 higher +16.49%
PLANB
SET
2026-06-05 0.4588 higher +16.35%
KTC
SET
2026-06-12 0.4565 higher +22.31%
SCANA
OSL
2026-06-12 0.453 lower -4.73%
ASK
SET
2026-06-05 0.4495 higher +14.59%
G07
SES
2026-06-12 0.4487 higher +8.73%
XPG
SET
2026-06-05 0.4483 lower -6.25%
MST
SET
2026-06-05 0.4469 higher +7.04%
AF
SET
2026-06-05 0.4456 higher +2.86%
MGT
SET
2026-06-12 0.4453 higher +2.81%
LPN
SET
2026-06-12 0.4449 higher +2.61%
300311
SHZ
2026-06-05 0.4438 higher +3.07%
PIMO
SET
2026-06-05 0.4434 higher +1.87%
605377
SHH
2026-06-05 0.4432 higher +5.13%
FSX
SET
2026-06-12 0.4419 higher +6.36%
ONEE
SET
2026-06-05 0.4404 higher +12.32%
BEC
SET
2026-06-05 0.4401 higher +1.90%
O1KE34
SAO
2026-06-12 0.4397 lower -2.08%
AH
SET
2026-06-05 0.4396 lower -2.10%
Distinctive drivers

What made this setup unusual

These are the largest deviations from the run baseline, renamed into plain market language.

  • Sector structure 2.26431
  • Volume pressure 1.10681
  • Trend Signal 0.63864
  • Next-week expectancy -0.3876
This is historical analogue evidence, not advice. It becomes more useful when historical rate, average return, chart evidence, and nearby examples all point in the same direction.
Sample audit ยท Selected examples

How much evidence stands behind the reading?

30 distinct examples across 28 companies and 2 entry weeks. 2026-06-05โ€“2026-06-12.

Median return3.62%4 weeks after entry
Middle 80% of outcomes-4.99% to 15.65%10thโ€“90th historical percentiles
Largest company share7%Of selected examples
Weight-equivalent sample30.0Accounts for unequal weights, not dependence

The distribution, including losses

< โˆ’10%
1
โˆ’10 to โˆ’5%
2
โˆ’5 to 0%
4
0 to 5%
9
5 to 10%
7
โ‰ฅ 10%
7

0 unchanged outcomes. Bins use equal counts, not similarity weights; positive means strictly above zero.

Does one observation change the story?

Selected-example mean
4.82%
Without the best outcome
4.22%
Without the worst outcome
5.44%
Non-overlapping within each company
3.78% ยท 28 cases

The two exclusion checks show how much the extremes affect the mean. Companies can still share the same market shock; these checks do not prove independence.

Peer comparison

Does the result travel across comparison groups?

The headline 4-week rate blends these stored groups. The selected example chart and one-week calculation use a combined shortlist, so they will not necessarily reproduce that blend.

Comparison groupPositive rateMean returnMatchesBlend weightMedian similarity
Global
76.6%
4.84%30100%0.45

Groups can share examples. Their counts must not be added together as independent observations. Similarity measures resemblance, not a likelihood of success.

Research playbook

What would strengthen or weaken this case?

Continuation

Historical lean meets chart support

A decisive weekly close that pulls the 1-week and 4-week evidence into agreement.

Look for the next completed weekly observation to sustain agreement in trend, relative strength and activity.

Conflict

Similar history, different conditions

If current price and direction oppose the historical lean, keep it as an unconfirmed case. Review the weaker peer group and whether the result depends on extreme examples.

Reassessment

The setup changes

A change in trend, fading relative strength or a materially different pattern can make the old matches less relevant. Recheck the pattern date and the next published reading.

Method and limits

Read the evidence in the right order

1. Compare like with like

Weekly price and indicator conditions are matched with resolved historical setups. Each listing appears once, using its most recent stored reading. Check the date and outcome horizon.

2. Separate frequency from payoff

The headline is a similarity-weighted blend of global, same-country sector and industry results where available. A high positive rate can coexist with large losses. Inspect the median, tails and individual examples.

3. Demand current confirmation

These are descriptive historical matches, not calibrated odds or an independently validated trading edge. Examples may overlap in time, share companies and reflect later data corrections. Costs, liquidity and delisted-company coverage need separate evaluation.

Evidence context