Market forecasts: More inputs must prove they add useful information
By OnTheRice · Sources and limitations · Report a correction
Research reference · source time Not applicable · not a current quote or execution price
Source confidence80% subjective research confidence in the stated pattern; not a calibrated forecast probability.
What’s changing?
A data field can move with prices without improving a forecast after other known information is included. OTR would compare new inputs with a simple method using untouched test data. Costs, repeated information and failures count too. A negative test does not prove the input is useless everywhere.
Key points
- A data field can move with prices without improving a forecast after other known information is included.
- OTR would compare new inputs with a simple method using untouched test data.
- Costs, repeated information and failures count too.
- A negative test does not prove the input is useless everywhere.
What could change this view?
- The tested models, assets and split are limited. Failed significance does not prove no effect exists anywhere. Better models or future data could give different results.
Bottom line
We would keep the simple forecasting baseline unless a new input improves performance on untouched data under a predeclared test. The OTR results support rejecting these tested additions, not claiming that all volume or spillover methods fail.
Its source facts and prices still need independent checks.
Terms in simple English
These explanations help you read this signal. The saved forecast, prices, deadline and outcome rules stay the same.
- Bullish
- OTR expects the price to rise during the stated time window. This is a forecast, not a promise.
- Bearish
- OTR expects the price to fall during the stated time window. This is a forecast, not a promise.
- Watch or on confirmation
- The signal asks readers to watch for a stated condition. If it also says bullish or bearish, that direction still counts in the recorded outcome check.
- Asset or investible
- Something that can be bought or sold, such as a share, bond, fund, currency or cryptocurrency.
- Entry or reference price
- The saved starting price used to check the forecast. A reference quote is not proof that a trade could be made at that price.
- Horizon or deadline
- The forecast time window. A saved deadline states when the outcome should be checked.
- Probability
- An estimate of how likely an outcome is. For example, 60% means about six chances in ten, not certainty.
- Uncalibrated estimate
- A judgement that has not been shown to match real outcomes reliably. It is not a proven success rate.
- Calibration
- Checking whether stated chances match real results across many forecasts. An estimate can still be untested.
- Volatility
- How widely and quickly prices move. Large swings do not show a clear direction.
- Issuer
- The company, government or organisation issuing a security. Evidence about one issuer may not apply to others in its sector.
- Bayesian framework
- Updating an earlier belief with new evidence. A posterior is the updated estimate; a likelihood ratio compares how evidence fits outcomes. A subjective estimate may have no tested calculation.
- De Finetti framework
- Keeping probability judgements consistent. Related calls may share evidence, so one story must not count as several independent reasons.
- Kolmogorov framework
- Basic probability rules: chances cannot be negative, and separate possible outcomes together total 100%.
- AI or machine learning
- Software learning patterns from examples to perform tasks. Its output can be wrong.
- Deployment
- Making software or a system available for real use. Deployment alone does not prove it works well.
- UTC and SGT
- UTC is a world time standard. SGT is Singapore time, eight hours ahead of UTC.
- Baseline
- The starting measurement or comparison used to judge change.
How likely is this call? · Three frameworks
This signal has no investible bullish or bearish call. A market-direction percentage does not apply.
Evidence behind this view
Checks whether separate pieces of evidence support the same view.
OTR’s 17:33 report records that adding relative volume to a price-based volatility model produced no corrected positive improvement across ten tested stocks. Its 19:36 report records that none of ten stock leader-follower tests passed the corrected significance threshold. These are reported study results, not independently rerun models.
What past price moves show
Compares today’s price moves with recorded past patterns.
Not applicable: no asset-return claim. OTR market-pattern results are context for evidence quality, not proof of a technology adoption rate.
Evidence against this view
Shows evidence and risks that could challenge this view.
The tested models, assets and split are limited. Failed significance does not prove no effect exists anywhere. Better models or future data could give different results.
Why the direction changed
Explains recorded changes from expecting a rise to expecting a fall, or the reverse.
Learning review before publication
The supplied direction was retained: No investible directional call.
Supplied synthesis: Technology synthesis
Published synthesis: We see a shift from collecting more market features towards proving that each feature adds value. A variable can be associated with movement without improving a forecast after information already known is included.
Reason: Six actual collaborating reviewers inspected supplied evidence and relevant retained lessons. Preserve the source direction, horizon, original cancelled/no-entry contract, subjective probability distribution and full source narrative. Dated issuer/official evidence supports mechanisms but does not verify an executable entry, a new outcome or empirical calibration. Replace a generic Technology synthesis card teaser with its existing source narrative where needed.
Lessons considered: Use lessons in the next call; Keep the clocks separate; Keep incomplete tests visible; Look for independent evidence; Separate direction from explanation; Compare matching quantities.
This review belongs to the new draft before publication. Earlier published calls and their original deadlines remain unchanged.
What people say vs the facts
Compares recorded public discussion with checked market facts.
We separate announcements, historical code fixes and company-reported controls from independently measured production outcomes. None of these categories alone establishes adoption, sales or a trading advantage.
How past forecasts performed
Shows recorded forecast results and what they can teach us.
No matched outcome cohort for this technology claim was established. OTR’s selected market accuracy is not transferred to this synthesis.
The pre-publication learning review is recorded under “Why the direction changed”, including cases where the supplied direction was retained.
Sources, original wording and price checks
Author-submitted research. The underlying facts and market prices have not been independently verified.
Original research record
OnTheRice’s Emerging Data
Technology synthesis
Category
Technology
Title
New market tests expose the limits of adding more forecasting inputs
Direction
Not applicable: no investment direction.
Horizon
One week for the evidence checks;24 hours for the dashboard check where stated.
Probability
Not applicable: no price forecast.
Research reference
OTR-TECH-EVENING-20261013-10
24-hour change
Source confidence
80% subjective research confidence in the stated pattern; not a calibrated forecast probability.
Forecast prepared
13 October 2026, 19:59:09 SGT
Research update
Within one week, require a predeclared baseline, new test dates and a stated pass threshold before accepting a higher-accuracy claim. A retrospectively chosen best result fails this check.
What’s changing?
We see a shift from collecting more market features towards proving that each feature adds value. A variable can be associated with movement without improving a forecast after information already known is included. The same problem affects AI products that add data, tools or reviewers and then claim higher accuracy. We would compare a simple baseline with the new component on untouched data and count costs, duplicated inputs and failed cases. Negative tests are useful because they stop weak patterns being packaged as a reliable product.
Key points
We connect the independent evidence channels set out under Evidence behind this view. We identify a practical operating need, an opposing explanation and a measurable check.
What could change this view?
The tested models, assets and split are limited. Failed significance does not prove no effect exists anywhere. Better models or future data could give different results.
Bottom line
We would keep the simple forecasting baseline unless a new input improves performance on untouched data under a predeclared test. The OTR results support rejecting these tested additions, not claiming that all volume or spillover methods fail.
How likely is this call? · Three frameworks
We assign 80% subjective confidence to the research pattern, based on the evidence and limits below. This is not a prediction of an asset price.
Favoured scenario
The connected operating need becomes clearer as users demand measured deployment evidence. We would require the result stated under 24-hour check.
Rise scenario
Stronger supporting evidence: a dated, independently checkable result meets the test above.
Decline scenario
Weaker supporting evidence: the result fails, the plan is delayed, or a simpler explanation fits better.
Neutral
No new comparable evidence; the research remains a watch item.
Bayesian · what changed the estimate?
The dated evidence under Evidence behind this view changes our qualitative assessment. We have no fitted posterior or calibrated probability series. The displayed percentage is subjective research confidence.
de Finetti · are the comparisons fair?
We keep source dates, software versions, workloads and methods separate. Shared issuer statements are one evidence channel, and historical code fixes are not new incidents.
Kolmogorov · do the numbers agree?
No rise/decline/neutral price probabilities are assigned. The confidence percentage describes evidence strength, so it is not a distribution to sum to 100%.
Evidence behind this view
OTR’s 17:33 report records that adding relative volume to a price-based volatility model produced no corrected positive improvement across ten tested stocks. Its 19:36 report records that none of ten stock leader-follower tests passed the corrected significance threshold. These are reported study results, not independently rerun models.
What past price moves show
Not applicable: no asset-return claim. OTR market-pattern results are context for evidence quality, not proof of a technology adoption rate.
Evidence against this view
The reasoning is set out under “What could change this view?” in this signal.
Why the direction changed
Not applicable: this is a new Technology synthesis, with no inherited bullish or bearish direction.
What people say vs the facts
We separate announcements, historical code fixes and company-reported controls from independently measured production outcomes. None of these categories alone establishes adoption, sales or a trading advantage.
How past forecasts performed
No matched outcome cohort for this technology claim was established. OTR’s selected market accuracy is not transferred to this synthesis.
Sources, original wording and price checks
OTR Database Data and the dated public sources listed below. We paraphrase factual content, identify our inference and do not provide a trading entry price.
Original research record
New synthesis OTR-TECH-EVENING-20261013-10; evidence retrieved this evening.
Publication time and per-investible results
Prepared 13 October 2026, 19:59:09 SGT; carried forward in this pulse-check report. No investment outcome applies.
Call and evidence
We would apply the mechanism under What’s changing? and test it using the stated check. This is a research synthesis, not a measured production outcome.
Signal evolution
New synthesis of historical statistical findings reported in OTR’s17:33 and19:36 research. This is not evidence of an industry-wide product shift or independent reproduction of the tests.
What changed
The current trigger and separately dated historical baseline are set out under Evidence behind this view. We keep announcement dates separate from deployment and measurement dates.
Why the view changed
The new connection concerns the operating mechanism under OTR reasoning. We have not changed an earlier investment contract or historical outcome.
Evidence, memory & review
We checked contradictions and designed a measurable follow-up. Reading coverage is disclosed in the report; this desk did not read every Drive file or replay every test.
Named assets or actors
OTR research teams, forecasting-tool developers, data users
Directional takeaway
Not applicable: no investment direction is required for Technology.
OTR reasoning
An input can move with the target while adding no information beyond existing features. Comparing models on the same untouched observations isolates that extra contribution and helps prevent a descriptive pattern being sold as predictive accuracy.
Counterpoint
The reasoning is set out under “What could change this view?” in this signal.
Source trail
OTR Database Data;
Review status
Research reviewed for source clocks, counterevidence and testability; production results and statistical calibration remain unverified.
Forecast prepared Publication time unverified
No explicit investible BULLISH or BEARISH call on this card.
What changed
Recorded public historyWe see a shift from collecting more market features towards proving that each feature adds value. A variable can be associated with…
This is the first recorded public view. Further public updates will appear here.
- New market tests expose the limits of adding more forecasting inputs
We see a shift from collecting more market features towards proving that each feature adds value. A variable can be associated with movement without improving a forecast after information already known is included.
Forecast prepared: 13 Oct, 21:35 SGTPublished by OTR: 13 Oct 2026, 22:41 SGT
Why the view changed
This is the first recorded view. Future source-backed updates will appear here.
Evidence, memory & review 0 linked sources
Directional takeaway: N/A
OTR reasoning: We see a shift from collecting more market features towards proving that each feature adds value. A variable can be associated with movement without improving a forecast after information already known is included. The same problem affects AI products that add data, tools or reviewers and then claim higher accuracy. We would compare a simple baseline with the new component on untouched data and count costs, duplicated inputs and failed cases. Negative tests are useful because they stop weak patterns being packaged as a reliable product.
Counterpoint: The tested models, assets and split are limited. Failed significance does not prove no effect exists anywhere. Better models or future data could give different results.
Review status: LEARNING REVIEWED WITH DISCLOSED LIMITS
This evidence panel does not assign a score or probability. Any forecast estimate is shown separately with its own source and limitations.