Truth Tracking vs. Trackrecord.info vs. Predictions.cc
Several websites now make public predictions easier to record and evaluate. But they do not all solve the same problem.
Some platforms audit predictions made by public figures. Some let people record their own forecasts. Others create shared events where a group can make predictions and later compare results.
This article compares three platforms that are especially relevant to public prediction records:
- Truth Tracking, which extracts and evaluates predictions from public media;
- Trackrecord.info, which builds permanent records of public forecasts;
- Predictions.cc, which lets users create timestamped predictions and prediction events.
The comparison focuses on product design, evidence, scoring, and use cases. It does not attempt to declare one platform universally “best.” The right choice depends on whether the reader wants to audit public commentators, maintain a personal forecasting record, or organize a prediction event.
Short version
| Platform | Best understood as | Main strength | Main limitation |
|---|---|---|---|
| Truth Tracking | An audit layer for public commentary | Extracts predictions from videos and builds channel-level histories | Extraction and adjudication are labor-intensive; scoring varies by domain |
| Trackrecord.info | A public forecast verification database | Explicit transparency principles, human review, and Brier-based scoring | Public dataset is still an early preview with many small-sample records |
| Predictions.cc | A self-service prediction and event platform | Easy timestamped entry, sharing, profiles, and user participation | Results depend more on user entry and self-verification |
The most important difference: where predictions come from
The platforms begin with different sources of data.
Truth Tracking starts with public content such as YouTube videos and other published material. The system transcribes or uses available subtitles, identifies candidate predictions, and turns qualifying statements into structured records. The relevant source video, quotation, and timestamp are preserved when available.
Trackrecord.info also focuses on public predictions, but its public methodology describes a logging process in which a prediction is recorded with its source, timestamp, and resolution criteria, followed by human review. The site’s stated goal is to create a permanent public record of forecasts and outcomes. Trackrecord.info Trackrecord.info methodology
Predictions.cc is primarily user-driven. A user creates an individual prediction or a public prediction event, and the prediction receives a timestamped public record. Users can later verify their own predictions, while event creators publish results for participants. Predictions.cc Predictions.cc documentation
This produces three different models:
- Retrospective extraction: What did a public commentator say?
- Curated public logging: What public forecasts should enter a permanent database?
- Self-service recording: What do users want to predict and track themselves?
None of these models is automatically superior. They are designed for different kinds of accountability.
Truth Tracking: public-media accountability
Truth Tracking’s central question is:
How has this public commentator performed across their historical predictions?
For stock predictions, the published methodology requires a specific instrument, a direction or target, a time horizon, verifiability against market data, and attribution to the host’s own call. It groups predictions by holding duration and reports outcomes such as hit, partial, miss, and pending. Stock calls can also include simulated P/L and comparison with SPY. Truth Tracking stock methodology
The politics workflow similarly focuses on dated, observable claims from public videos, with source links and timestamps when available. It distinguishes successful, partial, failed, pending, and difficult-to-verify cases. Truth Tracking politics methodology
Strengths
- It can track people who never signed up or agreed to make forecasts on a dedicated platform.
- It preserves the media context in which a claim was made.
- It can build channel-level and commentator-level histories over time.
- It supports domain-specific fields, such as ticker, direction, target, horizon, and benchmark performance.
- It can expose patterns in long-form public commentary that are not visible in isolated forecast entries.
Limitations
- Speech-to-text errors can affect the initial extraction.
- Deciding whether a statement is truly a prediction may require interpretation.
- Model-assisted extraction can produce false positives or miss implicit claims.
- Stock P/L, hit rate, and political verdicts are not one universal scoring system.
- Coverage is curated and grows over time rather than representing every public commentator.
- The current product is more useful for reading existing records than for entering one’s own forecasts.
Truth Tracking’s advantage is therefore not that it is a general-purpose forecasting platform. Its advantage is the attempt to create an audit trail for public predictions that already exist in videos and other media.
Trackrecord.info: formal public forecast records
Trackrecord.info is the closest of the three platforms to a direct conceptual comparison. It states that it systematically compares public predictions with real-world outcomes and publishes accuracy metrics. The public site describes logging, human review, primary-source resolution, and permanent publication of resolved results. Trackrecord.info
Its methodology places particular emphasis on:
- verifiability;
- independence from the forecaster’s identity or reputation;
- fine-grained evaluation;
- transparent resolution rationale;
- no retroactive adjustment after resolution except factual corrections.
The methodology uses mean Brier scores as the canonical backend metric and displays a transformed Brier Index where higher is better. This is a meaningful distinction from a simple hit rate because Brier scoring can evaluate probabilistic confidence rather than only whether a binary outcome was correct. Trackrecord.info methodology
Strengths
- It has a clear permanent-record philosophy.
- It uses explicit resolution criteria and human review.
- It is designed around public evidence and repeatability.
- Brier scoring is better suited than simple accuracy for probability forecasts.
- Its methodology is unusually detailed and version-controlled.
Limitations
- The public site currently describes itself as an early public preview.
- Its displayed leaderboard can contain very small samples; the current page states a minimum of two resolved predictions.
- The visible examples are concentrated in public-figure and sports-style forecasts.
- It does not appear to be primarily designed around extracting predictions from long YouTube videos or producing full channel histories.
- A Brier-based record may be harder for casual readers to understand than hit/miss categories.
Trackrecord.info may become a particularly strong comparison point for probabilistic forecasting and formally resolved public predictions. At the current stage, its methodology is more mature than the size and breadth of its public record.
Predictions.cc: personal and social prediction tracking
Predictions.cc takes a more accessible, user-facing approach. Its public documentation describes two main formats:
- individual, timestamped predictions that users later verify;
- public prediction events that other people can join.
User profiles display prediction history and accuracy statistics. Its homepage emphasizes permanent timestamps, shareable events, public profiles, and user-managed verification. Predictions.cc Predictions.cc documentation
Strengths
- It is simple to start using.
- Users can create a permanent timestamp before an outcome is known.
- Prediction events make the product useful for communities and groups.
- Public profiles provide a straightforward personal track record.
- It can cover any subject without waiting for a research team to add a channel.
Limitations
- The user or event creator has a larger role in entering and resolving the record.
- It is less focused on independently extracting what a public figure said in a long video.
- Self-verification can be convenient but may be less independent than an external adjudication process.
- The public documentation allows older events to be hidden under its deletion policy, which readers should consider when interpreting completeness.
- Simple accuracy statistics may not capture probability calibration, forecast difficulty, or financial risk.
Predictions.cc is best suited to people who want to make their own record public or organize a prediction game. It is not primarily a historical audit of media personalities.
Comparison by use case
| Use case | Most natural fit | Why |
|---|---|---|
| Audit a YouTube commentator’s past calls | Truth Tracking | The workflow starts from published media and builds channel records |
| Maintain a personal forecast diary | Predictions.cc | Low-friction timestamped entry and personal profiles |
| Organize a prediction event for a group | Predictions.cc | Shareable events and participant tracking |
| Study formal probability forecasts | Trackrecord.info | Brier-based scoring and explicit resolution criteria |
| Compare stock pundits’ ticker and horizon calls | Truth Tracking | Domain-specific extraction and market-data verification |
| Inspect a public figure’s permanent forecast record | Trackrecord.info or Truth Tracking | Both are oriented toward public-record accountability, with different collection methods |
| Get a live crowd probability | Neither is the primary fit | A forecasting platform or prediction market is better suited |
Where the platforms are not interchangeable
It would be misleading to compare the platforms only by asking which one has the highest accuracy percentage.
Truth Tracking may record a commentator’s statements after they were published. Predictions.cc users intentionally submit forecasts before the outcome. Trackrecord.info may select and formalize public forecasts using explicit resolution criteria. These records have different selection processes and cannot be treated as samples from the same population.
The platforms also use different types of evidence:
- a transcript and video timestamp;
- a manually logged public forecast;
- a user-entered timestamped prediction;
- a human-reviewed outcome;
- a self-reported resolution.
The comparison should therefore focus on transparency and fit for purpose rather than trying to produce one universal winner.
What Truth Tracking should improve
The comparison also reveals areas where Truth Tracking can become stronger:
- Add clearer score definitions across domains. Readers should immediately understand how a stock hit differs from a political or sports result.
- Expose extraction uncertainty. Candidate confidence, transcript quality, and human corrections could be more visible.
- Add probability and calibration support. This would make comparisons with formal forecasting platforms more meaningful.
- Publish more downloadable data. Machine-readable records would improve independent analysis.
- Make corrections and methodology versioning prominent. Historical data should be auditable when rules change.
- Clarify coverage boundaries. Readers should know which channels, dates, and prediction types are included or excluded.
These improvements do not weaken Truth Tracking’s positioning. They make its strongest claim—source-based public accountability—easier to verify.
Conclusion
Truth Tracking, Trackrecord.info, and Predictions.cc all make prediction records more visible, but they begin with different assumptions.
- Truth Tracking is primarily a retrospective audit of public-media commentators.
- Trackrecord.info is a formal public-record and scoring system for public forecasts.
- Predictions.cc is a self-service and social platform for creating timestamped predictions.
Truth Tracking’s clearest niche is not competing with personal prediction tools or live forecasting markets. It is preserving and evaluating what public commentators actually said, across complete channel histories and domain-specific outcomes.
That niche is valuable only if the extraction, scoring, and limitations remain visible. The fairest comparison is therefore not “which site is best?” but “which site provides the right kind of evidence for the question I am asking?”
This article is educational and is not investment, political, or sports-betting advice. Historical forecasting performance does not guarantee future results.