Scope of the comparison
AfricaTipsGuide compares public football prediction websites. We document their features, information formats and limitations. We do not publish our own daily match tips. These reviews are based on public-source research, not personal betting results or an independent prospective trial.
The current comparison was checked on 5 November 2026. A source link identifies where a feature or provider statement could be inspected. Provider claims and our own editorial interpretation are distinguished. Historical football scores are not treated as a record of original forecasts.
The same four tasks for every service
The 7 November 2026 check used the same questions for all nine services:
- Find a match: identify a named fixture from a public list or competition page.
- Inspect the analysis: identify the reasoning, statistical inputs, model output or contributor attribution actually visible.
- Find earlier material: follow a history, results, finished-fixture or dated-page route.
- Inspect results reporting: distinguish final match scores, retained forecast fields, contributor records and itemised betting ledgers.
Each review records a named example, the observation, a primary-source link and any retrieval limit. “Observed” means visible in the retrieved page or browser sample. A search snapshot can document source content but is not a successful live navigation test. Dynamic pages and opinions can change.
Fixtures were chosen as accessible examples of each format; the samples are neither random nor matched across all services. They support feature comparisons, not measured accuracy, full-season coverage or profitable-strategy claims. Account-only features, subscriptions, mobile performance, keyboard accessibility and in-country networks remain outside this check.
New evidence can change an assessment. In this round, SoccerVista’s completed-match view qualified its FAQ claim, and Tips.gg’s finished fixture provided more reporting evidence than navigation alone. Both results-transparency scores were updated to 2/5. HC Media AB remains the responsible publisher; research is AI-assisted and no individual expert audit is claimed. See editorial responsibility and process.
Separate previews, betting tips and their records
Match previews are assessed for the depth and clarity of their analysis, attribution and useful statistical context. Betting tips are specific selections whose results can be tracked. A public preview archive and a settled betting-tip ledger are recorded as different features.
For betting-tip transparency, an itemised log showing the selection, odds, stake, outcome and profit/loss provides meaningful evidence. Visible losing entries and monthly or cumulative reporting strengthen the assessment. An independent audit would add further assurance; the public reporting itself is credited as an observed feature.
For example, the VictoryPicks review assesses its written previews separately from its Betting Tips results ledger. Its betting-tip record is not applied to every prediction mentioned in a preview. Likewise, a competitor’s archived football score is not treated as a complete betting return record.
The six criteria
| Criterion and weight | What we assess |
|---|---|
| Verifiable results 30% | Original predictions and their outcomes, including losses. Look for publication time, market and selection, a defined reporting period, settlement rules and a complete accessible series. Independent reconciliation is stronger evidence than a provider’s summary. This criterion measures auditability, not a claimed win rate. |
| Analysis and methodology 25% | An inspectable explanation of how forecasts are produced. Check attribution, source data, market-specific reasoning, treatment of uncertainty and the limits of the model or editorial approach. A mathematical label or named author alone does not establish accuracy. |
| Free content 15% | What a visitor can actually inspect without payment. Check public predictions, supporting explanations or data, registration barriers, feature restrictions and the visibility of any paid-access boundary. Free content is assessed as access and usefulness, not as a promise of profit. |
| Coverage 10% | Confirmed leagues, fixture types and markets. Identify specific competition pages and market outputs, distinguish regular coverage from isolated examples, check dates and season scope, and record meaningful gaps. Do not substitute an advertised league total for a coverage audit. |
| User experience 10% | Readable information and usable navigation. Inspect the ability to find a match and previous records, understand labels, use a small screen, operate controls with a keyboard and access the content reliably. Measured speed and local mobile-network claims require actual tests. |
| Local relevance 10% | Evidence useful to the particular country. Check domestic competitions, clear timing and language, access from local connections, locally useful context and country-specific feature availability. A country name or flag on its own earns no credit. A country score must identify its country; it cannot be silently reused worldwide. |
Scoring rules and evidence thresholds
Each criterion can receive a score from 0 to 5, in whole points, only after a dated evidence record supports it. The reviewer must record the source, scope, date, observed strengths and gaps. A provider statement can establish what the provider says; a performance or access claim needs the corresponding supporting data or test.
| Score | Evidence standard |
|---|---|
| 0 | Evidence confirms that the criterion is not met in the stated scope. A missing search result is not sufficient evidence for zero. |
| 1 | A limited relevant feature is confirmed, with substantial documented gaps. |
| 2 | Some useful elements are confirmed, but important parts remain absent or restricted. |
| 3 | The core criterion is supported in the defined sample, with material limitations explained. |
| 4 | Strong, consistent evidence supports most elements across a documented sample. |
| 5 | Comprehensive, reproducible evidence supports all elements in the declared scope; important claims have been cross-checked. |
| Not verified | Insufficient evidence to choose a score. This is not converted to zero. |
Overall score out of 100 = Σ (criterion score ÷ 5 × criterion weight). For example, a hypothetical 3 out of 5 on the 30% results criterion contributes 18 points. This is an arithmetic example, not a rating of any service.
A fully assessed overall score requires sufficient evidence for all six criteria. Missing criteria are not removed or reweighted. A global overall score also requires a stated local-relevance scope; country-specific scores must state the country. Provisional totals with explicit assumptions follow the separate policy below. Criterion ratings were assessed on 7 November 2026 using the dated sources in each review. The ratings comparison separates assessed criterion scores from estimates. Comparable small-screen and keyboard assessments, and country-scoped local-relevance checks, remain incomplete. Fully assessed totals remain unavailable, while provisional totals use the explicit assumptions below. These editorial scores assess features and reporting, not prediction accuracy. A missing independent performance audit does not, by itself, prevent credit for a documented public results ledger.
If an outcome is disputed or a source becomes unavailable, the affected claim is marked and the score is withheld or reviewed. An unchanged page is not assigned a new check date just because the calendar has moved on.
Provisional scores, estimates and ranking
From 7 November 2026, the homepage ranks every reviewed service by a provisional editorial score out of 10. The six weights remain 30%, 25%, 15%, 10%, 10% and 10%. Supported criterion scores retain their existing evidence and rationale.
An unverified criterion is assigned 2.5 out of 5 for this provisional calculation only. This is the neutral midpoint of the scale: an explicit modelling assumption, not a claim that the feature was tested or found average. Every service uses the same rule. A confirmed absence receives an assessed zero rather than a midpoint; an apparent conflict between a provider statement and observed behaviour must be investigated before assigning zero. The SoccerVista results assessment was corrected after a completed-match check found forecast fields.
Provisional score /10 = Σ (assessed score or midpoint estimate ÷ 5 × criterion weight) ÷ 10. No criterion is removed or reweighted. Each total shows the percentage of its weight that is estimated. PredictZ has an additional unverified results criterion, so more of its total is estimated. Its neutral results estimate is not evidence of a verified forecast archive.
Local relevance currently uses a neutral estimate for this guide’s combined Nigeria, Ghana, Kenya and Uganda focus. It does not assert verified domestic coverage, in-country availability or a country-specific ranking. Country guides continue to describe their own evidence separately.
Each review also shows an assumption range: the minimum and maximum total obtained by setting all unverified criteria to 0 or 5 while holding assessed scores fixed. It is not a statistical confidence interval. Close positions can change when evidence replaces estimates, so provisional ranks should be read alongside the review.
Services are sorted by the unrounded calculated total. Equal totals share a rank and appear alphabetically; alphabetical order gives no scoring advantage. Shared ownership does not override this calculation. Estimated criteria are replaced when a dated assessment supplies adequate evidence. A fully assessed total is available only after every criterion has sufficient evidence.
What a fair performance comparison would need
To compare accuracy or financial returns, we would need a defined common period, the original pre-match selections and timestamps, a consistent market definition, odds available at publication, all losses and voids, a fixed staking rule and a documented settlement method. Small samples and different types of selections would need explicit treatment.
We have not assembled or independently audited that dataset. We do not turn a website’s estimated probability, confidence indicator, user vote or displayed historical profit into our own verified performance claim.
Evidence labels used in the reviews
- Observed public feature: the linked source showed the feature during the check.
- Provider statement: the service describes its own process or limitation; this does not independently validate it.
- Not verified: the required evidence or test is incomplete.
- Editorial interpretation: our explanation of the practical implications of the observed format.
Reviews link primary pages for concrete observations. We do not say we registered, paid, placed bets or tested local networks when we did not.
Review cadence and corrections
Correction record: 7 November 2026
- SoccerVista: results-transparency score revised from 0/5 to 2/5 after a completed match retained forecast fields. The categorical archive-absence claim was withdrawn.
- Tips.gg: results-transparency score revised from 1/5 to 2/5 after a finished fixture showed retained opinions beside its outcome.
- VictoryPicks: an outdated sentence saying no numerical total was assigned was corrected to match the displayed provisional score. No score increase was made for this copy correction.
Our editorial target is to recheck the main comparison and review sources at least once per quarter, and sooner when a material change or error is identified. This is a review policy, not an automated monitoring claim. Each page shows the most recent actual source-check date.
A correction should identify the exact page, the disputed statement and the primary source supporting the change. On review, the source note and affected assessment should be updated together. New services are included when they add a useful format and enough public evidence exists to describe them accurately.
Ownership and equal treatment
HC Media AB owns both AfricaTipsGuide and VictoryPicks. Shared ownership is shown on the homepage and the VictoryPicks review. All services use the same weights and evidence thresholds. Payment or ownership does not purchase a better assessment.
The homepage opens with a provisional ranking calculated under the policy above. VictoryPicks’ written match previews and separate Betting Tips ledger are assessed through the linked samples. Every service uses the same calculation. The detailed feature table remains alphabetical. Neither list measures verified forecasting accuracy. Read the editorial disclosure, return to the comparison table, or explore the distinct Nigeria and Ghana research guides.