How the Gubbing & Value Index works
Editor at Better Bet. Ex-bookmaker customer service. Knows what soft accounts look like from the other side of the screen.
The Gubbing & Value Index (GVI): Methodology
What the Gubbing & Value Index measures
The Gubbing & Value Index (GVI) is a 0–100 composite score per UK-licensed sportsbook, refreshed monthly, that estimates how suitable that sportsbook is for the specific activity of matched betting. It blends four components (gubbing risk, welcome offer value, reload offer frequency and value, and operator reliability) into a single letter-graded ranking. The index is rooted in public offer data we already collect, anonymised member reports, and a manual reliability rubric we maintain. It is not a recommendation to gamble, and it is not a "best bookmaker overall" comparison: it is an analysis of one specific use case.
The four components
Each component is scored 0–10 against fixed anchor descriptions. Component scores are then weight-combined into the composite (see "The composite formula" below).
1. Gubbing Risk (weight 30%, multiplier ×3)
What it measures: how long a profitable matched-betting account is likely to remain unrestricted at this bookmaker before being gubbed (limited stake size, stripped of promotions, or closed entirely). A higher score means the bookmaker tolerates promotional play for longer.
How it's scored:
| Score | Description |
|---|---|
| 0 | Bookmaker reliably gubs profitable accounts inside three months of consistent reload uptake. |
| 5 | Mid-tier risk: gubs after roughly 6–12 months for the typical matched bettor. |
| 10 | Effectively never gubs recreational matched bettors; exchange-style stake-laying tolerated. |
Data source: when the bookmaker_gubbing_reports table holds at least 10 anonymised reports for a bookmaker (member-submitted via the website dashboard, scraped from public forum threads we monitor, or recorded as internal observations), the score is computed from the count and severity of those reports inside the last 180 days. Severity is mapped from the report's trigger_pattern value: consistent_profit weights highest (a structural gubbing signal), qualifying_loss_spike lower (a softer warning), multi_account_flag zero (a KYC/identity issue, not gubbing). Below the 10-report threshold the engine uses a manual fallback score from seo_system/data/bookmaker_seeds.tsv (manual_gubbing_risk_score column), which is informed by our own internal observation and the operator's known historical pattern.
This is the component competitors cannot easily replicate without a member base.
2. Welcome Offer Value (weight 20%, multiplier ×2)
What it measures: the estimated cash value of the bookmaker's current welcome promotion, after accounting for typical exchange lay commission and qualifying-bet loss. Captures whether the bookmaker is currently competing for new sign-ups and how strongly.
How it's scored:
| Score | Estimated EV (£) | Description |
|---|---|---|
| 0 | No active welcome offer | Or negative-EV after qualifying loss. |
| 5 | £10–£20 | Standard "small free bet" tier. |
| 10 | £30+ | Top of the active UK welcome-offer market. |
Data source: most recent active row in signup_offer_data per bookmaker. EV is computed from the offer mechanics with a v1 hardcoded extraction rate of 0.78 (typical SNR free-bet conversion at 4–6% lay commission). Refined per-offer EV from the existing oddsmatcher pipeline is on the v2 roadmap.
3. Reload Frequency & Value (weight 30%, multiplier ×3)
What it measures: how much promotional volume the bookmaker keeps pushing to existing accounts, summed across the last 90 days. Captures the bookmaker's appetite for keeping recreational players engaged with profit-positive promotions.
How it's scored:
| Score | Total reload value (£, 90 days) | Description |
|---|---|---|
| 0 | £0–£10 | Effectively no reload activity. |
| 5 | £50–£150 | Steady mid-tier output (~one offer a week, modest EV per offer). |
| 10 | £300+ | Aggressive reload schedule: multiple offers per week, high EV per offer. |
Data source: reload_offer_data rows for the bookmaker in the last 90 days. Total reload value is computed as count × avg_ev_per_reload. Average EV per reload uses the existing oddsmatcher EV estimate where available; otherwise a v1 fallback constant of £4 (a conservative bet-builder reload baseline). On the v2 roadmap: per-offer EV pulled directly from the oddsmatcher pipeline at compute time.
4. Reliability (weight 20%, multiplier ×2)
What it measures: operational hygiene (payment speed, dispute resolution, T&C change frequency, KYC pattern, and whether the operator has been the subject of recent UKGC enforcement action). The only entirely manual component.
How it's scored:
| Score | Description |
|---|---|
| 0 | Chronic withdrawal problems, lost disputes, surprise T&C changes used to deny payouts. |
| 5 | Mid-tier: occasional friction at withdrawal time, KYC asks at unhelpful moments. |
| 10 | Pristine record: fast payments, sane KYC cadence, no recent UKGC action, predictable T&Cs. |
Data source: manual_reliability_score column in seo_system/data/bookmaker_seeds.tsv. Reviewed quarterly by the editorial team and updated when a material event occurs (new UKGC settlement, mass-payment-delay reports, etc.). The scoring engine reads this value directly. There is no automated component for v1.
The composite formula
GVI = round(GubbingRisk × 3 + WelcomeOffer × 2 + Reload × 3 + Reliability × 2)
Each component contributes a maximum of 30 (gubbing, reload) or 20 (welcome, reliability), summing to a 0–100 range.
Letter grades:
| Grade | GVI range | Reading |
|---|---|---|
| A+ | 90–100 | Exceptional fit for matched betting on every component. |
| A | 80–89 | Strong fit, one or two components below ceiling. |
| B | 70–79 | Solid mainstream pick, with a couple of meaningful weaknesses. |
| C | 60–69 | Workable but with notable drawbacks, usually sharp gubbing or thin reloads. |
| D | 50–59 | Marginal: only worth using for the welcome offer in most cases. |
| F | <50 | Avoid: either gubs immediately, has no offer activity, or has a structural reliability issue. |
Why these weights. The two highest-weight components (Gubbing Risk and Reload Value at 30% each) are the two factors that actually determine a matched bettor's lifetime return at a bookmaker: how long the account stays unrestricted, and how much offer volume the operator keeps pushing while it does. The Welcome Offer is one-shot (lower weight), and Reliability is a hygiene factor that mostly differentiates between B-grades and C-grades (lower weight). The weighting is intentionally conservative: small differences in the welcome offer don't drag a strong-on-everything-else operator into a lower bracket.
Update cadence
GVI snapshots are generated monthly. On the first of each month, an automated run of seo_system/scripts/compute_gvi.py --all queries the current state of signup_offer_data, reload_offer_data, and bookmaker_gubbing_reports, recomputes the four component scores per bookmaker, writes a new row to bookmaker_gvi_snapshots, and triggers a static rebuild of the bookmaker hub pages. Historical snapshots are retained for trend lines.
Manual fields (the reliability score and the manual gubbing-risk fallback) are reviewed quarterly. A snapshot can also be triggered manually by a content reviewer if a material event happens between scheduled runs (e.g. a major UKGC settlement, a new welcome offer launching mid-month, a mass-payment-delay incident).
The scoring engine includes sanity checks: no bookmaker can score above 9 on more than two components without a manual flag review; reload counts must come from at least four distinct weeks of data within the 90-day window; and a gubbing-risk score with fewer than 10 anonymised reports automatically falls back to the manual seed value rather than computing a noisy data-driven score.
Worked example
Bet365, GVI snapshot for 2026-05-09 (the first snapshot run).
Underlying data on the snapshot date:
- 31 reload offers in the last 90 days.
- One active welcome offer in
signup_offer_data(£23 in free bets). - Fewer than 10 anonymised gubbing reports in our system at v1 launch, so the manual fallback applies.
Component scoring:
| Component | Score | Reason |
|---|---|---|
| Gubbing Risk | 4.0 | Manual fallback. Bet365 has a strong industry reputation for restricting accounts that establish a consistent profit pattern; this score reflects that profile. |
| Welcome Offer | 5.6 | £23 free bet × 0.78 extraction = £17.94 estimated EV. Linearly interpolated between the £10 → 4 and £20 → 6 anchors. |
| Reload Frequency | 5.6 | 31 offers × £4 v1 fallback EV per offer = £124 total reload value over 90 days. Linearly interpolated between the £50 → 4 and £150 → 6 anchors. |
| Reliability | 9.0 | Pristine UKGC record, fast payments, predictable T&Cs, no material enforcement action since brand inception. |
Composite:
GVI = round(4.0 × 3 + 5.6 × 2 + 5.6 × 3 + 9.0 × 2)
= round(12.0 + 11.2 + 16.8 + 18.0)
= 58
Letter grade: D (50–59 band).
Reading: a solid mainstream pick with consistent reloads and pristine reliability, but the v1 reload fallback (£4 per offer) understates the real value of a steady reload cadence at this operator, and the gubbing-risk profile holds the composite down. Both reflect the matched-betting-specific reality that a profitable Bet365 account is unlikely to remain unrestricted for long. The reload component will rise into the 7–8 range once the v2 oddsmatcher per-offer EV pipeline replaces the fallback constant. (The live snapshot value will fluctuate as reload counts and welcome offer state shift; see the live /bookmakers/bet365 page for the current figure.)
Limitations and what GVI is not
The GVI is narrow by design. It scores one specific use case, matched betting, and does so against the bookmakers in our active scraper coverage. It does not capture:
- Product breadth beyond matched-betting-relevant markets: in-play depth, casino, poker, virtual sports, novelty markets.
- Customer service experience for non-matched-betting users.
- Mobile app quality or website UX (though we sometimes flag these in the per-bookmaker reliability notes).
- Recreational-bettor value: the same operator that gubs profitable accounts may be excellent for a casual once-a-week bettor.
- Casino-bonus value for bookmakers with a strong casino product.
The score is an analysis based on observable public data, anonymised member reports, and our manual reliability rubric. It is not a definitive verdict, and the operator's commercial decision to restrict any individual account remains entirely within their published T&Cs. A "low" GVI is not an accusation of bad faith; it is a reading of suitability for a specific kind of player.
The GVI is also explicitly not a recommendation to gamble. UK Gambling Commission–licensed operators must be 18+, must include the standard responsible-gambling messaging, and gambling can be addictive; see GambleAware and BeGambleAware for support. We publish the GVI to help readers who have already decided to engage with matched betting (a lawful, tax-free side activity in the UK) make a more informed bookmaker choice.
Why we publish this
We publish the GVI for the same reason a personal-finance comparison site rates credit cards on APR, fee schedule, and rewards quality. The methodology is documented, the data sources are public or transparently anonymised, and the score is updated on a fixed cadence. Without a published methodology the index would be opinion; with one, it is data journalism, and we believe matched bettors deserve at least the same standard of comparison data that personal-finance readers take for granted.
The index also serves a practical product purpose for us. Better Bet's tutorial and oddsmatcher tools both depend on members understanding which bookmakers are sensible to engage with at which stage of their matched-betting career. The GVI hub pages give that audience a substantive public comparison they can read before signing up, and a place to return to as the offer landscape shifts.
Related docs
- docs/seo_system/scripts/compute_gvi.md: scoring engine implementation.
- docs/seo_system/scripts/seed_bookmakers.md: seed script.
seo_system/data/bookmaker_seeds.tsv: manual rubric inputs (reliability, gubbing-risk fallback).seo_system/_audit/bookmaker_hubs_and_gvi_plan.md: full project plan (TODO 7 + TODO 3).
Last verified
2026-05-09