The three grade tiers
The single highest-confidence pick per market per day — selected by our model as the pick with the strongest and clearest data signal across all inputs for that matchday.
- One Banker per market per day
- Full analyst note on every Banker pick
- Supporting statistics for every selection
- Confidence score typically 70%+
The top 25% of confidence scores for a given market on a given matchday — picks with a strong data signal but not quite the top pick for their market.
- Typically 4–6 Sure Tips per market per day
- Good for accumulator building
- Confidence score typically 60–79%
- Shown on all market pages and the Sure Tips hub
All remaining qualifying picks below the Sure threshold — picks that passed minimum data quality requirements but sit in the lower 75% of the day's confidence distribution.
- Typically 10–15 Free picks per market per day
- Still published with odds and confidence score
- Confidence score typically below 65%
- Volume picks for fixture coverage
What does the confidence score mean?
Every pick carries a confidence percentage — for example, 78% on a Match Winner pick. This figure represents our model's estimated probability that the pick lands, based on the inputs described below. It is not a guaranteed win rate — it means the model assigns a 78% chance of that outcome occurring given the available data.
The confidence score is calibrated against the market's implied probability (derived from the available odds). A confidence of 78% on a selection priced at 1.60 (62.5% implied probability) indicates positive value — the model rates it more likely than the market does. A confidence of 78% on a 1.28 selection (78.1% implied) indicates the market and model are broadly aligned with no additional value edge.
What goes into each pick
Recent form
Each team's results, goals scored and conceded, and expected goals (xG) across their last 5–10 matches at the relevant venue — home and away form are tracked separately.
Head-to-head record
Historical results between the two clubs at this specific ground — not just overall H2H. We look at scoreline frequency, goal counts and result patterns over the most recent 6–10 encounters.
Expected goals (xG)
Each team's xG per game at home and away this season, giving a more accurate picture of attacking and defensive quality than raw goals scored and conceded alone.
Squad availability
Known injuries and suspensions. A key striker or goalkeeper being absent can shift the model's probability meaningfully for goal-related markets.
Venue and context
Home advantage, crowd factor, and fixture context — whether the game has high stakes (title run-in, relegation battle, cup elimination) or is a low-motivation end-of-season fixture.
Market alignment
How the model's probability compares to the market's implied probability (1 ÷ odds). Strong alignment adds confidence; large divergence is flagged as either a value opportunity or a caution signal.
Referee data (cards/corners markets)
For cards and corners markets only — the assigned referee's seasonal booking and corner-award rates. This is the most predictive single input for the cards market.
Schedule and fatigue
Days since last match, upcoming fixture congestion, and whether either team is likely to rotate for a less important fixture — relevant particularly in cup competitions and fixture pile-ups.
What grades do not mean
Banker refers to our highest-confidence grade for that market on that day — not a guarantee. Banker picks still lose. Football results are inherently uncertain, and no prediction system eliminates that uncertainty.
"Sure" is a confidence tier label, not a promise. It means the pick sits in the upper tier of our daily confidence distribution — it does not mean the outcome is assured. We use the term to signal relative confidence within our grading system.
Grades are based on confidence, not odds. A Banker pick can be short-priced (1.30) or longer-priced (2.10) — what makes it a Banker is the strength of the data signal, not the potential return.
FREE picks are lower in the daily confidence distribution — not low quality. On a matchday with strong fixtures, a FREE pick may carry the same absolute confidence score as a SURE pick on a quieter day. The grade is always relative to that day's full distribution.
How to use the grading system
Use Banker picks as confident single bets or as the anchor leg of an accumulator. Each comes with the full analyst note and statistics so you can verify the reasoning before backing it.
Sure Tips are well-suited as acca legs — each has a strong enough confidence basis to justify inclusion in a 4 or 5-fold. Pick legs from different fixtures to reduce correlation risk.
For value bettors, the High Value Tips page identifies picks where the confidence score significantly exceeds the market's implied probability — regardless of grade tier.
Free picks fill out the full picture for a fixture — useful when you want to see what our model says about a specific match across all markets before making your own call.
Value edge — beyond grade alone
The High Value Tips page operates on a different axis from the grade tiers. Value edge is calculated by comparing model probability against the market's implied probability — a pick with a 70% confidence score at odds of 2.50 (40% implied) carries a value edge of +30%. A pick with a 90% confidence score at odds of 1.10 (90.9% implied) carries no edge at all despite being very likely to land.
Value and confidence are complementary, not the same thing. For long-run profitability, consistently backing picks with positive value edge is the most reliable strategy — but high-value picks carry more risk and higher odds than pure confidence picks. Both pages serve different objectives:
- Lower odds, higher probability
- Better for short-term consistency
- Anchor for accumulators
- Best as singles or safe acca legs
- Higher odds, positive edge
- Better for long-run expected return
- Requires larger sample to evaluate
- Higher variance in short term
Our editorial standards
100Tip is operated by Beta Web Analysis Limited. Our prediction model draws on publicly available match data — results, xG, league tables, squad news from official club sources and licensed data providers. No pick is influenced by commercial relationships with bookmakers or operators. All tip grades are assigned algorithmically from model confidence scores — no pick is manually upgraded to a higher grade for marketing purposes.
We publish full analyst notes on every Banker pick so you can evaluate the underlying reasoning. Where our model is uncertain — for example when a key squad news item breaks close to kick-off — we note the caveat directly in the pick. We do not publish picks on fixtures where data quality is insufficient to support a confident recommendation.
Football predictions are inherently uncertain. We publish grades and confidence scores to help you calibrate how much weight to give each pick — not to replace your own judgment. Every pick page includes responsible gambling guidance, and no tip on this site should be interpreted as guaranteed to win.