Every tip ZoraTips publishes passes through three analytical layers before going live: Poisson distribution modelling, Expected Goals (xG) data and live pre-match intelligence. This guide explains each layer in full, walks through all nine markets we cover, and provides the bankroll and staking rules we recommend for using our predictions intelligently. This is not written for search engines — it is the exact process our analytics team uses every day, and the same process behind our publicly recorded 78% overall hit rate.
Part 1 — Our Analytical Methodology
Poisson Distribution Modelling
The Poisson distribution describes the probability of a given number of events occurring in a fixed time interval. In football, that interval is 90 minutes and the events are goals. It is the most widely used and rigorously tested model in professional football analytics.
For each fixture we calculate two input values: each team's attack strength (their goals scored per game relative to the league average) and defence strength (goals conceded per game relative to the league average), then apply a home advantage adjustment specific to that league and season. These produce an expected goals figure for each team in that specific fixture.
We then run the Poisson formula across every scoreline from 0-0 to 9-9. Each scoreline gets its own probability. We aggregate these to derive the outcome probabilities — home win, draw and away win — shown in the probability bar on every prediction card.
The critical final step is comparison against bookmaker implied probabilities (100 ÷ decimal odds). When our Poisson probability meaningfully exceeds the bookmaker's implied probability, there is positive expected value in backing that outcome. This is what "model edge" means — and why we only tip when the gap is at least 5 percentage points.
Expected Goals (xG) Analysis
Raw results tell you what happened. xG tells you what was likely to happen based on the quality of chances created and conceded. Every Premier League club, Champions League team and top-tier analytics firm uses some form of xG modelling internally.
A goal scored from a penalty carries approximately 0.76 xG. A corner header carries around 0.08. A clean one-on-one through on goal carries roughly 0.45. The xG value for each shot is derived from its location, angle, whether it was headed or struck, and whether it was assisted or self-created. A team generating 2.4 xG in a match created chances that would produce 2.4 goals on average — regardless of what the goalkeeper did that night.
We use xG in two ways. First, as the quality-adjusted input to our Poisson model instead of raw goals — this strips lucky goals and costly goalkeeping errors out of the numbers. Second, to detect form mispricing: a team on a five-game winning run built on xG of 0.7, 0.8, 0.6, 0.9 and 0.8 is living on borrowed time and likely overpriced by the bookmaker.
Live Pre-Match Intelligence
The best statistical model is worthless if it ignores the latest information. Our pre-match check runs within 24 hours of every kick-off across five areas:
Part 2 — The Nine Markets We Cover
Back the home win (1), draw (X) or away win (2). Our strongest market because Poisson modelling handles directional outcomes well when the quality gap is clear. We avoid 1X2 when the winning side's model probability falls below 50% — the draw risk is too high. We never tip draws from this market.
Both sides must score at least one goal. The key input is defensive xG-against away from home. A team allowing 1.4 xG per away game has roughly a 75% chance of conceding — regardless of recent clean sheets. Well-suited to evenly matched fixtures where no clear directional edge exists.
Three or more goals required. Primary trigger: combined season-average xG above 3.0. Best candidates are high-tempo fixtures in the Premier League, Bundesliga and Eredivisie. We also require Over 2.5 to have landed in at least 3 of the last 5 H2H meetings before publishing High confidence.
Two or more goals required. Our highest single-market hit rate. Both teams must post a season-average xG above 0.7 and the fixture must have produced Over 1.5 in at least 4 of the last 5 H2H meetings. Primarily used as a low-variance accumulator anchor — standalone odds of 1.10–1.25 rarely justify a single stake.
Covers two of three outcomes: 1X (home or draw), X2 (draw or away), 12 (home or away). Preferred when the model gives a team a 45–55% win probability — too thin for a straight Match Result tip but strong as a combined 1X or X2. The 1X option is particularly powerful for strong home sides where win + draw probability exceeds 78%.
Stake returned if the match draws. Ideal when the model gives a team a 48–62% win probability and the draw probability exceeds 24%. DNB removes the draw loss at a small odds reduction and provides better risk-adjusted value than the straight win in volatile derby fixtures where draws are historically common.
Predict the score at 45 minutes. A specialist market — only tipped when a team shows a clear first-half pattern across recent fixtures. We use first-half xG specifically (not full-match figures). A team leading at HT in 65%+ of home games facing a side allowing early away goals is the ideal candidate.
Predict both the half-time AND full-time result — nine possible combinations at odds of 3.00–8.00. Lower hit rate by design: positive expected value works at the higher price range. We calculate joint probabilities for each combination and tip only where our probability materially exceeds the odds implied.
A goal handicap applied to one team creates a two-outcome market. Allows us to back dominant teams at better prices than Match Result — a 58% win probability may price at 1.65 in 1X2 but 1.90 at AH -0.5, representing far better expected value. We use three lines: -0.5 (win only), +0.5 (win or draw), -1.5 (win by 2+).
Part 3 — How to Read a Prediction Card
The Probability Bar
Every prediction card shows a three-colour bar beneath the fixture. Left (green) = our Poisson model's home win probability. Centre (gold) = draw probability. Right (red) = away win probability. These are independent of bookmaker odds — they represent what our model gives as the true probability before any bookmaker margin is applied.
The value gap is visible here: if our bar gives the home side 54% and the bookmaker's odds imply 38%, there is a 16-point positive edge. That gap is why the tip exists.
The Five Confidence Dots
Five circles appear beneath the odds. Each corresponds to one analytical signal:
Five filled dots = Very High conviction — full unit stake. Three dots = moderate signal — half unit recommended. Two or fewer = speculative only.
The Analysis Modal
Clicking any card opens the full analysis popup with: home and away xG, BTTS probability, H2H average goals, last five result dots for each team, the probability bar, a written rationale, and the positive and negative factors behind the tip. Everything in the modal was written before kick-off — we never edit after results are known.
Part 4 — Understanding Value in Football Betting
What Value Actually Means
A bet has positive expected value when the probability of it winning is higher than the probability implied by the bookmaker's odds. If your model gives an outcome 55% and the bookmaker implies 45%, you have a 10-point edge. Back it consistently over hundreds of bets and the edge materialises as profit. A tip that loses can still have been a good bet. A tip that wins can still have been a poor one.
The Bookmaker Margin
Every set of bookmaker odds contains a margin — the percentage by which implied probabilities across all outcomes exceed 100%. In a typical 1X2 market with a 5% margin, the three implied probabilities sum to 105%. To profit long-term, you need to find outcomes where the bookmaker has applied that margin unevenly — where a specific outcome is underpriced relative to our model. We only publish tips where the gap is at least 5 percentage points.
Part 5 — Bankroll Management
The Flat Stake Model
Stake the same unit amount on every tip regardless of odds or recent results. Flat staking allows the long-run edge to express itself cleanly, without the variance amplification of variable staking. Strategies like betting more after winners, less after losers, or more on tips you "feel strongly about" introduce subjective bias that consistently underperforms flat staking over meaningful sample sizes.
Defining Your Unit
At 1–2% unit size, a losing run of ten consecutive tips reduces your bankroll by 10–20% — painful but survivable. At 5–10% per bet, the same run is potentially catastrophic.
Staking by Confidence Level
Very High / High (4–5 dots): full unit stake. Moderate (3 dots): half unit. Low (1–2 dots): quarter unit or skip. This is the only subjective staking adjustment we endorse — all other staking systems (martingale, progressive, chasing) are statistically harmful.
Weekly Budget Rule
Set a weekly betting budget before the week begins and treat it as spent the moment you allocate it. Never replenish mid-week after losses. If your budget is exhausted, stop until Monday. A hard weekly budget prevents the most damaging pattern in recreational betting — the mid-week loss chase.
Part 6 — Accumulator Strategy
When Accumulators Add Value
Accumulators multiply the odds of several legs together, producing a higher return at the cost of requiring all to win. Casually constructed accumulators are poor value. But when every leg carries genuine model edge, the accumulated expected value is the product of each individual edge — potentially excellent. The rule: never include a negative expected value leg. A Low confidence tip in an otherwise strong accumulator reduces the expected value of the entire bet.
Optimal Accumulator Size
Three to four legs is the recommended maximum. A five-fold at 1.35 average odds per leg returns 4.44x — meaningful at manageable variance. A seven-fold returns 6.05x but requires seven consecutive tips to land. The additional variance outweighs the marginal return increase beyond four legs for most bettors.
Which Markets to Use as Legs
Best accumulator markets: Over 1.5 Goals (88%), Double Chance (82%) and Match Result for clear favourites (81%). These have the highest base hit rates. Over 2.5 Goals (74%) is strong at the right price.
Avoid Correct Score (28%), First Goal Scorer (22%) and HT/FT (52%) legs in regular accumulators. Their low individual hit rates compound the loss risk dramatically.
Accumulator Staking
Stake 2–3% of your weekly budget on any individual accumulator — not 2–3% of total bankroll. A sensible weekly split: 70% of budget on flat-stake singles, 30% across one or two pre-built accumulators from our accumulator page.
Part 7 — Responsible Gambling
The Limits of Any Model
A Poisson xG model is the most rigorous freely available tool for football probability estimation. It is also imperfect. A single defensive error, a goalkeeper performance, an injury in the warm-up or a red card in the first five minutes can overturn the most statistically sound pre-match case. Our 78% hit rate means roughly one in four tips loses. No model eliminates variance.
Warning Signs
Please seek support if you recognise any of the following: betting with money you cannot afford to lose; chasing losses by increasing stakes; lying to family members about gambling activity; feeling anxious or preoccupied with betting outcomes; or gambling to escape negative emotions.
Support Resources
ZoraTips does not accept advertising from bookmakers and does not earn referral commission from links to gambling operators. Our commercial model is based on VIP subscription — not steering users towards specific bookmakers.