Our Testing & Prediction Methodology

Our testing and prediction methodology is the least glamorous page on this site, and it's the one that matters most. Every pick published on NomaPlay follows the same process: logged with a timestamp before kickoff, never touched afterwards, and marked win, loss or void once the match finishes. That's it. No exceptions for bad weeks.

How a pick gets made

We look at form, head-to-head record, team news, market odds movement and, where it's relevant, the kind of context a casual bettor might miss. Mid-table FKF Premier League fixtures, for instance, are notoriously low-scoring, and unders land more often on those matches than most bettors expect going in. That's the sort of pattern we build into the reasoning behind a pick, not a hunch we bolt on afterwards to explain a result. Each published tip includes the reasoning that led to it, expandable on the page, so you can see the logic and judge it yourself rather than taking our confidence percentage on faith.

Timestamping and the no-edit rule

This is the core of the whole system. A pick is published with a timestamp before the match kicks off. After that, it cannot be changed. Not the tip, not the odds quoted, not the confidence figure. If we got it wrong, that mistake stays visible exactly as we made it, which is the only way a track record means anything. A site that can quietly edit a losing pick into a winning one after full time isn't running a track record, it's running a marketing page.

Publishing wins and losses equally

Every settled pick appears on the results feed. Wins and losses both. We don't archive bad runs or bury them below the fold, and we don't stop posting results when a market goes cold on us. There have been stretches where correct-score picks ran badly for two or three weeks straight, and those weeks are on the site same as the good ones.

How we calculate win rate

Win rate is settled wins divided by settled picks. Voided matches, postponements, abandonments, are excluded from both sides of that calculation, since a void isn't a prediction outcome, it's the match not happening as expected. A pick that pushes doesn't count as a win and doesn't count as a loss. This is a simple ratio deliberately. No weighting, no massaging by market or by confidence tier.

How we track profit

Profit is tracked at flat 1-unit stakes across every pick, regardless of the confidence percentage attached to it. We do this because staking systems (doubling up after losses, varying stake by confidence) make a track record easy to dress up and hard to compare against anything real. Flat staking is the boring, honest version. One unit in, one unit tracked, whatever the market.

What this doesn't promise

A verified process is not a guarantee. Football outcomes carry genuine uncertainty, and a well-reasoned pick still loses a meaningful share of the time, that's what odds of 1.85 or 2.10 actually mean. What our methodology guarantees is that the number you're looking at on the track record page reflects real, timestamped, unedited history. Nothing more, nothing less. Betting is 18+ only, and it should stay entertainment, not a plan.