PredictZ, audited against the closing line
Below the closing line In plain words: the best odds on PredictZ's predictions, when we read them, were on average 1.76% below the fair closing price.
As of 27 Oct 2026, PredictZ has 690 settled predictions on matches we price. Of the 639 that were won or lost at a bookmaker price recorded when we read them, 43.5% were right, against 44.4% needed to break even at those odds. Flat-stake return at the recorded odds: -1.4% on 639 priced predictions. Closing line value: -1.76% (95% interval -2.35% to -1.17%, 628 predictions). The interval sits below zero: the best price available when we read them was worse than the fair closing price, so on average these predictions cost more than they were worth.
Measured by market
| Market | CLV (95% interval) | Hit rate | Break-even | ROI | Settled | Read | On our matches |
|---|---|---|---|---|---|---|---|
| Match result (1X2) | -1.76% [-2.35; -1.17] · n=628 | 43.5%n=639 | 44.4% | -1.4% | 690 | 1,168 | 794 |
| Correct score | Right 7.8% of the time; the estimated fair chance of the scores they chose was 8.2% (550 settled with a fair-chance estimate). | 690 | 1,168 | 794 | |||
Hit rate and break-even use the predictions won or lost at a bookmaker price recorded when we read them (n under the rate; the best soft-bookmaker odds, exchanges excluded). CLV is measured against Pinnacle's closing price on the same selection with the margin removed; the 95% interval appears from 30 predictions and the return from 100 priced predictions. Correct score has no closing line: we compare its hit rate with the estimated fair chance of the chosen scores when we read them, from our score model fitted to Pinnacle's 1X2 and over/under 2.5 prices with the margin removed. It is an estimate, not a bookmaker price.
What they claim
“PredictZ is one of the most established and popular tipsters on the web.”
— predictz.com/windrawwin-tips · read on 14 Oct 2026
“Picks do not guarantee winning bets or profits.”
— predictz.com · read on 14 Oct 2026
Quoted word for word from their public pages. We found no measured hit rate, return or closing line value for PredictZ's own predictions on the pages we read; the figures on this page are ours.
How we measure
We read their public prediction pages every day before kick-off; we publish only aggregate figures, never their picks. Each prediction is matched to a fixture we price, settled on the 90-minute result, priced at the best bookmaker odds available when we read it (exchanges excluded) and compared with Pinnacle's closing price on the same selection with the margin removed. A prediction read after kick-off does not count.
Coverage since 14 Oct 2026: 2,336 predictions read on all markets, 1,588 of them on matches we price, of which 1,380 are settled and 204 still pending. Since 15 Oct 2026 we also look up the predictions we cannot match in our data provider's fixture list for all competitions: of the 736 looked up, 428 turned out to be fixtures in competitions we do not cover. Pages we read every day: predictz.com/predictions, predictz.com/predictions/tomorrow.
Full methodology → Are free football prediction sites accurate? We measure them in public → What is closing line value (CLV)? →
Frequently asked
Is PredictZ accurate?
As of 27 Oct 2026, PredictZ has 690 settled predictions on matches we price. Of the 639 that were won or lost at a bookmaker price recorded when we read them, 43.5% were right, against 44.4% needed to break even at those odds. Flat-stake return at the recorded odds: -1.4% on 639 priced predictions. Closing line value: -1.76% (95% interval -2.35% to -1.17%, 628 predictions). The interval sits below zero: the best price available when we read them was worse than the fair closing price, so on average these predictions cost more than they were worth.
How many of PredictZ's predictions did TipsAudit check?
Since 14 Oct 2026 we have read 2,336 public predictions from PredictZ. 1,588 are on matches we price: 1,380 settled, 204 still pending, 2 read after kick-off (these do not count) and 2 void or without a usable result. The other 748 are not, or not yet, matched to a fixture we price. Since 15 Oct 2026 we also look these up in our data provider's fixture list for all competitions: of the 736 looked up, 428 turned out to be fixtures in competitions we do not cover. We read their public prediction pages every day before kick-off and publish only aggregate figures.
What is PredictZ's closing line value?
Across 628 predictions with a measurable close, PredictZ's average closing line value is -1.76% (95% interval -2.35% to -1.17%), measured against Pinnacle's closing price on the same selection with the margin removed. The interval sits below zero: the best price available when we read them was worse than the fair closing price, so on average these predictions cost more than they were worth.
Which market is PredictZ best at?
Only one market has enough predictions with a measurable close to judge, Match result (1X2): closing line value -1.76% (95% interval -2.35% to -1.17%, 628 predictions). The interval sits below zero: the best price available when we read them was worse than the fair closing price, so on average these predictions cost more than they were worth. Correct-score picks have no closing line; in the table above their hit rate is compared with the estimated fair chance of the chosen scores (our score model, fitted to Pinnacle's prices).
Right of reply: if PredictZ disputes a figure on this page, write to us and we will publish the response next to it.
We have no affiliation with PredictZ or with any bookmaker, sell no picks and measure our own predictions the same way.