Paris Saint Germain (0:2)
Confidence67.1%
Base rate18.1%
Lift3.70×
Fair odds1.49
Radar
Scan any mix of football markets and leagues for the strongest selections
Best predictions — top 8
1 / 8
Atlético Nacional (DNB)
Confidence88.4%
Base rate35.2%
Lift2.51×
Fair odds1.13
Over 3 (void on exactly 3)
Confidence62.3%
Base rate16.9%
Lift3.69×
Fair odds1.61
Benfica to win
Confidence79.6%
Base rate31.3%
Lift2.54×
Fair odds1.26
Arsenal (DNB)
Confidence88.4%
Base rate40.2%
Lift2.20×
Fair odds1.13
SCO-Premiership 
Celtic (DNB)
Confidence81.2%
Base rate40.3%
Lift2.02×
Fair odds1.23
Barcelona over 2.5 goals
Confidence68.3%
Base rate30.8%
Lift2.22×
Fair odds1.46
KAZ-First League 
Turan (DNB)
Confidence81.3%
Base rate44.1%
Lift1.85×
Fair odds1.23
New here? How to use the radar
Set how many picks and which leagues — everything else has a
sensible default. Then read the Calibrated column, not Model: it is the
probability after the model has been corrected by its own track record.
- Rank by Signal (default) gives picks that actually say something. Most likely to land returns safer but more obvious selections.
- Lift is the quickest read:
3.0×means three times likelier than that selection normally is in the league. - Fair odds is break-even. Only worth backing at a longer price.
- Keep one pick per match on — two picks from the same game tend to win or lose together.
- Set Look ahead to Today or This week for the day's or week's best, then use Back and Next below the table to move past the top ten.
On this page
10 of 658
Average confidence
72.9%
All correct
3.89%
Combined fair odds
25.73
“All correct” multiplies the picks together as if they were independent. That holds reasonably across different matches, but it would break if two picks came from the same match.
Top 10 · 1–10 of 658 this week, by strongest signal
| Rank | Match | Pick | Chance | Edge over base | Fair odds | Evidence | |
|---|---|---|---|---|---|---|---|
| 1 |
 Cailungo v
 Tre Penne
 SMR-Campionato results-only
|
Tre Penne (1:0) | 72.3% | +58.8 | 1.38 | 4,775 ≠ 78% | |
| 2 |
 Chippa United v
 Mamelodi Sundowns
 RSA-Premiership results-only
|
Mamelodi Sundowns & over 1.5 | 65.9% | +48.8 | 1.52 | 736 ≠ 55% | |
| 3 |
 Paris Saint Germain v
 Slovan Bratislava
 EUR-Champions League results-only
|
Paris Saint Germain (0:2) | 67.1% | +49.0 | 1.49 | 1,718 ≠ 87% | |
| 4 |
 Boyacá Chicó FC v
 Atlético Nacional
 COL-Primera A results-only
|
Atlético Nacional (DNB) | 88.4% | +53.2 | 1.13 | 1,161 ✓ 87% | |
| 5 |
Mamelodi Sundowns v
 Siwelele
RSA-Premiership results-only
|
Over 3 (void on exactly 3) | 62.3% | +45.4 | 1.61 | 8,425 ≠ 55% | |
| 6 |
 SC Cambuur v
 NEC Nijmegen
 NED-Eredivisie results-only
|
NEC Nijmegen over 2.5 goals | 58.9% | +43.7 | 1.70 | 7,978 ≠ 76% | |
| 7 |
 Osijek v
 Dinamo Zagreb
 CRO-HNL results-only
|
Dinamo Zagreb (DNB) | 88.4% | +52.1 | 1.13 | 1,161 ✓ 93% | |
| 8 |
 Elversberg v
 Bayern Munich
 GER-Bundesliga xG
|
Bayern Munich over 1.5 in 2nd half | 67.0% | +46.2 | 1.49 | 488 ≠ 56% | |
| 9 |
 Shkupi v
 Pelister
 MKD-Second League results-only
|
Pelister to win | 79.6% | +49.0 | 1.26 | 605 ✓ 78% | |
| 10 |
 Moreirense v
 Benfica
 POR-Primeira Liga results-only
|
Benfica to win | 79.6% | +48.2 | 1.26 | 605 ≠ 73% |
How these are ranked
- Model is the raw probability from the scoreline distribution.
- Calibrated corrects it by how often the model was actually right, out of sample, when it made claims in that probability band for that market. The correction is shrunk toward the raw number when the evidence is thin, so a handful of past results cannot swing a pick.
- Base rate is how often that selection has simply been true in that league. Edge is calibrated minus base rate — it separates a pick that knows something from one that is merely obvious. “Under 4.5 goals” is 92% likely in every match ever played; that is arithmetic, not a read.
- Fair odds is 1 ÷ calibrated. It is the break-even price. A bookmaker offering longer than this is the only situation where a bet has value — and they rarely do.
Ranking is by calibrated probability, so the top pick is the single most likely outcome, not the most profitable one. Those are different questions, and only the first can be answered without bookmaker prices.