What happened to the 64 percent
Four days ago we graded six midweek picks and all six won. That pushed the season number from 58 percent to 64 percent in a single afternoon. We wrote at the time that a six pick sample deserved caution rather than celebration. With 29 more picks now settled, the number has come back to 57 percent.
Nothing changed about the method between those two updates. What changed is the size of the sample. At 39 graded picks the base was small enough that one strong week moved it six points, and one ordinary week moved it back. At 68 picks it is steadier, and it will keep getting steadier as the season runs.
This is the ordinary shape of an honest prediction record. A tipster page whose accuracy only ever climbs is a page that is quietly deleting the misses. Ours is on the front board, every pick, win or lose.
The round in full, by league
Twenty nine picks settled across seven competitions. The Bundesliga and the Betway Premiership were both graded for the first time, so those two columns start from zero and have very little history behind them yet.
| League | This round | Season | Accuracy |
|---|---|---|---|
| Serie A | 2W – 1L | 7W – 2L | 78% |
| Champions League | 2W – 1L | 7W – 3L | 70% |
| La Liga | 3W – 2L | 9W – 5L | 64% |
| MLS | 2W – 1L | 7W – 6L | 54% |
| Bundesliga | 3W – 3L | 3W – 3L | 50% |
| Premier League | 1W – 4L | 5W – 6L | 45% |
| Betway Premiership | 1W – 3L | 1W – 3L | 25% |
| UFC | no picks | 0W – 1L | 0% |
The Premier League was the damage. One win from five settled picks, and three of those four losses were matches where we named the right favourite and the game finished level. Our first four Betway Premiership picks returned one win, which is a poor start and worth saying plainly rather than burying.
The confidence bands still separate cleanly
The single most useful thing in this dataset is that the confidence score continues to mean something. High confidence picks win far more often than low confidence ones, and the gap between the bands is wide.
| Band | Season record | Hit rate | Was |
|---|---|---|---|
| Strong, 70% and above | 11W – 2L | 85% | 86% |
| Good, 55 to 69% | 17W – 10L | 63% | 69% |
| Close call, under 55% | 11W – 17L | 39% | 50% |
Read that bottom row before you read anything else on this site. Our close calls lose more often than they win. They are published because a 45 percent call is honest information, not because they are good bets. If you only ever follow the strong picks, you are following the part of the record that has actually held up.
Only the second strong pick we have ever lost
Liverpool at home to Nottingham Forest, 74 percent confidence, finished 2-2. Under our grading rules a draw counts as a loss, because a pick that names a winner has to produce that winner. That is now two strong pick losses in the entire record, the other being Manchester United at 71 percent losing 2-0 at Hull.
Neither was a case of backing the wrong team on paper. That distinction matters for how much you should read into it, which is not very much from a sample of thirteen.
The draw problem has not gone away
Sixteen of our 29 season losses were draws. That means more than half the time we lose, the match finished level rather than the other side winning. We named the wrong winner only thirteen times in 68 picks.
This round alone produced seven drawn losses, including Liverpool against Forest, Bournemouth against Everton, Mainz against Paderborn, Union Berlin against Frankfurt, Toronto against New York City, Richards Bay against Kaizer Chiefs and Polokwane City against Milford. If you want the market that exists specifically for this problem, we wrote about it in what draw no bet actually means.
Three MLS fixtures that never existed
While grading this round we found that three fixtures published on last week’s board were not on the official MLS schedule at all. Seattle against Chicago, DC United against LAFC and Atlanta against Charlotte were never played because they were never scheduled.
Those three picks are voided rather than counted as losses, because grading a match that did not happen would be meaningless in either direction. The rows have been deleted from the board. That takes our total voids to fourteen, and it is the third time a bad fixture feed has done this to us, so the source has been replaced.
We are flagging it here rather than fixing it silently. A prediction record is only worth reading if the mistakes behind it are visible too.
What is on the board now
The front page has been rotated for the week ahead. There are 57 live picks running from 31 August to 9 September, covering Premier League Matchweek 4, La Liga Matchday 4, Serie A Matchday 3, Bundesliga Matchday 2, a full Betway Premiership round and the opening Champions League league phase fixtures on 8 and 9 September.
- Real Betis against Real Madrid on Friday 4 September, 58 percent to Madrid
- Schalke against Bayern Munich on Saturday 5 September, 75 percent and one of only three strong picks on the board
- Inter against Napoli on Saturday 5 September, 57 percent to Inter
- Kaizer Chiefs against Siwelele on Sunday 6 September, 63 percent, our sixth Betway Premiership pick
- Real Madrid against Inter in the Champions League on 8 September, 59 percent
You can see the full set on today’s football betting tips, and the single highest conviction call each day sits on bet of the day. The previous round’s picks were laid out in our Matchweek 3 preview, and the six from six run that started all of this is in last week’s midweek recap.