3. Liga relegation battle 2026/27: who will go down?
Germany · Rest of the season simulated 10,000 times · updated 18 Oct 2026
Lowest survival chances in 3. Liga: TSV Havelse (4%), Grossaspach (30%), Fortuna Koln (47%). 45 points kept a club up in 90% of 10,000 simulations.
enough to stay up in half the simulated seasons
enough to stay up in nine seasons out of ten
Survival chances, club by club
| Club | P | Pts | Proj. pts | Survive | Needs (90% line) |
|---|---|---|---|---|---|
| TSV Havelse | 9 | 5 | 3123–39 | 4% | 40 from 29 (1.38/game) |
| Grossaspach | 10 | 12 | 4032–48 | 30% | 33 from 28 (1.18/game) |
| Fortuna Koln | 10 | 10 | 4234–50 | 47% | 35 from 28 (1.25/game) |
| Meppen | 9 | 10 | 4536–53 | 59% | 35 from 29 (1.21/game) |
| Wurzburger Kickers | 10 | 11 | 4637–55 | 67% | 34 from 28 (1.21/game) |
| Mannheim | 10 | 13 | 4840–57 | 80% | 32 from 28 (1.14/game) |
| Stuttgart II | 10 | 14 | 4940–58 | 82% | 31 from 28 (1.11/game) |
| Preussen Munster | 10 | 12 | 4940–58 | 83% | 33 from 28 (1.18/game) |
| Regensburg | 10 | 10 | 5041–59 | 85% | 35 from 28 (1.25/game) |
| Dusseldorf | 10 | 14 | 5042–59 | 87% | 31 from 28 (1.11/game) |
| Wehen | 10 | 12 | 5243–61 | 91% | 33 from 28 (1.18/game) |
| Hoffenheim II | 9 | 13 | 5243–62 | 92% | 32 from 29 (1.10/game) |
| Ingolstadt | 10 | 14 | 5647–65 | 97% | 31 from 28 (1.11/game) |
| Verl | 9 | 12 | 6051–69 | 99% | 33 from 29 (1.14/game) |
Above 99%: Duisburg, Viktoria Koln, RW Essen, Hansa Rostock, Saarbrucken, Alemannia Aachen.
What does each club need?
Pick a club and test how many points from its remaining games are needed to stay up.
Loading the scenario…
How these chances are made
A Dixon-Coles model rates every club's attack and defence from the league's results, then the 283 remaining scheduled fixtures are played out 10,000 times. A club's chance is the share of those seasons in which it finishes outside the relegation places (top 16). The points line is what the club finishing 17th ended on in each run; ties on points count as a coin flip (goal difference decides them). Point deductions, head-to-head rules and play-offs are not modelled. A statistical projection, not a tip.
The model's match probabilities are scored in public on the calibration record. Free to republish with credit and a link (CC BY 4.0).
