Super League 1 · Greece · Relegation Round - 1

Atromitos–PAS Giannina

3–2

HT 1–0

Kick-off: Saturday, 16 March 2024 at 17:30
Status: Match Finished
Venue: Peristeri Stadium, Peristeri

Goals and red cards

37'AtromitosE. Gonzalez (assist: D. Valencia)
49'PAS GianninaG. Pamlidis (assist: F. Gino)
61'AtromitosG. Tzavellas
85'PAS GianninaRodrigo Erramuspe (assist: M. Tsaousis)
90+5'AtromitosN. Vergos (assist: P. Kunde)

Match statistics

Match statistics
Atromitos		PAS Giannina
8	Shots	17
6	Shots on target	5
5	Shots inside the box	12
3	Goalkeeper saves	3
2	Corners	6
0	Offsides	5
56%	Possession	44%
482	Passes	358
394	Accurate passes

What we expected, and what happened

Expected goals before kick-off: Atromitos 1.57 · PAS Giannina 1.11. On the pitch: 3-2.
Shots Atromitos 8 · PAS Giannina 17; on target 6 and 5.
Atromitos kept the ball more (56%).

Referee

E. Manouchos→

4.48

yellows per match

0.32

reds per match

matches of data

Model prediction

Why this prediction?

The model puts the most likely outcome at 47.5%: “Atromitos to win”. Second comes “a draw” at 27.0%.
Behind that distribution are the expected goals: 1.57 for Atromitos, 1.11 for PAS Giannina.
The highest double chance value is 74.6% (1X — Atromitos win or draw).
These values were computed from 397 past matches.

This is a prediction, not a guarantee; a probability is not a certainty.

These sentences are built by a fixed template; every number in them comes straight from the model's computation, none invented while the text was written.

Atromitos to win47.5%

Draw27.0%

PAS Giannina to win25.4%

When the model has said 45–50% in the past, 2,196 of 4,685 predictions were correct (46.9%). Accuracy record →

12	Fouls	13

1	Yellow cards	2

0	Red cards	0

First half

Actual: 1–0

Atromitos lead at half-time22.8%

Level at half-time42.4%

PAS Giannina lead at half-time34.8%

First-half double chance

1X — Atromitos ahead or level65.2%

12 — not level at half-time57.6%

X2 — level or PAS Giannina ahead77.2%

These values are not a fraction of the full-match prediction; they come from a separate model trained on the first-half scores of 396 matches. Its edge over the base rate is markedly smaller than the full-match model's — the measured figure is on the accuracy record. Read them more cautiously. First-half goal markets are not published because they failed to beat the base rate.

Expected fouls and corners

Fouls

Atromitos: at least 13Actual: 12
PAS Giannina: at least 13Actual: 13

Corners

Atromitos: at least 4Actual: 2
PAS Giannina: at least 4Actual: 6

These figures are set to hold in at least 60% of matches; in the rest the count comes in below them. No upper bound is stated — we make no "at most" claim. Yellow cards were measured and are not published: there our model did not beat the historical average.

What these probabilities rest on

Expected goals comes out of a team's attack weighed against its opponent's defence. A strong attack is damped when it meets a strong defence — which is why comparing the ratings on their own is misleading.

How many goals Atromitos score1.57

Attack: Atromitos score below the league average (24% fewer).

Opponent's defence: PAS Giannina concede 25% more than the league average, which pushes the expectation up.

Result: These two effects give an expectation of 1.57 goals (including home advantage).

How many goals PAS Giannina score1.11

Attack: PAS Giannina score below the league average (25% fewer).

Opponent's defence: Atromitos's defence is close to the league average (1.09×), so it doesn't move the expectation much.

Result: These two effects give an expectation of 1.11 goals (including the away penalty).

Most likely scorelines

1–113.7%2–110.3%1–09.9%

A single scoreline is always less likely than an outcome: “home win” is the sum of 1-0, 2-0, 2-1 and dozens more. So the most likely single score can be a draw while the most likely outcome is a home win; the two do not contradict each other.

Week 1 of the season. This prediction was computed from previous seasons. Summer transfers, budget changes and squad rebuilds aren't in the ratings yet — the model can only see that a team has been rebuilt once new results come in.

We measured this and we're not hiding it: in the opening week of past seasons the model came in 3.3 points below the accuracy it claimed (n=220) — for the rest of the season it beats its own claim. The tendency points this way, though it is not proof on its own: in the opening week it slightly over-calls home wins (said 42.7%, got 39.5%) and slightly under-calls draws (said 26.1%, got 28.2%).

This sample is small and not statistically established, which is why we do not correct the model for it. The correction we tried (pulling probabilities toward the league average) didn't win on holdout, so it wasn't added. Read opening-week predictions with a little more caution.