← Suggestions
Match analyses HVC - Alicante United (9)
Hello,
Recent day I played 3 friendslies a day to test how consistent the engine seems to be when using the same tactic against the same opponent: Alicante United, a team with no human manager.
Note: Alicante did not always play with the exact same starting 11 (injuries?), but a sample suggested one or two changes.
When I ran those matches 12 in total, on 27-9 to 30-0, a few things caught my attention:
- HVC won the first 6 matches, but only won 1, drew 1 and lost 4 in the last 6 matches. Maybe the opponent changed tactics?
- The first two matches had 3 red cards, none after (good development)
- Shots on target varies, but does not show any significant deviations.
- Shots on target converted to goals dropped from around 30%-40% to 0-38% in recent matches from my side.
Below the table with data
A = Date
B = Result
C = Goals home (HVC)
D = Goals away (Alicante)
E = Shots home
F = Shots on target home
G = % Shots on target home
H = % Shots on target is goal
I = Shots away
J = Shots on target away
K = % Shots on target away
L = % Shots on target is away
M = % Posession home
N = % Posession away
O = Yellow cards home
P = Red cards home
Q = Yellow cards away
R = Red cards away
A B C D E F G H I J K L M N O P Q R
27-9-2026 Win 2 0 20 7 35% 29% 7 4 57% 0% 50 50 1 1 1 1
27-9-2026 Win 5 0 38 13 34% 38% 3 1 33% 0% 48 52 0 0 3 1
27-9-2026 Win 6 1 34 17 50% 35% 3 3 100% 33% 49 51 3 0 1 0
28-9-2026 Win 2 0 14 5 36% 40% 16 4 25% 0% 49 51 0 0 2 0
28-9-2026 Win 3 1 28 10 36% 30% 11 4 36% 25% 49 51 4 0 2 0
28-9-2026 Win 10 1 42 20 48% 50% 6 4 67% 25% 49 51 1 0 1 0
29-9-2026 Loss 1 7 22 10 45% 10% 14 10 71% 70% 47 53 0 0 2 0
29-9-2026 Win 3 0 16 8 50% 38% 11 4 36% 0% 49 51 1 0 2 0
29-9-2026 Loss 0 3 11 2 18% 0% 25 6 24% 50% 51 49 1 0 0 0
30-9-2026 Draw 1 1 25 10 40% 10% 8 6 75% 17% 49 51 1 0 3 0
30-9-2026 Loss 0 4 8 2 25% 0% 28 8 29% 50% 49 51 1 0 1 0
30-9-2026 Loss 1 3 16 6 38% 17% 21 10 48% 30% 52 48 0 0 2 0
I’m genuinely extremely curious about your findings, but the table doesn’t read well. Would you be able to join our discord and post it there?
Really appreciate your effort!
How do I share it? I'm not really familiar with discord, but created an account.
If you join our discord by clicking on the link in the footer, you can join our channel.
You can simply copy paste it in our suggestions channel over there.
Would be great, your feedback is really valuable.
I'll send some coming days (adding some more data). I am a bit worried about the match engine. Yesterday I played a top tier team (Zwolle) and won 8-0. Today I played Ajax, they were last and lost 2-10? It feels like a lottery now and that takes the pleasure away a bit.
And yesterday I was dominating with a 45/19 shots vs 6/2 and today 6/3 for me vs 34/19 of Ajax. Both no red cards.
It was a very balanced match engine to be honest. But i must admit some changes were made which might have influenced it a little bit.
The coming week I’ll put a lot of my attention into running thousands of analysis to make sure the changes are working as intended.
Hi again and thanks a lot for raising this, and you were right. We dug deep into the match engine after your post, and results really were more random than they should be. Too many goals, too few draws, and the better team didn't win often enough.
We've reworked the engine with one goal: matches should feel like real football. Before releasing anything, we simulated thousands of matches with the new engine and compared the numbers to real top-flight football:
Goals per match: 2.7 (real football: about 2.7)
Draws: 27% (real: about 25%)
0-0 draws: 6% (real: about 7%)
Home win / draw / away win: 43% / 27% / 30% (real: about 45% / 26% / 29%)
Matches with a hat-trick: 5% (real: about 3-4%)
Shots per match: 29 (real: about 25)
Red cards: about 1 every 6 matches (real: about 1 every 7 to 10)
For comparison, the old engine had 3.1 goals per match, only 15% draws, and a hat-trick in roughly one in every six matches.
Better teams now win most of the time, but not always:
- Two evenly matched teams: any result is possible, and about 1 in 4 games ends level.
- A clearly stronger squad wins about two thirds of its matches and still loses about 1 in 7.
- Even a far stronger squad loses now and then. Giant-killings happen, just like in real football, roughly 1 in every 15 to 20 of those matches.
That last point was important to us. A game where the strongest team always wins is boring, but one where squad quality doesn't matter is worse. We think it's now close to how real football behaves: your squad and choices decide most matches, and on any given day an upset is still possible.
A few other things you'll notice:
- Playing at home now gives a small edge.
- Teams react to the scoreline: a side that's comfortably ahead eases off, and a side chasing the game pushes forward.
- Players get more tired during a match, so stamina and fresh legs from the bench really matter.
- Fullbacks are much more involved in the game.
- The match report will tell you when a team had a special day, good or bad.
The full list is in the devlog under v0.18.0.
Thanks again for the feedback. Posts like yours genuinely make the game better, so please keep them coming. If anything still feels off after the update, let us know here.
Thanks for the update!! You guys are working hard and delivering a promosing game! I wasn't able to upload the file yet, so sorry for that. But I'll keep organising friendly matches to see what happens and share the results.
I created an add image button to the forum. So if that helps to post it more clearly, feel free to use it
Log in to reply to this thread.
