Reported duel: Tengri vs. Rojava

We recieved several reports of duel between tengri and Rojava (aka BeSerok) about possible 'battle phase bug' or 'hack'.
As this duel brought many (justified) questions and pretty high interest of community, I will provide you what happened there in this way, so staff doesn't have to answer individually on each question about it.

It's neither 'battle phase bug' and neither it was any kind of 'hack'. It's game mechanic which occurs pretty rarely, but still, it exist.
Let me add, same thing happened in duel against Witch-Doctor almost 2 years ago, but without any proof, just words, so nothing could be investigated there.

---

What happened?

Tengri (Turkey) attacked Poland which was held already by Rojava (Ukraine). Tengri noticed that in battle phase Rojava lost all units in Warsaw and after he closed battle phase, Rojava still had 1 unit in Poland and Tengri didn't conquer it.

Here are proofs of it:

1. Tengri attacking Poland (Battle phase results by tengri)

2. Outcome after the battle


Also, Tengri provided a video proof, so we can more easily investigate on what actually happened:

---

Explanation:

Thankfully for having a video record of that duel, we found out what actually happened over there.
We watched video and carefully looked on battle phase several times. At the pretty last second of battle phase, you can notice how Tengri lost his last unit as well:


Dave took some time and looked thoroughly into codes to possibly find an answer of what happned.
And he really found the existing code for this pretty rare thing in game mechanics.

---

Now let me explain what it really is (clovis also explained it years ago, but only few people ever heard of it)

We all knew it is possible when both players attack same neutral city, that they both can end with 0 units and neutral city remains uncaptured.
Well, it's found out that it's possible when player attack player as well.

I'll repeat, it's the matter of game mechanics.
There is existing code where both sides fight to death and when it happens, defending side stays alive.
What that means is - if both stacks die, it gives 1 unit back to the defending side, to prevent tie, as that would be contradictory for already captured land.
In cases like this, those kind of battles where both side units are completely dead, it always ends in favor of the defender.

We still don't know exactly in which circumstances this battle outcome occurs, but it's neither visual or any kind of battle phase bug, neither any kind of hack.


Thanks on understanding and thanks to Dave and LukeTan for their time and help on solving this 'issue'.