Pathing Test DE vs AoC

Hey guys (and girls), I made identical Scenarios for DE and AoC in which I test the pathing in several tests

The test starts with simply tasking 4 Villagers to the goal (King on Ice), through a maze where our Scout blocks the way

One manual test is the 4 Monk test. Simply select the 4 monks and tasks them to the opposite Jungle Tree, after all, 4 monks don't have the "stop" symbol, task them back to the other Jungle tree, the results are truly not what you would expect

All other 3 tests are automatic and start once the stone wall is destroyed. We have 2 Patrol tests with a group of Tarkan's and one with ranged Unit + melee units of different speeds

The last test is simply tasking 2 villagers to collect wood from a tree which is only accessable by one tile.

Disclaimer:

Even though most of the tests are done by triggers, the scenarios are not consistent, as pathing is not consistent. You might load the same scenarios twice and get 2 different results.

Results:

First of all, I would kindly ask for any Youtuber or people that are familiar with video editing (anyone) to run the 2 scenarios side by side (if possible) or one test after another (Test 1 DE, Test 1 AoC, Test 2 DE etc.). This would really show the results more perfectly.

Lets move on

First test

The first test is really just a simple go from A to B test with a maze and a scout as obstacle

DE:

2 Villager makes it to the set Location, while the 2 other bug in the scout and stand still. One of the villager also doesn't walk the fastest way to the location but eventually, find its way.

AoC:

All Villager make it to the set Location, The majority of Villager (3) just walk through the scout as if it doesnt exist (In DE only one does that), the last one is equal to the last one in DE, while walking in the wrong direction, in my tests he was able to correct himself sometimes early than in DE, sometimes they went the same. In AoC the Villagers move less smooth and are stuttering at the start to turn from a 2x2 villager grid to a 1x4 (as the path is only 1 tile wide).

Second test:

The monk test is one of my favourites. It checks how the pathfinding reacts when units are tasked to locations that are occupied or cannot be reached. I stumbled upon this test when I tried to test the conversion resistant of 4 different units (scout, light cav, hussar, generic knight) but I digress.

DE:

In DE they stutter, not all monks make to the closest tile, but most of the time they do.

AoC:

In AoC the situation doesn't get better, both versions struggle to compute the right pathfinding. In my tests AoC was able to move all 4 monks more consistently but way slower. Nonetheless its performance is catastrophic and both Versions fail massively

Third test:

The third test consists of a pathing test. A group of Elite Tarkans are patrolling to into a building which has a building next to it laying down a trail of buildings to destroy

DE:

In DE you can see that the game uses the waypoint system to command the units. If you select all Tarkans you can see the flags it sets from building to building (Should you see the flags?) and the Tarkans do a great job, often also stacking by themselves. After all buildings are razed, the Game simply lets them patrol from the starting point to the patrol point, as intended

AoC:

In AoC several weird things happen. First of all the units dont stack and patrol as usual if they dont have enough space to attack the house, but some units appear to bug out and just stand in a line while not moving. After destroying all buildings, the tarkans patrol in an for me not understandable way, by just patrolling in front of the last building razed, while one tarkan stands still at the starting point.

Fourth test:

The fourth test is also a patrol test, this time its a mix of diffrent units (2 Teutonic knights, 2 plumed archers, a regular archer, and a mangudai are tasked to patrol to the King on the Ice. 4 immobile Longswordsman are in the way.

DE:

The Definite Edition does an great job and executes what you would expect. When the Longswordman, who can only be hit by ranged units, showed up, the archers took them down while the TTK's went on patrolling back and forth, after the archers were done killing, they joined the TTK's.

AoC:

Same as DE, however in some instances the TTK's bugged out at the king and stood still, the archers did the same.

Fifth test:

2 Villagers are tasked to cut a tree that is only accessible by one tile. While you might think this is simply, it's by far the weirdest one, on AoC.

DE:

In DE it was apparently coded that villagers are really using the closest way to the tree, following the edge of the nearest tile and standing on the left outer edge of the tree that gets cut, allowing the second villager to chop on the outer right. Both Villagers are working, DE passed.

AoC:

Now, this is one of the weirdest unexpected results I have ever seen. In AoC the villagers try to chop trees in the middle, clogging up the whole tile and not allowing the second villager to share this tree. However, if you task the first villager to move to the outer edge manually, both can chop, but we don't interfere. So, instead of the second villager taking a nearby tree as it cannot access the intial tree, it simply walks 50 tiles and cuts a bamboo on the other side of the map. Wow 11.

​

Conclusion:

Both pathings are not ideal, DE has improved Patrol in my tests. I play a lot, I find DE to have more pathing errors in practical situations (Military often cannot decide what to attack next and go back and forth between 2 buildings instead of choosing one of them).

I also found bugs which never happened to me in AoC, and I have 800 games on Voobly and 100+ games on DE. I already reported 3 of them in the forums (except that bug that Villager stop working sometimes).

Also, Foot Archers have a higher range in DE.

​

In DE the Plumed Archer and the regular Archer can attack the outpost, while in AoC they can't. Mangudai can attack the outpost in both.

Should in the DE version any archer attack the outpost stop them move them to the starting position and put them on stand ground. I would have done this with trigger but you can only put them on passive stands (the Trigger from AoC with Userpatch that can give them any stand ground doesnt work in DE)

​

Download Scenarios:

The DE scenario is the AoC scenario where I exchanged the .scx to an .aoe2scenario.

DE Version

AoC Version

​

If you have any problems comment here