Is it just me, or do battle outcomes appear more random lately?

Like the title asks, I've experienced more random outcomes from battles the past few days. I'm not sure if it started with the leader board/attack pool update or slightly before, but I feel like I notice both a more random attack outcome with more frequent replay errors.

I am aware of the replay bug. I'm just noticing the replay bug seems much worse, and much different. As one example of many, I had a 86% victory against a max base, replay as a 36% loss. But hold on, this all ties together.

The difference was subtle, but the change in outcome was dramatic. During the actual attack, I spent the first 50ish seconds using gunners and juggs to take out peripheral buildings to crate a "path" for my Bombers to follow. By path, I just mean using the AI to make them run straight to the closest building BEHIND the walls so they would pop the walls.

It worked perfectly. I used 17 Bombers to cut through two layers of level 11 walls, followed them in with the gunners (brutes/juggs were ahead of Bombers so they're drawing fire) and the WD. Got the PH, 1m grog and a nice win.

The replay however was a different story, because of one small difference.

The first minute was accurate, periphery buildings went down, path was cleared, etc. Then it came time to watch the bombers get to work, but this time, on the replay, the bombers got excited, hung a sharp right, ran through the flame gate and under the ground pounder and they were dead. Then everyone else died.

Now to the point of all this. The only explanation I can think of for replays not imitating the actual game is because some random number generator is creating a different number in the replays, which creates a different battle outcome. So my theory being the only thing saved from a replay is who fought, loot, rank, where the buildings/walls are and where and when troops were deployed. The math, including the random number generator takes over from there and creates a replay. I'm speculating here, only a dev can confirm this.

Continuing, if there is a random number generator at all in the battles, then that would also explain varried battle outcomes. Now lately it seems the randomness has become more random. I'd consider myself an intermediate skilled player with a solid understanding of what to expect out of the AI. I know where brutes will go, I know where Juggs will go, I know what gunners will target and how they move.

Well, at least I did up until a couple days ago. Now, I have no idea where my gunners will go. They will pass right by an open hole in a wall to walk through a flame gate 10 tiles further away. They will stand right next to the Brutes and Juggs under mortar fire where it used to be easy to keep them at range. Bases with poor layouts I was 100% smashing a few days ago 9/10 times are now a 50/50 crap shoot.

I have really enjoyed this game, the puzzle of figuring out how to use the different troops is really interesting. But I feel like all my knowledge now is useless. I can't make it work. It's just too random.

Anyone else notice this, or am I having a very long string of bad luck?

 

Something has changed. I cannot pinpoint it though. Lately the troops for are not always go in the direction i want. Could be i attack sloppier but dont think thats it. For replays, i hardly watch because they are too far off reality of the attack

 

I felt I'd lost my attack rhythm. I can normally look at a base and know how to go about the takedown. My bombers are now totally in predictable, my juggs love to hug the walls and walk the longest way into the GPs. My WDs love to tag someone further away from the troops they are protecting and walk off. Or worse get stuck somewhere. Expensive.

 

It's not just me then.

 

I agree there has been an upswing of an unpredictable huge variance in battle outcomes. As noted the replay inaccuracy is well known & has been long standing. I've had surprising losses, where I'm baffled that troops didn't 2 star or even 3 star a base. Then to have multiple replays of that same battle all show verification of decimation of that base (again I know replays aren't accurate, but they used to be roughly in the ballpark of how the battle went down). Vice versa I had my base attacked with my win, but replays show a drastically different story where my base is 97-98% decimated. Very, very peculiar Watson... Any weighing in @Lynsey [Midoki] ?

 

Ignoring the fact replays are hopeless for the moment, OP raising an interesting question.

Are battles random?

I'd assumed not and just a straight logic flow. If they are random is this just paths or damage dealt/received as well, how about who defences target?

Back to replays.

Even if it is random, that is no excuse for the replays sucking. "Random" events on computers are entirely reproducible with random number seeds. Computer generated "random" numbers are anything but, they just pull the numbers from a fixed list and you just need to record the seed value (i.e. start point on the list) for the battle to regenerate the outcome, n.b. this is kind of a lot less onerous than remembering time, place and type of each troop deployment and base layout. Even more complicated randomisers are still fixed in reality.

 

Nope, it's not you. Murmurings / complaints / ponderances going on here too: Did something change in AI?

 

So @Lynsey [Midoki] are matches random or fixed based on rules given starting points etc?

 

Can you share exact who likes what and preference ?

 

Thank you very much @Lynsey [Midoki]

 

I've noticed the same thing of late. The behaviour has changed but in a subital way that's hard to describe. For instance, I've recently had bombers run away from walls, this has never happened before. I'm in agreement with the OP, it feels like something's changed although I can't fully explain it.

 

this information is really useful! Thx Lindsey!

 

I just experienced a group of gunners that were already inside a walled structure merrily destroying things and then instead of destroying the next defense in line (which they were right next to) they decided to walk a long distance back outside of the walls. Then they walked all the way around the wall to a gold mine that they literally would've been able to hit from their original spot (albeit they would have had to shoot over the wall but they would have been within range probably without even taking a step or two).

It seemed really weird and out of character. It also wasted quite a bit of time.

 

Maybe someone could put this into the Tips & Tricks thread or something?

 

Thanks for the reply @Lynsey[Midoki]. Path errors explain a lot of what's happeing. I've experienced the same thing Cap'n Sparkle has with gunners making these long walk u-turns to destroy things on the other side of a wall they could have reached by making one or two steps forward inside the walls.

And today, I broke through a wall by killing one bomber at a time, but when I launched my Brutes and Juggs, they didn't go through the hole, they went around! and I cleared the side buildings, the closest things to attack were through the hole, there were no buildings left outside the wall to draw them around like that. it's as if they thought the wall hadn't been broken. And sure enough, watch the replay, and in the replay the wall isn't broken. Could these be somehow related?

 

Just happened again, and this time the replay is accurate. clear ALL building on the ousted of the wall, poke hole in wall, launch attack and everyone, brutes, Juggs, Bombers and gunners ignored the hole, took a longer path measured in tiles to their chosen target, and stood under the ground pounder and died.

The short path through the hole to the target was 8 tiles to the point they aimed at, six tiles to the closest point of the target. The path they chose was 15 tiles and required an additional wall breech. Makes no sense.

 

Again, have seen this too.

 

I made a thread in the battle strategies section with this information. Good thinking!