Something fishy is going on with the Echo Web Creation

:warning:Disclaimer:

So, this is a very elaborate problem, which needs some explaining and a lot of context.

I am almost sure, that this is a bug or maybe a limitation of the underlying systems, but I felt like putting it in the feedback section would most likely cause more discussion around this, since I am very curious what other people think about this or if they noticed this as well.

TL:DR will be at the bottom

Over the last couple of weeks, I have done a lot of “reveal as much of the echo web as possible” for some really advanced Vessel Of Memory and Vessel Of Chaos uses.

While uncovering the echo web, I usually start going outwards as fast as possible and when I hit dead ends (Shade Of Orobyss Nodes) I will try to wander “sideways” and try finding deeper branches, since the chance of getting Exalted and Specific Unique/Set Item Nodes does increase, the deeper I am in an Echo Web.

Here are a couple of images to show what I mean

Going outwards to reach “the edge” asap

The first few shades I discovered only were +5, +9 and +10 Corruption

Then after going sideways for a extended period of time it seemed like it does no go deeper anymore, but instead even going backwards (closer to the starting point again)

Also some branches started to becomes very linear (lots of nodes in a row without any branches splitting up.

Orange is 11 nodes in a row without any off-branch

Then I started to get another 7 nodes in a row linear branch, which also leads to a dead end (Shade)

Next branch with 9 linear nodes and also a dead end

And now the absolute cherry on top, which made me start this post.

23 linear nodes, which were even decending in depth and resulted in a dead end, that is just 10 echoes deep into the echo web from the starting point.

This branch was literally wasting 30min playtime for nothing

This issue, once the echo web is too much developed seems to somehow corrupt the entire echo web, because even when starting from the startign poijnt again, even only 3-5 echoes deep the branches already start getting very linear.

Just for reference, here is the initial starting path I took, with a very “spider-webby” pattern.
This is how the entire echo web should look like IMO.

TL:DR After a large portion of the echo web is uncovered already by going “sideways” for an extended period of time, the creation of “branches” seems to be corrupted to only spawn very linear branches anymore.

This issue takes away everything, what makes MoF great: planning and strategizing around Loot vs Difficulty. Evaluating longer branches after revealing several one with a beacon is one of the coolest things, but some of the beacosn you see here, just revealed like 1 or 2 branches, which were linear. So no real “choice” there

Thank you for you time reading this and if anyone has more specific questions or wants certain images or descriptions elaborated even more, just ask.

Cheers,
Heavy

5 Likes

Love seeing things like this. I have actually got some reasons as to why this is happening that I can share.

Aspects of this are working as intended.

  1. There’s a maximum depth beyond which you only really get shade echoes.
  2. When the web is particularly large it creates fewer branches.
  3. When creating a new branch close an existing branch it’ll have fewer side branches because the more echoes there are around a specific spot, the lower the chance it has of containing an echo.

These are in place to keep the web from becoming too unwieldy. You found some unintended consequences of the system.

7 Likes

That is VERY interesting!
In Heavy’s post, do you see anything that does NOT match the original intentions?
And now more difficult, is there a way to prevent forking sometimes, so that the web forks at better places and gives another kind of final result?

First off, thank you Mike for the insight, I highly appreciate this.

That is a very basic behaviour, but in my particular examples also one of the major points, why this can be very annoying i ncertain situations.

When you actually want to find the highest shade possible, getting a lot of dead ends (Shades) very early, can leads to massive de-tours, which cause the echo web to grow fast.
In my example the first 3 dead ends where depths 13, 14 and 15, which is not the worst, but also not the best outcome.

I already had webs where I had onyl dead ends way earlier and some were it was way later)

That one is very apparent if you follow my images. Even on spaces where there is “enough space”, it gets linear very fast.

That makes sense and is fine as well.

I can understand the goal, but except Vessels of Memory and Vessels Of Chaos (which only start appearing regularly, when you have a lot more experience with the system already, you will only uncover the web step, by step and uncover a few ndoes at a time without searching through dozens and dozens of nodes.

Now to a few more questions:
Do you have ways and/or the desire to tune/adjust a few of the causing factors of this?

I just want to throw out some ideas, because I obviously don’t know how the system works and what would be the easiest to implement.

  • Can the “creating of fewer branches” be limited to regions of the echo web? And not the entire echo web? So only if you go further out. (So if you start agai nfrom the startign point ,when you already uncovered half of the echo web, the behaviour of echo creation is the same. as a fresh started one)
  • Could branches, that are already linear for a couple of echoes have a chance to split up again.
  • Could there be a higher tendency for branches to go deeper (further away from the starting point) to avoid situations like on my 2nd last image, with the 23 echoes in a row, that effectively were going shallower
  • Could there be less tendencies to spawn a shade, when there are no alternative paths present?

I do have one ideas, but I don’t know how good or bad that would be.

What if the entire echo web is actually pre-generated and you just “uncover it”. Right now every echo you “uncover” is created on-the-fly.
A pre-generated web, created by a good algorithm probably has less "coliding branches, that block each other.

This currently would probably have one disadvantage for offline play, since people could very likely view/access the game files and say how many Nodes of X or Y are in a particular echo web.

We actually had set up a pre-generation system and I can’t remember why we didn’t use it. We ran into some design conflict that made it not work at the time.

In that case, when you fight a Shade, you get the exact same web. Is something you want? Discover a web that has always exactly the same shape? I don’t know, just asking. It may be interesting though.
Unless you re-pregenerate at that moment?

It would still be random.

Every time you start a new timeline or reset with the Shade mechanic a new web.

Then I really like your idea. Sad Mike said it did not work. Yet.

Very interesting post.

I have noticed similar but never followed it to the level you had - usually just gave up and reset the Echo by doing an Orobyss…

One thing you touched on vaguely, that has not been mentioned here is the issue around Orobyss Corruption +/- levels based on the distance from the starting point… From my experiences (subjective obviously) I have noticed that sometimes you can follow a long linear branch (with no secondary branches) that loops back toward the starting point and the corruption levels dont increase although the path was particularly long and actually FAR from the start…

If seems to me like some sort of line distance from the start island rather than number of actual node links to the start. I.e. its almost as if you draw a straight line from the start to the Orobyss echo to calculate the corruption change vs taking the branch distance…

Not sure if its intended this way, but if it is, it does change how one would navigate the echos to get to the biggest corruption increments…

That’s exactly how it works and how it’s supposed to work.

It’s always the relative depth from the starting point.

1 Like

I understand that in theory… but when you end up running down a rabbit hole path without any branches to the eventual Orobyss, the arbitrary direction that the each node has taken dictates the corruption increase or decrease… i.e. just because it curved back in the direction of the start means the number of nodes you ran to get there is irrelevant and its purely based on the crow flight…

I suppose it simply means that if you end up down a path without branches like that you should stop and go elsewhere… which is a little counter intuitive to the number of nodes away from the start…

But you can always easily see if a path is going “deeper” or “shallower”, just observe the Timelines Stability gained.

The amount of Corruption a Shade gives and how deep the Node is within the echo web is correlating.

If you see the Timelines ability is staying the same or going down, stop following that branch.

In a nutshell…

I would also like to add that this system gives strange behaviour with echoes.

This is my map before I do the Beacon: https://i.imgur.com/kXQ7wGF.jpeg
As you can see, there is only 1 way to go beyond the Beacon.

And then I complete the actual Beacon: https://i.imgur.com/TLFM5Db.jpeg
Suddenly I have 3 choices, for no logical reason as to how the web discovery should work.

1 Like

I just checked and this is expected behaviour.

1 Like

Yeah both Beacons and Vessels sometimes reveal “hidden paths”, which I think is fine, because it makes them worth doing even if you “think” there is no other path.

This topic was automatically closed 90 days after the last reply. New replies are no longer allowed.