ratajmods

Hello! Welcome to my blog. I know, I know.. what is this? 2008?

It's 2026, but who's gonna stop me? Some things are easier to share when we don't require everyone to join a Discord. Blogs are cool and internet is dead. Not that anyone cares, but there is some value in putting something on the internet written by human, especially now when not many things are that way anymore. Let's do some chaotic writing.

I won't make this a habit to write blog posts, but I might treat is as FAQ. If I get a question that I answer "it's complicated" to too often, it can end up here.

And mandatory excuse: English is not my first language. Mistakes will happen.

In this first post, I'll briefly talk about my attempts at comparing performance of Ratchet to other popular methods of modding and scripting on Conan Exiles.

Why am I doing this? Because it would be really nice to have some metrics, to have some number to show, so we can say that Ratchet is 275x faster at math operations than devkit mods.

But the real reason is plain curiosity. Curiosity of my own, and of the people who keep asking me whether it's faster or slower than blueprints.

Although I made Ratchet available to everyone last Christmas (December 25th, 2025), the first public server running it was Avardon. It was a jump in the darkness. We had no way of testing how will Ratchet handle real load of many players until we launched. It could've failed spectacularly. And it did in some ways - first few days, it was crashing quite a bit. Few hotfixes later, we had 227 players in a server and even building was smooth.

That's when I realized that off-loading everything into Ratchet which runs on a separate thread was a really good idea. But to this day we have no measure, no real comparison of this performance.

So that's why I'm doing this, or why I tried to.

My method was going to be very simple, as to make sure I can easily replicate it across all mods that give us some scripting capability.

Loop 10 thousand times over addition of two numbers, measure the number of milliseconds it took to process, compare them in a table, add some nice graphs, and write conclusions.

As it soon turned out, I can't really do that. This comparison was doomed from the beginning and it was an unfair competition. These other mods are designed to be simple and let even people without technical skills create systems for servers. And they do really good job at that. Ratchet does not, Ratchet is for coders.

And of course, other mods are limited by what they can do with UE blueprints, and building another abstraction on top of that is never going to be performant. But let's go through my results.

Pippi

First, I wanted to try Pippi. It's a very old and very popular mod that needs no introduction. It even placed among winners in the recent modding contest hosted by Funcom.

I've never used it before, so I tried installing it to see how well it will perform in my test. Unfortunately, I'm doing this test one week after the 2.2.0 update of the game and Pippi isn't updated yet at the time of writing. I guess we'll have to skip it.

Thrall Wars Utilities

Another old and popular mod for scripting is TWU, and I do remember playing on first season of Halcyon when this mod was the go to solution before we had scripting in TotAdmin.

So I installed TWU on an empty server, messed around with it's bands, as they call them, and managed to get it working. I had to use recursion as there's no for loop, but that would do.

Running the loop 10 times at first for testing worked just fine. Running it 10 thousand times crashed my game. I guess we'll have to skip TWU as well.

To be fair, it was never designed to be used like this.

Tot ! Admin

I love the scripting in TotAdmin, it's really easy and I've made so many scripts with it that I was very excited to build this test case in it as well.

Then I realized that it has intentional throttle to only execute certain number of nodes per second. It was never about speed, it was about not breaking the server by having people spam all sorts of scripts. Fortunately, there are some pre-run options to burst through multiple nodes without stopping, and so I set that up and built my test case, albeit on smaller window of just 200 iterations through the loop.

But I ran into another problem, there was no way to measure execution time. There is a time node, but that gives us in-game time and there's also a date node but it's not very helpful either. There was no precise way to measure it.

And so I built the above, a few nodes to measure it with in-game time. Then I did some math on the day night cycle to get some estimation of the execution time and I was left with 5s, which is a lie because it definitely took under 1s. We're likely in hundreds of milliseconds here.

Apparently, we can't measure this either. Not precisely enough.

More importantly, yet again, scripting in TotAdmin (just like any of the other mods) was never designed to loop through so many nodes.

I actually endorse and very much recommend this mod, it's the best visual scripting option out there, and it's potentially a good way to get into Conan modding as well, as it's not too different from UE blueprints.

UE5 Blueprint

Finally something I know for sure that we can measure.

Blueprints in Unreal Engine are the baseline, this is as much of a raw power and performance as you can get with any mod on Conan, since this is how they are all made, in the official devkit.

I built my test case and ran it a couple of times.

The results were always either 11 ms or 12 ms.

Ratchet

And now to measure it in Ratchet. Here's the Lua code I used:

local start = os.clock()
local num = 0
for i = 1, 1000000 do
    num = num + 1
end
print(tostring(os.clock() - start))

Why is there suddenly a million instead of 10k? Uhh... I tried it with 10k, and the result was 0ms, and that's not good for comparison. I had to make it 100x more difficult to get out of sub millisecond numbers as I can't reliably measure those.

And so, with a million iterations, we ended up with 4 ms.

Theoretically we can now divide it by the 100 again, and assume that 10k would take 0.04ms. And then we'd get to that foreshadowed number of being 275x faster.

So here's the funny table.

Is that even possible? Well, the math says so.

Does it mean that Ratchet is actually 275x faster than devkit mods? No, not really.

What we need to consider is that the test case was just a simple math, and in case of Ratchet it ran entirely on a separate thread from the game's main thread that all other mods share, which is insanely unfair advantage for Ratchet.

It does actually perform that much better in this niche case, but if we interact with the game (such as manipulating inventory items, buffs, RPR data), then we'll have to enter that game thread anyway, and that will be enough of a bottleneck for any real implementation.

Additionally, Ratchet (on default settings) ticks at 50ms to run its scripts, so there is this small overhead that would actually cause it to lose against blueprints. Then it would be 4.5x slower. But if we're taking that into account, then we should also take server FPS into account, which would on any full server cause absolutely any mod to take a second anyway.

Now you know why Avardon runs so smooth. Because where other servers use complex scripts in second layer of visual abstraction that takes hundreds of milliseconds up to entire seconds of processing time, resulting in so called red boxes when used by too many players at once, Ratchet has very minimal overhead. It isn't better for everything, but it is performant.

In conclusion, Ratchet won't make your systems magically faster than devkit mods, but it can do heavy operations on the side, which does help server performance and that's where it shines.

Please note that I was comparing incomparable, but people kept asking me about how Ratchet compares to these other methods too often, so now I have something to redirect them to.