Monsters & Memories

[GhostCow]

Full findings from the test, as well as the server specs and some other stuff can be found in this pdf: https://files.catbox.moe/qt4zfs.pdf

[GhostCow]

Update 31 includes all the stuff from the pdf I posted earlier and more

https://monstersandmemories.com/updates ... st-results

[GhostCow]

Update 34 is out. I skipped out on posting 32 and 33 because they weren't very interesting. I'll quote my favorite parts below and here's a link to the full post https://monstersandmemories.com/updates ... -july-2023

Character Material System

The Character Material System got a big upgrade this last stint. If you remember from previous updates, we had been clothing our characters by generating our textures then attaching them to the model.

This was done in a layered approach, where each texture layer (picked up from a server generated texture layer definition) would be essentially composited into a final single texture, which is then attached to the model.

In theory, this would have been a great solution because each texture variation would be just one final texture attached to the model (ie. All Night Harbor guards would share one texture for their bodies). In practice, this was problematic for a couple of reasons.

1) Texture layering was achieved via “blitting” a render texture. This means that for every layer, we are sending the layer texture to the GPU then asking the GPU to overlay it over the composite texture. Every step would have to be done discreetly with no chance of doing anything in parallel. In general, we have to repeat this 8-16 times per texture composite. This led to this process taking too long when you are entering an area and must composite 300-500 new textures in a couple of frames.

2) The resulting texture could not be compressed and must be held along with all its mip maps in an uncompressed state on GPU memory per variation. This was expected to be more efficient than having to hold multiple textures per material, but in practice was less efficient than the new method and led to situations where graphics cards that are memory constrained performing a lot worse than expected, even such cards as an RTX 3060 Ti.

Therefore, we pivoted to a new method that was in effect the same type of layered texture approach, but was something that could be handled in real time on the GPU.

We handled this by shifting over to a pretty complex shader that takes all the source textures as is and builds the result on the GPU without the need to keep feeding it textures in discreet steps. We thought this would be better, but we were not prepared for HOW MUCH better it would be.

In terms of loading speed, it could now do all the feeding of textures in ONE STEP and can handle a lot more in parallel (all the compositing would happen using the entire power of the GPU).

In terms of graphics memory utilization, we could now feed a single mip level in a highly compressed format (DHT5) directly into the GPU. We could stream in different mip levels as required. If something is further away, it would use a much smaller texture size. If something was closer, it would stream in higher resolution mips. Overall, it brought graphics memory utilization down by over 80%!

Combat Tuning

While combat formulas were created after a great deal of thought, discussion, and iteration, we were finding some unexpected results, especially when we started testing them over the entire range of levels.

We knew that in order to really get to the bottom of this, we needed a way to very quickly simulate a variety of combat situations and validate our formulas for errors and for intent.

“Instantaneous Combat Logs!”

And so that is what we got busy doing! In order to achieve this, a few foundational changes needed to be made. We had initially had different formulas governing player and NPC combat. In order to simplify simulation (which would require basically copying a player or NPC into a simulant or a group of simulants) and in order to reduce the chances for error, both technical and in design, we had to unify those formulas.

Unifying the formulas meant that we had to change the way NPCs worked and how their stats worked. We had to figure out in player terms, what those stats needed to be. We then converted everything to be on player scale and switched it over to using the same formulas as a player exactly. This immediately introduced a ton of bugs!

We were having NPCs in the higher levels hitting players for BILLIONS of damage. Even though that was hilarious, especially when the target was a gnome, it was something that needed to be fixed. We traced it to an unsigned integer underflow, which wraps a negative number to a very large positive number, which came about due to conversion NPC damage ranges into “player weapon” terms.

In the end, the work was worth it, because we found a lot of compounding issues in our formulas which were causing unintended effects at higher levels. We were able to tune the formulas not only for our entire level range, but to support years of combat scaling in the future.

Plus, it’s pretty cool to pick a player from our database, pick a target, and make them fight against each other thousands of times. It’s like a little fast-forwarding gladiator arena!

Getting the foundation firmly in place, we were able to balance AC budgets, stat budgets, NPC AC and stats, and base hit and miss rates to make combat feel tight.

The next targets on our list were % chance combat events. With a good baseline, a simulation engine, and quite a few parses in our logs, we were able to balance parries, dodges, shield blocks, spell channeling, and weapon procs. They now also feel tight over the range.

Finally, we wanted to round off this stint with some love for pet users. We worked out how weapons should affect NPC damage, which allowed us to let pet users give weapons to their pets. This necessitated building out the underlying persistent NPC inventory system, which we are going to enjoy playing around with in the future.

We also wanted to prevent AFK farming with a pet, which led to working out all the rules of kill credits with regards to pet credit, party member credit, damage shields, and edge cases that come about due to Feign Death and other disengages.

In the course of all this, we had multiple rounds of rebalancing threat generation in the context of all these changes.

[Nevergrind Online]

I just don't like the concept of raids altogether. Herding cats isn't fun, small groups are fun.

Ahem, yes, I agree with this!! Raids were epic, but, all too often you were just a cog in a machine. Overall, I think small groups facilitate a better social experience. That said—I think M&M looks interesting... I wish them the best of luck.

[GhostCow]

Update 35 is out. It's too much to paste here or summarize. Enjoy.

https://monstersandmemories.com/updates ... ember-2023

[GhostCow]

Saw this while catching up on the MnM discord

I'm a big fan of this approach, but I know most people here are going to hate it.

More info:

[Tweed]

I'm a big fan of this approach, but I know most people here are going to hate it.

I HATE HATE HATE this shit.

Might as well just give people the ability to change race or toss in fursuits if you're going to do that shit.

[rusty_shackleford]

Saw this while catching up on the MnM discord

I'm a big fan of this approach, but I know most people here are going to hate it.

If I were a less merciful admin I'd nuke this thread.

[GhostCow]

I'm definitely going to be playing a bard

[Acrux]

I don't understand what that means. Like the 1950s dance, the Twist?

[Tweed]

I don't understand what that means. Like the 1950s dance, the Twist?

Song twisting is when the bard activates one song, waits for it to tick, stops it, and then starts another, and so on and so forth. A good bard can chain together 3 or 4 song effects instead of using one song repeatedly. You internalize how long it takes for a song effect to activate, how long it lasts, and how many you can fit in so you're always hitting keys to start and stop songs.

[GhostCow]

I'm a big fan of this approach, but I know most people here are going to hate it.

I HATE HATE HATE this shit.

Might as well just give people the ability to change race or toss in fursuits if you're going to do that shit.

Saw this while catching up on the MnM discord

I'm a big fan of this approach, but I know most people here are going to hate it.

If I were a less merciful admin I'd nuke this thread.

If you guys would like to go into detail on why you don't like this, I will link the devs to this thread so that they can check it out.

After some thinking, I don't really like it either. I liked it in Age of Wonders 4 for lore reasons, but it doesn't really work for me in any game that doesn't have basically the same setting as AoW 4.

It definitely doesn't make sense for something like night vision to not be tied to race. That's like an evolutionary thing based on DNA, so it doesn't make sense to be a selectable trait unless they come up with a lore reason and whatever lore reason they come up with will be dumb. It's also kind of a weird decision when you take into account that they don't want transmog in the game because they want players to be able to have a good idea of what gear you are using by looking at each other. Their current way of doing race is just transmog for races.

[H-H-Holmes]

No offense but I am surprised it needs any explanation.
It trivialises the whole process of choosing and building a character that you intend to play the role of, making it feel somewhat pointless, which is never a good start in an RPG or MMORPG.

That design choice would be more fitting for an FPS or action game where roleplaying is not a major element and your racial choice is merely cosmetic; a skin.

[Dead]

Aesthetics should be related to function for the game's world to make sense.

[GhostCow]

I do like the idea of some traits being selectable. I don't see why something like fishing skill should be tied to race rather than background, but it's definitely dumb for things like night vision and the iron lung trait that they mentioned. There should be racial traits that are set in stone as well as background traits you can choose from.

[GhostCow]

[maidenhaver]

Aesthetics should be related to function for the game's world to make sense.

I hate races as much as I hate aliens. God made one race: White Man. The rest are other species and shouldn't be playable.

[GhostCow]

My favorite MnM dev being based again. I often think he's hiding a pretty big power level.


Also the October update is out: https://monstersandmemories.com/updates ... tober-2023

I won't try to summarize it because there is a lot of media in the post that goes with the text and you should all check it out.

[GhostCow]

Monsters and Memories is having another test tomorrow at 12pm EST if anyone wants to join. You can make an account and download the game here: https://account.monstersandmemories.com/

[GhostCow]

This info might be useful for some people


Races have different stats you fags.

[Tweed]

Wow, stats, no other MMO ever had stats. We so back MMO bros.

[GhostCow]

Looks like I may have had the time wrong. I think it's 1pm EST.

edit: New patch just went up, so be sure to patch your client in advance

[GhostCow]

Full client download if anyone has issues with the launcher https://files.monstersandmemories.com/m ... 3.11.1.zip

[Tweed]

Nostalgia will carry the final product for about six months or so and then people will start quitting when they relive all the shittiest aspects of classic EQ that they scrubbed from their memories.

[GhostCow]

Nostalgia will carry the final product for about six months or so and then people will start quitting when they relive all the shittiest aspects of classic EQ that they scrubbed from their memories.

Most of the people who buy this will be people who are currently playing P99 and Quarm, so I doubt it.

[Tweed]

Nostalgia will carry the final product for about six months or so and then people will start quitting when they relive all the shittiest aspects of classic EQ that they scrubbed from their memories.

Most of the people who buy this will be people who are currently playing P99 and Quarm, so I doubt it.

Tick tock, tick tock.

Also when you have to pay for the pleasure of being frustrated it's a whole different issue.

[rusty_shackleford]

[Mon Nov 13 19:05:55 2023] You tell HQFrens:1, 'MnM reminds me nothing of what I like about EQ tbh'
[Mon Nov 13 19:06:17 2023] You tell HQFrens:1, 'seems like it's designed for people like GC whose memories of EQ amounted to sitting in a dungeon room f or 8 hours'
[Mon Nov 13 19:06:48 2023] You tell HQFrens:1, 'grinding in one spot for xp was always my least favorite part'
[Mon Nov 13 19:07:13 2023] You tell HQFrens:1, 'liked the wide array of spells, interactive/reactive world, lots of skills, lots of race/class unique ability things'
[Mon Nov 13 19:07:15 2023] You tell HQFrens:1, 'exploring'
[Mon Nov 13 19:08:48 2023] You tell HQFrens:1, 'but when I think of what I like about eq, sitting in one room for 8 hours killing the same mob over and over doesn't come to mind'
[Mon Nov 13 19:09:02 2023] You tell HQFrens:1, 'loved all the various secret things'

[GhostCow]

[Mon Nov 13 19:05:55 2023] You tell HQFrens:1, 'MnM reminds me nothing of what I like about EQ tbh'
[Mon Nov 13 19:06:17 2023] You tell HQFrens:1, 'seems like it's designed for people like GC whose memories of EQ amounted to sitting in a dungeon room f or 8 hours'
[Mon Nov 13 19:06:48 2023] You tell HQFrens:1, 'grinding in one spot for xp was always my least favorite part'
[Mon Nov 13 19:07:13 2023] You tell HQFrens:1, 'liked the wide array of spells, interactive/reactive world, lots of skills, lots of race/class unique ability things'
[Mon Nov 13 19:07:15 2023] You tell HQFrens:1, 'exploring'
[Mon Nov 13 19:08:48 2023] You tell HQFrens:1, 'but when I think of what I like about eq, sitting in one room for 8 hours killing the same mob over and over doesn't come to mind'
[Mon Nov 13 19:09:02 2023] You tell HQFrens:1, 'loved all the various secret things'

Sounds like you'd really like what they are doing with what they call MUD actions.

[GhostCow]

Surprise MnM test in 3 hours. Starts at 1pm EST

[rusty_shackleford]

@GhostCow are you aware of any videos or articles that goes over the connection between Unity & their database they've setup? It looks interesting.