This is going to be an in-depth and highly detailed explainer/behind the scenes thread for everything related to Project Neltharion's current Mythic system, including the idea behind it, how it came to be, etc.
If for whatever reason Kain's guide isn't enough information for you, you are in the right place.
Let's dive into the backstory first.
On an endgame realm, endgame content has to be the focus point of the entire game. What is endgame content, you might ask? In 4.3.4, that generally refers to Dragon Soul. So you run Dragon Soul for a few months, you get BiS gear, then what? Exactly. We didn't like the idea of the "then what" aspect to the game, so we had to come up with something. The nature of the game dictates that as you get more and more gear, content becomes easier and easier. At the top level on the original Neltharion we could see groups clearing Dragon Soul in less than an hour consistently. The main content was repetitive and boring. If you didn't have multiple characters ready for Dragon Soul, you would have no incentive to play the game for more than 1 hour a week. This was a big contributing factor to the realm's inevitable death (among other things). What have we learned from this is that we need a brand-new idea to revive the endgame with, one that doesn't get stale as you delve deeper and deeper into it, one that keeps the challenge up for every player, one that's rewarding, etc.
This chat was created for the sole purpose of "Making Neltharion Great Again". Where four like-minded individuals came together and brainstormed everything that would eventually become one of, if not the main focus for the longevity of endgame content that would shape our realm. This is the first documented appearance of the word "mythic" in the entire discord of almost 600k messages. Fun fact, #temp is almost 100k out of that with only 4 people in it. Just for comparison, #neltharion-chat barely has 70k (stats as of 09.07.).
Anyway, a lot of dev and design decisions get conjured in that chat. Mythic just so happened to be one of them, and less than 3 months after that we already had the core system up and running on the realm. The pre-alpha version (showcased April 16, 2020) already included full core support, database columns, variable key data, level scaling, quest timers, and the Challenge Mode NPC to tie it all together. The only "main" things missing were the rewards and the affixes.
From here on, I took it upon myself to start actually designing the system and come up with some concepts about how everything should work, while also leading and setting up keys for tester groups to feel out the way everything was at the time. 4 days later, I sent the first version of the "mythic notepad" to Radioactive, head developer at the time. In its 164 lines, it already explained the core system, scaling, rewards structure, and gave a basic idea on how the loot should be distributed into different tiers and awarded to the players.
The next couple of months consisted of a lot of mythic and dungeon testing in general. We knew that since the entire system is based on Cataclysm dungeons, they have to be working as close to flawless as possible, so the amount of problems that can come from mythics later on are minimised. By late July / early August, the 164 line notepad expanded to 912 lines, and got reviewed by some veterans as well. Shortly after, the "mythic reward system" screenshots image was put together based on that notepad, which is still pinned in #neltharion-chat to this day.
The very next changelog (released August 24, 2020) reflects this time period perfectly, in which over 75% of the documented changes are in relation to dungeons or mythics themselves.
The next period (late 2020) consisted of revisiting and reviewing previous work, and laying some small but very important foundations. This was also the time when we opened an entirely new chat for mythic related discussions, and put out a teaser video to hype the whole thing up. A lot of concepts and my personal vision for the system (what the original design reflected) had to be slightly adjusted and rebalanced. Loot was among one of the things that were discussed and introduced during this era. Here are some parts of the chats from then.
It also included a lot of visual representations that had to be made in order to better showcase our ideas and get our points across for the team. We each have made several videos, spreadsheets, drawings, such as this one (was too big to fit on the previous pic; early concept with diagrams thrown together in a few minutes on the spot):
There are some items that already exist without a patch and could be polled to be added to very high level (25+) level rewards, those being the 410 valor items (Heroic cloaks, rings, trinkets). They are mentioned in the original design and are the most likely get polled in the future, as they are the only non-standard items that still give a fair advantage, but are not completely overpowered. They would also fit the "mythics giving BiS gear rewards" theme from the actual retail system.
By this point, most of what the system is today had been implemented. Small, gradual changes were done to dungeons and mythics themselves. Recently, however, we decided to put some more focus on the visualising aspects of the system and the ideology behind its concepts.
For ideas that resemble the retail version of mythics, some things had to get reworked due to our limitations. Namely, the Enemy Forces counter and interface. In 4.3.4, the objective tracker can only have up to 4 different things tracked per quest. This means that in a dungeon with more than 4 bosses, we wouldn't be able to list all of them for you to kill. We also can not have multiple trackers running at the same time, as you would get rewarded more than once. So to combat this, we have predetermined lists for each dungeon, containing all NPCs and the amount you have to kill to get to 100% after which you receive your rewards. All bosses are included, some trash packs are still skippable, and the order in which you complete the list does not matter.
For the parts that are up to us to design, we have taken a lot of inspiration from other games with similar systems, that players love and enjoy, so we could skip the feedback part as that's been already done for the most part. One notable example, in particular, is RuneScape. With its 20+ years of history and Guinness World Record of the largest and most updated MMORPG, not to mention several members of our staff having vast amounts of experience in its world, it is a perfect candidate for a muse. Those familiar with the "retail" game and on the endgame side of things, you might have heard of Telos. He's an incredibly unique boss with groundbreaking design that still holds up more than 5 years after initial release. There are a lot of similarities between our mythic system and Telos.
To start things off, there's an "enrage" mechanic, which allows the player to fight a more powerful version of the boss, and have better chances at loot. Familiar enough? But wait, there's more. There's also a very defining learning curve, phase unlocks above certain levels, a lot of maths and statistics, not to mention the design of the reward system and streaking. This part is the one that had to get slightly reworked due to our limitations. Let me explain.
With Telos, you can "risk" all of your obtained loot by keeping a "streak" / "pushing enrage" / "continued challenge", and have increased chances at better loot and faster accumulated progression. However, if you "fail your streak" (die or lose the kill by other means), you can only cash out 75% value of your potential rewards. I personally felt like this was too big of a drawback, and since our theoretical cap of ~30 levels with the possibility of +3 max level upgrades is nowhere near comparable to the 400+ near perfect kills you need to do to in order to reach the enrage cap, this mechanic is completely redundant and therefore didn't even make it into my original concept. (You can read more about Telos and his loot mechanics here.)
Drop rates and rarity often operate on a mathematical process called binomial probability. The chances of getting your desired drop "on rate" (at least once in the expected amount of attempts) is only about 63.2%, which correlates to a constant quite commonly found in science and mathematics overall being e (Euler's number, being the base of logarithmic functions, probability calculations, compound interest, etc.), or in this case 1 - 1/e, with its curve (lim x->∞ (1-((x-1)/x))^x) representing the likelihood of getting what you want over time. If we take a look back at Telos for a second, here's an explainer thread from reddit that explains the way streaking works and how likely you are to receive drops.
Many links in that massive post link to images that look similar to this one:
It follows the same e^x function type curve as expected, no surprise there.
Here's how that system compares to our current one.
What I've done is slightly adjusted the scale from one of those graphs by overlaying it on top of ours just so that the x-axis matches up better, and put the ~63.2% on our theoretical maximum level key being 32.
From this image, the following things can be pointed out:
- The rates at any given point always add up to 100%, this means players will always receive some kind of reward at every level.
- There are compound effects in play, as you have to progress low levels in order to get to the higher ones, and you will always get better loot over time.
- The low tiers start off strong at low levels, but drop off gradually over level progression, first rapidly, and then slowly fade out.
- The med tiers are the most defining ones and best aligned with the curve, with slightly boosted values in the early and middle levels (to draw players in), with not much change or fluctuation overall.
- The high tiers start coming in around the middle, with no decline.
- There is no optimal point to stop pushing levels at, as the reward chances are strictly better at higher levels compared to lower ones.
After a lot of feedback, a lot of discussion, and a lot of number crunching, this is what we came up with and what's going to stay with the current system as we know it.
A couple key things to point out with the reward system itself:
- Your bag will upgrade to the next tier if you have obtained everything from that tier. This theoretically means you can obtain everything we have to offer from spamming level 1 - it would take a very long time, but still possible.
- Majority of the sought-after items, such as rare transmog sets, weapons, mounts, etc. start appearing at the middle point, and they are significantly easier to obtain from higher levels than lower ones.
- The system is dynamic enough where there's no optimal "farming point" for any tier of loot, the likelihood of receiving better loot is always increasing over time.
- The dark red line represents your average loot expectancy over time as you delve deeper into pushing keys.
- There are no duplicates (unless you sell/delete/use rewards) - which means when you get something you don't want, your chances of getting something you do want are increased the next time you are rewarded.
- Gold, Justice, and Valor Points are also rolled with each bag, but do not influence other drop chances.
- Lower tiers include a lot of BoE transmog sets, which can be traded/sold on the auction house to gather full sets.
- Some rewards are categorised by armour type and player class, meaning you can only obtain items that you can actually use.
Those are just some basic rules, there's a lot more to it that you will find out for yourself by actually playing it.
Aside from loot, the other important thing to talk about is the multiplier/level scaling. Here's what that currently looks like:
As it can be clearly seen, the damage scaling is exponentially more difficult for each level. Every new level is a flat 10% harder than the one below it. Precisely, the equation used is y = 1.1 ^ (x-1), where y is the damage and health multiplier, and x is the key level. The only exception is made at level 1, where there's a flat 5% boost difference from Heroic.
Affixes are introduced at different intervals as well, resembling a more logarithmic scaling than a linear "+1 every x levels" type deal. It is important to mention that our keys can have 4 total affixes at the same time, instead of the 3 as per retail. This was a deliberate decision for balancing purposes. The affixes themselves are split into 4 "groups", and players can only have up to 2 affixes present from the same "group" for their keys. This idea groups together similar affixes and makes "mechanically impossible" combinations way less likely to occur. Also, because keys are individually generated with random affixes (as opposed to retail's weekly rotation), there's a lot more variety and choices to be made by the players. You can read more about our current affixes at https://forum.projectnelth.com/topic/50-affixes-explained/. Keep in mind that these are subject to change and can be replaced/reworked entirely based on feedback. For the most part, however, we are satisfied with all of them.
During initial testing, we have found that 3 affixes were just not cutting it for the difficulty levels in most cases, hence why we opted to include a 4th slot in the generation. With the way everything works at the moment, we are overall quite happy with the scaling.
This is a good time to further clarify that apart from the existence of the system itself, nothing is custom about it, meaning every single affix, mechanic, voice line, visual effect, spell, etc. is already present in the game. This feature allows us not to force players to download any patches or custom game files in order to experience mythics. We will continue to work with this in mind as far as we can, however if there is a point where it's no longer possible to replicate the retail version with enough accuracy, we will put a poll out for it. With it also, we would be able to add custom downported models and visual effects that otherwise wouldn't be possible with the basic 4.3.4 client, and its' limitations.
The only remaining things missing from the current system are:
- Some loot tiers being filled in
- Certain pitched items have to get discussed whether or not to be added to high level satchels or the website's store
- The timer has to get changed to a dynamically scaling one, instead of the 4+ hour one that we currently have for testing purposes (already have some ideas)
- Some minor affix and logic corrections
- Base heroic dungeon fixes
It is also important to remember that your actual physical item that Aggra can give you by talking to her in any main city does not relate to your "key" at all, instead it is just an item with an onuse effect that displays information about your current key, which is nothing more than a few numbers in the database. The two are not connected in any other way, therefore deleting the item will not get rid of your progress (if you wanted to make bag space, for example).
Having all that said, that pretty much explains everything behind the current Mythic system and the way it was designed and built. It is months (if not years) of work boiled down to a simple click of a few buttons to start your adventure, which will probably resemble this flowchart originally made by Cryptid in a discussion.
Thank you for reading, and we wish you good luck in Mythics.
try - design and mechanics
Icnatsepll - loot tables and theorycrafting
Radioactive - system code and implementation
Hyjarion - affixes and dungeon functionality
Cryptid - Aggra
Redstorm - testing and feedback
& everyone else who has helped us with some runs along the way so far.
PS: The main mythic explainer threads have also been updated along with this post and the most recent changelog.