Lots of issues with layering abuse have been popping up, and there has been lots of talk about whats a better solution. I have also been giving it some thought ans was curious as to what people thought of the proposed solution below:
Firstly, I know that all the same issues of layers that exist today will still be in place with this idea. I feel that changing how assignment to, and migration between layers takes place will greatly reduce how noticeable it is, and reduce its ability to be easily abused.
I propose a hybrid layering system with a little invisable queing involved.
1. There will be a dynamic character 'cutoff level' that dictate which characters are assigned to the 'primary layer'. This level will be calculated via the following method:
1a. There will be concurrent 'logged in' snapshot taken every 5 minutes, where the layering system grabs the level of every currently logged in player across all layers for a server.
1b. It will then count the number of characters starting at 60 working backwards to 1, until the count reaches 50% of the primary layers max population.
1c. Whichever character level the count stops at is set as the primary layer's cutoff level for the next 5 minutes until the next recalculation.
1d. minimum level cuttoff is 20 (Negotiable).
2. Characters over the current cutoff level are assigned to the primary layer on logon, and cannot leave the primary server while their level > primary layer cutoff level.
3. The primary layer will assign additional characters below the cutoff level to the layer dynamically until primary layer population = 50% max layer population. (50% is negotiable)
3a. This dynamic assignment will occurr at logon only.
3b. As population on the primary layer approaches 50%, it will slow down the assignment of characters that are below the cutoff to the primary layer. Assigning 100% of below cutoff character logons to the primary layer when the layer pop = 0, Assigning 50% of below cutoff character logons to the primary layer when the layer pop = 25% max population, and Assigning 0% of below cutoff characters logons to the primary layer when the layer pop >= 50%+
4. When the primary layer hits 80% population cap, the layer will look for appropriet events to reassign sub 'level cutoff' characters off the primary layer to another layer.
Note: There is ALWAYS a 30Sec out of combat or post node interaction delay, and if you die from combat, the 30sec counter doesnt start until ressurection. (time length negotiable)
Layer Migration events are:
4a. taking a flight (instant)
4b. riding a boat (instant when boat rezones)
4c. entereing / exiting any instance area (instant)
4d. teleporting/hearth stoning (instant)
4d. character's distance from other players is greater than the max view distance for character rendering. (instant if ooc >0s)
4d. character's distance from other players is greater than the clients currently set view distance to render other characters. (if over ooc 30s+)
4e. Entering a city (remember the ooc 30s+)
4f. Dieing in certain areas while not in a party. Such as a city, or x meters away from mob spawn points, etc.
4g. Character logs off for 15m+
4h.Joining a party of all sub 'level cutoff' characters will have you migrated to the party leaders layer after 30+ secs ooc and are on the same contienent a a party member, or you trigger an instant condition. (so migrate off the primary layer)
4i. If a character on the primary layer invites or joins a party on a different layer, all characters not already on the primary layer will move to the primary layer after 30+ secs ooc and are on the same contienent a party member or you trigger an instant condition.
(Additional suggestions for reassignment)
5. If a logged in character crosses the primary layer cutoff level while logged in and not on the primary layer already, the layering system will look for appropriet events to reassign the character to the primary layer.
6. In the rare event the layer population hits 100% and a character over the current cutoff tries to log on, they are instantly assigned to the next available layer and then treated like an individual who had just leveled up past the level cutoff to be reassigned to the primary layer.
7.There is a queing system for both moving onto and off of the primary layer.
7a. If you are below the level cutoff when you enter the primary layer, you will enter a deport queue instantly, your order is based on the /played of all characters for the account. Highest /played bottom of the list, lowest /played top of the list. When population on the primary layer > 90% individuals at the top of the queue will qualify themselves for an "ungracefull" migration, Where all layer migration triggers in #4 will move you, but also Ungracefull conditions in #8.
7b. Characters who are not on the primary layer, but over the cutoff level will enter into a queue, prioritizing /played of all characters on the account. Highest /player top of the list, lowest at the bottom. When a spot on the primary layer becomes available, instead of migrating the next person to trigger an migration event, the layering system will reserve the available position to the character at the top of the queue, remove them from the queue, and wait for the character to trigger a migration event in #4.
7c. If a character has a reserved spot and has not yet migrated to the primary layer when the cutoff level is recalculated, and they now fall below the new cutoff level, the reserved spot will removed and they will no longer migrate to the primary layer if they trigger an event in #4
8. If a player qualifies for an ungracefull migration the following events can also trigger a layer change:
8a. Character enters ooc for 0s+ while not in party
8b. Character changes zones
8c. Character Dies
8d. Character exits of leaves rested area while ooc 0s+
(additional suggestions?)
As stated earlier, I know this solution wouldnt solve all issues with layering, all current layering issues would exist in this solution. But I feel that the changes to layering assignment and migration between the layers would reduce the frequency of the issues, make it harder to abuse. Also, you could introduce a secondary layer that acts like the overflow for the primary layer. so you have a primary layer cutoff of 40, and a secondary layer cutoff of 20, and all other layers. players will flow from the other layers, into the secondary layer if they meet the cutoff conditions, and then into the primary.
The goal is to reduce the layer hopping conditions for the player of levels where abuse of the layering system would most effect the econ. (Higher levels). Sure, it would effect the econ slightly if peacebloom was farmed and flooded the market, but that effect would have minimal effect on the end game econ compared to purple lotus.
Thoughts?
Nice ideas!
Tipsout posted something this morning, not sure what Blizzard already have done.
I am not sure if any of us can make constructive contributions without understanding the underlying network technology in detail. But that being said, this is my suggestion (cobbled together from other people's ideas)... Put a 20 minute timer on joining a party ( or maybe just completely block ) whenever someone is: flagged for pvp or in a contested zone unless they are in an inn, on a gryphon, at a meeting stone, or already on the same layer. Also make sure that a character stays pinned to a layer for at least 20 min after logging out. I think that would prevent all of the shenanigans, but unfortunately sometimes makes grouping up in the world for quests or pvp a bit of a pain.
Scrap layering. Create 3-5 individual servers (whatever their intended max layering number was) for each server name. Tichondrius 1, 2 and 3. These servers will cap at the max pop (3-4.5k, whatever Blizz decides they wanted on a layer). These servers get merged based on their active population status. For as long as each "layer" can survive without merging, it will. As soon as populations dip to 2.5ish players, they get merged permanently Tich 1, 2 and 3 are now just Tichondrius. One realm. No layers. This is SUCH an old idea. I began seeing examples of this popping up on the Blizz forums on the day layering was announced. Hell, this is how retail worked for years! Create new servers, merge dying servers. Having individual servers completely removes ALL of the issues we currently have with layering. Players will no longer be able to manipulate their layer. Players will STILL exist in a cohesive world without sharding. This solution STILL addresses server populations and will offer an easy merger experience after hype has died down and the tourists return to Retail. It performs ALL of the functions layering would have, without ANY of the drawbacks.
While I can acknowledge that they have invested a lot into layering, there is also a simple fix to remove it from the game. Do I think they will do this? No. Any further development costs that are spent on a failed idea are a waste in my opinion. Throwing duct tape on a terrible idea (as I have stated since the announcement of layering) is probably not good. This is the point of a beta right? Try things. When they don't work, change them. Layering didnt work. There are ~60 days left til Classic launches. Scrap it. If we were still 4-6 months out I could see spending more time on this idea, but we are not. Its crunch time.
edit: will also say that I appreciate the effort you put into your post OP, I just think that with 60 days left to launch (leaves very little time for the development and testing of new concepts) there simply is not enough time to mess around with layering anymore. It was an awful idea right out of the gate as was stated by most of the more knowledgeable crowd. All of these issues that we are currently seeing with layering, were foreseeable and predictable. See some of my old comment threads regarding this topic. I gave examples of ways that this would be exploited and within a few days we seen Monkeynews' infamous layer hopping method. Even though I would have preferred temp sharding confined to level 1-10 zones, there also is not enough time to explore that option. As a secondary option I would have preferred a polished version of dynamic respawns, but there is nowhere near enough time to explore this option. We are left with layering or multiple servers functioning as layers. These are the only two options we have time to implement as I see it.
I'm not a fan of layering but blizzard is expecting MASSIVE player drop across all servers. I would estimate 60-70% within the first month and its already apparent in the beta. Merging servers might appear to be a great idea however its very likely that the number of concurrent players will increase at the beginning of each phase. Lets say we merged server 1, 2 and 3 together how do you expect a single server to now deal with twice or three times as many concurrent players without outrageous queue times. You cant.
Its unfortunate but you just have to deal with it for the time being.
Nice ideas!
Tipsout posted something this morning, not sure what Blizzard already have done.
https://twitter.com/TipsOutBaby/status/1143897554469322752
My apologies to aeh and Stfuppercut, I did not realise you had already posted the tweet in this thread. I included a bit more of his replies in my post in the other thread for some more context/info.
They appear to be sticking with layering and looking to patch up some of the exploitable holes that are present. They seem to be making some pretty decent changes (obviously its all in a testing phase, and I would imagine a lot of it based on feedback from the stress test).
I think it's pretty obvious no one wants layering long term, so if they can make it functional, bearable and temporary then that is a decent compromise. Certainly is too late to make any drastic changes to a completely layer free solution, as mentioned by Stfuppercut.
I think layering is here to stay (at least for the duration that Blizzard deems fit) because all other alternatives have downsides that are far worse than what layering has to offer. This has been discussed to death on everywhere so I won't go to their details here.
We can only hope that they continue on improving it so it (1) doesn't break immersion and (2) not abusable (so no "layer hopping").
If Tips' tweet is accurate, that's quite hopeful indeed.
Choosing a layer would be functionally identical to choosing a server from a players perspective and would allow Blizzard to add and merge layers as they are needed, without wasting money on servers they don't think they'll need in a few months.
This would also allow layering to continue past phase 1, as they would no longer have any effect on gameplay.
I understand the argument for layering, what I can't get my head around is why players need to be able to jump between them.
Very well written. This was basically what I was trying to say. I feel like dialing in layering and collapsing it properly later on for a final merge is going to be REALLY, really difficult for Blizz. I'm also really concerned about their player count estimates. Layering requires a lot of really accurate estimations for it to end well, where as separate servers do not. I'm certain that the merge phase of layering will be different on every realm and for some realms, they will knock it out of the park. For other realms this could be a nightmare. Ever since layerings announcement I have been primarily waiting to hear two numbers. What is the final decision on players per layer (we have a 3k estimate) and how many layers are possible per server? These are the two most important variables in this entire discussion.
There is absolutely zero chance of them "scraping layering" and going with something else this late in the release cycle.
Now I'm a bit conflicted if I should copy paste my reddit (*gasp*) blurb or just link to it.
Improvements to the layering system to hide the immersion breaking and community breaking side-effects of it better is the name of the game at this point.
Layering
What is it supposed to fix?
(imposed by technical limitations but with game design to match)
The good
The bad
Look to Blizzard 2004 for why you want an open seamless world in your mmo-rpg with no invisible walls or loading screens (which sharding and layering both are essentially a "kind of")
Suggestions for improvement to the core system.
This means update the layer affiliation on all the characters under an account to the last layer any of the characters played on. This will stop players playing a solo "layer roulette" by parking disposable alts to different layers and using them to layer hop their main.
I.e. try to do layer transitions in loading screens. It makes sense from a gameplay perspective as well, there is no reason to layer a player out of their immediate environment to a player that is in another continent.
How could Layering work to mitigate launch period early zone congestion?
Layer the most populated brackets aggresively and shunt the least populated brackets into the same layer.
This would spread lower levels to layers and group all the powerlevelers together "naturally" fixing a bunch of side-effects in one go (for example abusing layering to get a huge leg up early on for scarse resources) as well allow the low (proportionally) population of powerlevelers to interact without layering getting in the way.
Finally; what happens if (some) megaserver populations do not "automatically" settle down to concurrency numbers that allow to keep both 2006-like server population and a single layer?
Do queues come into play?
Free transfers?
Increase layer caps to double or triple the 2006 realm population?
Open new mega-servers?
Keep Layering?
I think we will probably end up with some form of Layering extending its stay.
Unfortunately I can only think of "bad" solutions.
I'd definitely be interested to hear suggestions.
"Megaserver populations will magically settle down to 3k single layer concurrency by P2" is imho *not a valid answer*.
Banking your game integrity on failure is not a good strategy.
What happens if they do not?
What happens if Classic doesn't deflate but inflates instead?
PS. I considered traditional *per zone sharding*, *login-queues* or *Alonsus-1, Alonsus-2, Alonsus-3 servers* using the same name pool and ready to be merged down the road as alternatives but I think that a properly implemented megaserver + Layering system is superior to those for reasons I'd rather not add to this wall of text but happy to discuss in comments.
Wall of text also assumes most people are familiar with the similarities and differences between traditional per zone sharding vs per continent layering vs server clusters etc.
PS. I considered traditional *per zone sharding*, *login-queues* or *Alonsus-1, Alonsus-2, Alonsus-3 servers* using the same name pool and ready to be merged down the road as alternatives but I think that a properly implemented megaserver + Layering system is superior to those for reasons I'd rather not add to this wall of text but happy to discuss in comments.
I agree with most of what you wrote. I also believe that most of what you wrote lends itself to the argument against layering. As you have outlined your version of pros and cons, I think your cons VASTLY outweigh the pros in regards to layering. I would be interested in hearing your opposition to multiple servers "*Alonsus-1, Alonsus-2, Alonsus-3 servers*" instead of layering in its current form.
Don't ever worry about creating a wall of text here. This is a text based forum, any and all efforts to add to the conversation and stimulate something interesting for others to read should be encouraged.
I would be interested in hearing your opposition to multiple servers "*Alonsus-1, Alonsus-2, Alonsus-3 servers*" instead of layering in its current form.
The short version is that the same thing that causes most undesirable effects of Layering (dynamic nature) is what provides the benefits.
But the devil is always in the details, implementation matters.
Servers are inflexible.
If you get discouraged by login queue on Alonsus-1 and decide to roll on Alonsus-2.
Blizzard have a long history of "dragging their feet" when it comes to free transfers or server merges.
So you have a good chance to be stuck on a low pop realm for a while or to have to bite the bullet and buy transfer as a way to fix their problem.
At the same time you have no in-game means to communicate with the other servers (no common chat, no trade, no guilds or in-game friend list).
Assuming server clusters that share a name pool (ie when you create a character in Alonsus-2 it checks if the name is available in the whole cluster of Alonsus-1-2-3) same for guild names etc, an eventual server-merge will not be as jarring as forced renames but we still have examples of friction when 2 established communities merge even if no one had to give up their name or guild tag.
Servers develop "personalities".
On the other hand a properly implemented mega-server and layering system should be a case of "If a tree falls in a forest and no one is around to hear it, does it make a sound?".
With a pre-requisite that layer transitions are "hidden" good enough that players do not notice them in the course of their regular play:
You never have a login queue.
You can interact with the entirety of your mega-server using in-game chat, trade/ah/mail, grouping and forming guilds.
Both daily fluctuations of population (primetime) and over time get smoothed out for all intents and purposes "immediately" (compared to having to go through a server-merge)
That's the main things I could say, now granted a *bad implementation* of Layering where layer transitions are obvious and frequent is obviously not preferable.
You can stop here but to expand a little on a side that's probably not very interesting to players themselves but is good to keep in mind as a factor on the business / developer side and what's reasonable to expect.
The era of the Realm Serverblade where each realm was running baremetal is far gone and not coming back 
It's all about virtualization and cloud nowadays (for good reasons)
Their CRZ is the "most efficient" in terms of allocating resources to match the population and minimize the cost across the board.
Layering as described is not as efficient (so they are already taking something of a hit for game preservation reasons - and obviously not scaring away their potential customers)
Because both the "world-unit" is larger (continent vs zone) and the pool of players to fit into each world-unit is smaller (mega-server vs server-cluster)
So you can still end up allocating resources (spinning up) a Kalimdor-Layer that is underused.
Anyway from what we can reasonably assume there's 2 kinds of layer transitions happening at the moment.
We do not actually know that it is split in two but that is the most logical way to do it both to avoid "dead overflow layers" (what would happen if a layer was a direct 1:1 replacement for a queue) and for efficiency reasons.
If layer was a direct replacement for queue instead of seeing "position in line 400: estimated time x" you'd automatically login to a layer that's severely underpopulated.
What happens if a couple raids from overflow layer #2 get invited into a layer #1 that's "at capacity". Does it keep expanding, do 20-40 ungrouped people get shunted into layer #2 to compensate etc?
So it's safe to assume that what is happening instead is that if a layer reaches 3000 (or whatever their "this layer is at capacity" number is) it gets split in 2 while the server goes through a list of group/guild/friend and possibly distance criteria to decide which part goes to Layer #1 and which goes to Layer #2 (keeping groups on the same layer, trying to keep guilds on the same layer, trying to keep friends on the same layer in descending priority) with unguilded/ungrouped players having the lowest priority and being used as "filler".
Similarly the login server shuffles the deck so to speak assigning new logins to Layer #1 or Layer #2 trying to keep them both at a decent population.
Now if for example primetime is over (it's now 2AM and the megaserver population dipped) if your megaserver has 2900 players total they will do the opposite and move them all into Layer #1 freeing all the resources used by Layer #2.
TL;DR; IF layer transitions are hidden well enough the megaserver + layering system offers the best of both worlds.
Big communities able to communicate in-game, a much wider range of population that remains within the 2006-healthy population at all times (both primetime and dead-hours) as well over the megaserver life.
Servers are inflexible.
If you get discouraged by login queue on Alonsus-1 and decide to roll on Alonsus-2.
Blizzard have a long history of "dragging their feet" when it comes to free transfers or server merges.
So you have a good chance to be stuck on a low pop realm for a while or to have to bite the bullet and buy transfer as a way to fix their problem.
At the same time you have no in-game means to communicate with the other servers (no common chat, no trade, no guilds or in-game friend list).
Assuming server clusters that share a name pool (ie when you create a character in Alonsus-2 it checks if the name is available in the whole cluster of Alonsus-1-2-3) same for guild names etc, an eventual server-merge will not be as jarring as forced renames but we still have examples of friction when 2 established communities merge even if no one had to give up their name or guild tag.
Servers develop "personalities".
I see these as strengths. The servers should be inflexible. Each community is its own. The difference here is that the merger is inevitable as is implied by the shared name. When Blizzard does choose to merge these servers, the merge is seamless. I have more faith in Blizzard merging realms that were always intended to be merged than implementing layering. Servers do develop personalities, layering will prevent this though. Perhaps these personalities will clash when servers are finally merged, but isn't that a good thing? Doesn't that mean that these servers had a personality? There will be no personality on a layered server. Raid groups running to get world buffs and then going to raids, running RIGHT through phantoms on the opposing faction (either by choice or by chance). Instead of a mass battle that helps shape the personality of the server, players feel a subtle chill and continue about their way. "What was that? It was probably nothing..." Tons of missed opportunity for interactions. Chatting with players who you cannot see. Leveling side by side with phantoms you will never meet during your leveling experience. That guy that keeps killing the ratchet flight master and is such a jerk? He only exists on layer 5. He wont define your servers personality. He has no impact. That server first ironfoe wielding warrior standing on Ironforge bridge? You cant see him. Your server has no personality. You are just a layer and layers are temporary. Layers are constantly changing. This is a massive issue in my opinion.
Having played on private servers I realized one major reality. People are valuable when there are less of them. At around 2-3k population, you remember players and guilds. The interactions you have with others are meaningful and you remember them. At around 3-6k players are less valuable, there are a lot more of them and you only interact with the players you know. You have such easy access to other players that you do not expand your inner circle. When servers are at 6k-12k players are not only useless, but they serve to hurt your experience. Like any resource, players are more valuable when there are less of them. If you are living in a large city, you likely avoid those around you because people are nuisances but in a small town your neighbors become your support network. Big server populations kill community. They kill a servers personality. Take this big server population and start fragmenting it into invisible layers and it becomes even more meaningless. At around the 8-12k mark on private people start spamming racial obscenities, ninja looting without any concerns and displaying generally poor social behavior. Why though? Because the social ramifications for your actions are non-existent. Add layers to this and I can only imagine how this will impact a servers personality or the lack there of.
On the other hand a properly implemented mega-server and layering system should be a case of "If a tree falls in a forest and no one is around to hear it, does it make a sound?".
With a pre-requisite that layer transitions are "hidden" good enough that players do not notice them in the course of their regular play:
You never have a login queue.
You can interact with the entirety of your mega-server using in-game chat, trade/ah/mail, grouping and forming guilds.
Both daily fluctuations of population (primetime) and over time get smoothed out for all intents and purposes "immediately" (compared to having to go through a server-merge)
I see these as cons. Interacting with phantoms sounds awful. We have also seen layering be abused AND we have seen layering limit players to very small numbers in their own layers. Even if Blizz is able to straighten out these wrinkles with layering we still have the collapsing phase to worry about. What happens if the retention rates aren't as high as we hope? Will they get rid of layering early enough? They didn't on the beta... Which seems to indicate that this is not a fully automated system as Blizz had implied. What happens if the retention rates are higher than Blizz think they will be? How many layers with how many players? Layering offers a VERY high potential for this system to have an extended stay in the game. Separate servers are functional communities offering a Blizzlike experience from day one.
That's the main things I could say, now granted a *bad implementation* of Layering where layer transitions are obvious and frequent is obviously not preferable.
Well over a year into development and that is what we have. Bad implementation. Layering in its current state has a ton of issues. Issues that Blizz are addressing but we are also 60 days to launch.
You have some interesting insight into transitions. I will say that transitions will always be apparent to a certain extent and will ALWAYS be the worst thing for immersion. Whether this is watching someone disappear due to a group invite or traveling on a boat full of people, hitting a load screen and loading into a layer with an empty boat. Due to this, having one persistent world on a separate server just seems better for the game. Better for the community. Better for the player. The only variable that we need to be concerned with when considering multiple servers, is the merge phase from Blizzard. While I understand people having reluctance in Blizzard merging at the appropriate time, this same concern exists within layering which also has a collapse/merge phase. The reality is that layering requires A LOT more interaction by Blizzard. Consistency. I would rather be on a consistently low pop server than a server that shifts wildly within every play experience.
World of Warcraft was about the world. We hear this all too often. Layering adds worlds. The game becomes Worlds of Warcraft. I have less faith in Blizz addressing ALL of the issues with layering than I do with them using the same solution they always used before sharding. Separate servers with merges to dying realms.
Your entire post was a pleasure to read, very insightful. It sounds like you have put a lot of thought into layering.
I get the contrasting viewpoint on the rest of your post and I don't really have something to say, layering is not perfect (especially in its current state).
The reality is that layering requires A LOT more interaction by Blizzard.
For this part I'll take a small issue.
Layering after they tweak it to how they want it will be an automatic process, it's nowhere near the deliberation required to decide to merge 2 servers (or open free transfers).
"Is this dip temporary, will it pick up because ZG in 1 month? shall we delay it" and while Blizzard is trying to weigh options some of the more casual players will be gone, there's a special kind of players that will be patient for 2 months waiting for a fix to their realm.
With Layering the server software will decide when a megaserver is at Layer1 capacity and follow an algorithm to split it to two layers or more.
Layer1 = 3000 -> Layer1: 1500, Layer2: 1500 > server pop = 6.000 > Layer1: 2000, Layer2: 2000, Layer3: 2000, total pop 9000 (4 layers at 1500 each again)
The reverse process when population goes from 5k to 3k, it'll automatically collapse the layers into 1.
There won't be anyone flipping switches on the other side :)
At the moment that algorithm that decides who gets assigned to which layer is pretty rough why we see all those side-effects (and I'm acutely aware)
That's not manual layer hoping, it's most likely pop rebalancing and they get shunted to another layer as they're ungrouped/ low prio.
Getting "layered" off the gryphon (extra funny)
or
https://clips.twitch.tv/PatientGloriousMomTBTacoRight
The algorithm used to decide layer assignment both on login and during the session is super important and some of the suggestions made are to make it so you do not see layer transitions.
For example you can see how that first clips wouldn't even be possible if combat and player/npc/gameobject interaction "locked you to your layer" until it's over.
The last clip wouldn't be possible if layer was saved for the entire account and not each character. (it wouldn't be impossible but it would require a ton more setup, find someone on another layer, get an invite etc, at some point the fuss is not worth it for a node)
The internal cooldown they already put on layering will help somewhat but it's like doing surgery with a hammer, more precise tweaks are necessary.
Because it is tweaks and not a complete overhaul I believe they have time to improve it in the 2 months to release.
Roadblock
Interestingly enough not all of those clips were the result of layering. One in particular was the result of phasing (as it was labeled). Zones in Vanilla actually had jurisdiction boundary lines and the phasing you experience while crossing the boundary between zones WAS part of vanilla. In the clip where the guy is fighting the croc in the river separating Durotar and the Barrens, this is the result of phasing between boundary lines, not layering. You will actually notice the zone names flash on his screen. This particular zone boundary is weird, it stays VERY close to the Durotar shoreline all the way Northbound until you approach Org and then shifts wildly to the west, creating a very odd boundary line that does not match the coast of the river. I used to know most of these boundary lines by heart because I was highly invested in world PvP. While the world was an open concept, your computer was always phased into a specific zone to reduce the strain on performance. This same experience can be replicated on Blizz like private servers that do not have layering tech. Conversely, the priest in the tower fighting in the barrens who had extra guards spawn DID experience a layer shift. The similarities between these two experiences are notable but the implications of both are vastly different. You will likely remember chasing someone into another zone while in Vanilla and when the player crossed the boundary line he would disappear temporarily, until you also crossed the boundary line. This often created odd interactions where a player could jump back and forth between boundary lines to avoid being ganked. I used this method in vanilla WoW for world PvP.
Side note: knowing your boundary lines will be key to defeating certain opponents. Being chased by a hunter and want to deadzone him? Run away until you phase at a boundary and then wait for him RIGHT on the barrier of the phase, he will walk straight into you allowing you to gap close without using any abilities or cooldowns.
The reason why I see layering as more manual process; despite the fact that Blizz sells layering as an automated system (which it mostly is) is that Blizz will still be responsible to shut layering off. At a certain point, Blizz will need to begin the collapsing process regardless of realm pop if we are to believe that layering has an end date. This process will be a systematic shut off of layers on each server, slowly collapsing the server to its final population. We have been told that this will be a slow process. To me, this means that the algorithm for layering will need to be adjusted during this phase to slowly bring a server down to one final layer. Separate servers are their own healthy experience and would only need to be merged when their population dipped to a concerning level. I guess what I am trying to say, is that regardless of how automated layering is, it will still be someones responsibility to shut layering off. This human interaction will require the same amount of intervention as merging two servers. Either way, it takes a human to turn off the tap. While layering will naturally collapse or expand based on player count, the system itself will slowly need to be decreased or removed. I assume (big assumption here) that Blizzard will shift the layering algorithm as we approach the shut off date and they will aim to increase the quantity of players on any given layer, reducing the overall layer count. They will continue to do this over time as we edge closer and closer to their final merge date. When the time is right, they will reassess and turn the system off. Regardless of how automated layering is, someone will ultimately need to turn it off. Though it becomes very clear that the needs of each individual server will be VERY different. What about the streamer server that will ALWAYS boast a massive player count? It will be really interesting to hear more about layering as we approach the release of Classic. I think you have a very level headed view of layering though.
I am not at all convinced that Layering will be "turned off" by P2, despite statements to the contrary.
This is entirely dependent on how megaserver populations trend closer to P2 and all we've heard so far is the equivalent of "fingers crossed" megaserver populations automatically settle down to 2006-like pops.
What happens if your megaserver still has 9-10 or 15k players at the end of P1?
I'm also 100% sure 2006 vanilla did not have phasing of any kind, this boundary (in actually map cell) "phasing" is a private server thing since Nostalrius.
It is however possible that some of those clips do indeed show early builds of the classic beta where they were still transitioning from their early plan (Blizzcon beta time) to just shard the early zones to the per-continent layering they ended up with later.