PVE redesign

Unknown

Demiurge,

On the whole, I've found you to be a responsive, eager-to-please studio. You release frequent updates with quality of life improvements, and requested features, to say nothing of all the new heroes.

That said, I have to point out that your implementation of PVE is in need of serious revision. I understand that you're revising it constantly, which I'm happy about. But you appear to be working within a set of constraints that's almost dooming you to failure from the start. Here are the two difficult constraints that I think are making your lives (and ours) hard:

1) PVE events should be competitive.
2) All players, no matter how new to the game, need to be able to play competitively through all PVE content.

Both of those sounds great, but their combination is causing you to implement bizarre rules of ever-increasing complexity and opacity, such as rubberbanding and scaling. These bizarre rules are going to cause your inventive player base to keep coming up with increasingly tedious, unfun, and probably unintended ways of gaming your system. I suggest that you stop trying to tweak or add such rules, and start looking at ways to simplify the whole design.

The biggest simplification would come from eliminating the competitive nature of PVE; you already have PVP, after all. But I think that's going too far. You've shown us that competitive PVE has the potential for being a lot of fun, even if you haven't found the correct implementation just yet.

A more reasonable simplification would be to eliminate the requirement that players of any level be able to play through all PVE content. It's a nice idea, and speaks well of you, but I think it's been way more trouble that it's worth. New players don't even expect it. I played for weeks before I took a good look outside the prologue, and when I did was completely surprised to find that there was actually stuff I could participate in. During those three weeks, while I was hacking through the prologue and totally unaware of your thoughtful gift to me, more advanced players were suffering because of that gift.

But if you're truly committed to working within both these constraints, then the next post has an idea.

Unknown

Do away with the whole notion of accumulating piles of points through repeated completion of missions; grinding isn't fun. Keep the story-driven nodes down to one play apiece as you have them now, and introduce a set of special nodes where all the competition happens. These special nodes would appear in little consecutive chains of three, sort of like the Savage Land layout. The first node is meant for 1* players, the second for 2* players, and the third for 3* players, with point values of, let's say, 1000, 2000, and 3000 respectively. These point values never change. No rubberbanding or any such nonsense.

This is the hard part: the point value of a node represents the maximum score you could get for a completion of that node. What determines the actual score you get for completing a node is your performance in the fight. You'll have to come up with a clever system of evaluating a player's performance in a fight based on several criteria, then award them points based on their score. Some thoughts on a scoring system:

-First of all, nothing idiotic like looking at whether the player completes the fight at full health. Do that, and you're just begging everybody to start gaming the system with healing characters. Instead, look at total damage taken.

-Consider how many turns it took to win. Fewer is better.

-Consider how many abilities the AI team used. Fewer is better.

-Probably avoid considering actual time elapsed; careful contemplation of a move shouldn't be punished!

-For players that have beaten the 3* node, retroactively grant a perfect score on the 1* node to avoid incentivizing the repetition of trivial content.

-Remember the player's highest score, and don't replace it with a newer, lower score.

Repeated plays of a node don't accumulate points, they just give you a shot at a better score. If you're close to the max score, your time is better spent elsewhere. Thus a chain of nodes with values 1000, 2000, and 3000, given absolutely perfect play, would be worth 6000 points. Realistically, that should be unattainable. Top players should be getting in the 5000-6000 range. These point values are just examples, but they're large enough so that we don't get too many ties, which are bad in competitive play.

Each map has several chains, and the chains can be of varying difficulty. However, their difficulties should be static. No going to bed and waking up to find your chain has outgrown you, and now you can't compete.

So, static point values. Static enemy levels, but with a wide range. Incentives to focus on the most challenging nodes.
No rubberbanding. No scaling.
No 36-hour events with only two important hours. An hour of good play at the event's start is as good as at any other time. No more setting alarms for weird hours.
No more grinding; if you had an awesome fight and got near the highest score possible, it's time to move to another node! Your work here is done. It's like that nice feeling of breaking through a point threshhold in PVP and finally shielding up. You get to relax and move on.

I realize that this is a big change, but if we're all being honest, your current system needs a big change. Please consider this one, or any other that addresses the problems that your community has been very vocal about.

NorthernPolarity

Susra wrote:

Do away with the whole notion of accumulating piles of points through repeated completion of missions; grinding isn't fun. Keep the story-driven nodes down to one play apiece as you have them now, and introduce a set of special nodes where all the competition happens. These special nodes would appear in little consecutive chains of three, sort of like the Savage Land layout. The first node is meant for 1* players, the second for 2* players, and the third for 3* players, with point values of, let's say, 1000, 2000, and 3000 respectively. These point values never change. No rubberbanding or any such nonsense.

This is the hard part: the point value of a node represents the maximum score you could get for a completion of that node. What determines the actual score you get for completing a node is your performance in the fight. You'll have to come up with a clever system of evaluating a player's performance in a fight based on several criteria, then award them points based on their score. Some thoughts on a scoring system:

-First of all, nothing idiotic like looking at whether the player completes the fight at full health. Do that, and you're just begging everybody to start gaming the system with healing characters. Instead, look at total damage taken.

-Consider how many turns it took to win. Fewer is better.

-Consider how many abilities the AI team used. Fewer is better.

-Probably avoid considering actual time elapsed; careful contemplation of a move shouldn't be punished!

-For players that have beaten the 3* node, retroactively grant a perfect score on the 1* node to avoid incentivizing the repetition of trivial content.

-Remember the player's highest score, and don't replace it with a newer, lower score.

Repeated plays of a node don't accumulate points, they just give you a shot at a better score. If you're close to the max score, your time is better spent elsewhere. Thus a chain of nodes with values 1000, 2000, and 3000, given absolutely perfect play, would be worth 6000 points. Realistically, that should be unattainable. Top players should be getting in the 5000-6000 range. These point values are just examples, but they're large enough so that we don't get too many ties, which are bad in competitive play.

Each map has several chains, and the chains can be of varying difficulty. However, their difficulties should be static. No going to bed and waking up to find your chain has outgrown you, and now you can't compete.

So, static point values. Static enemy levels, but with a wide range. Incentives to focus on the most challenging nodes.
No rubberbanding. No scaling.
No 36-hour events with only two important hours. An hour of good play at the event's start is as good as at any other time. No more setting alarms for weird hours.
No more grinding; if you had an awesome fight and got near the highest score possible, it's time to move to another node! Your work here is done. It's like that nice feeling of breaking through a point threshhold in PVP and finally shielding up. You get to relax and move on.

I realize that this is a big change, but if we're all being honest, your current system needs a big change. Please consider this one, or any other that addresses the problems that your community has been very vocal about.

I agree that PvE needs to be revamped in a major way, but there are some major problems with making the points be based off of your "performance" in the level simply due to how hard it is to actually quantify performance. The devs method of having it based off of health isn't working well, and I don't think your suggestions of it being based off of turns played and enemy abilities used would work well either. Doing this would just completely disincentivize the players to use characters that are meant for drawn out fights such as LazyCap and just focus on using characters that can win as fast as possible. I think total damage taken would also be even more "idiotic" than net health lost as this would just encourage players to game the system by taking damage down to a sliver of health, then heal up, then take more damage down to a sliver of health, heal up, etc. At least the current system doesn't encourage you to endlessly take damage to lower your MMR, which is what that would do.

I think d3 is just faced with a hard problem. Any attempt to quantify how well the player is doing in a match is futile because that forces the player to play in such a way to maximize his performance in the event, whether that be the player forced to use characters such as Patch over LazyCap to win faster, or repeatedly take damage to make it seem like he's having a hard time to reduce scaling. The only fair way to do this is to make difficulty scale identically for everyone and scale based off of some static value, whether that be the global point value that the player has earned in the event or the amount of times the player has beaten a particular node. The problem with this is that it will probably lock non-3* rosters out of top rewards (which they probably dont deserve anyways, but thats a separate topic since d3 wants to keep this property in their game) and force players to grind to get to a level of scaling that suits them, so I don't know how d3 can approach this without screwing someone over.

Unknown

I personally crave variety. Most PVE's lately have been repetitions of the same formula dark avengers + goons and sometimes maggia. I would be so happy to see an event that released skrull copies of everyone on the battlefield and I never saw a combination that was just three of the dark avengers again.

Unknown

Susra wrote:

-First of all, nothing idiotic like looking at whether the player completes the fight at full health. Do that, and you're just begging everybody to start gaming the system with healing characters. Instead, look at total damage taken.

-Consider how many turns it took to win. Fewer is better.

-Consider how many abilities the AI team used. Fewer is better.

Except that the latter two considerations beg everybody to game the system with infinite turn and stunlock characters.

Unknown

My wording on the damage-taken thing was bad. I meant the opposite of what you interpreted, NorthernPolarity. We should give more points to players who take less total damage. But now I'm having second thoughts about even that. A player who manages to soak all their damage with Patch or Daken, or who manages to heal it all with OBW or Spiderman deserves something for their skill. Ideally, these criteria will encourage the sort of gameplay that we mostly want already. We already want to get through matches quickly without taking much damage.

I would dearly love to remove the grinding and scaling and such as stated above. If it can be done without my individual-performance-measurement thing, then great! But as you said, it's a hard problem. Hence the crazy, opaque rules Demiurge has come up with.

Infrared

I like your idea of having your PVE score based on your best performance of each node, rather than accumulating points from grinding the same node repeatedly.

What if we combine this with a form of scaling? The higher the scaling, the higher the score. So under this system you would keep winning a node until it scales past your ability/roster strength. And your score for that node would be from the highest scaling you could beat. That is one easy way of determining performance. It also makes scaling desirable, since it directly improves your event ranking, rather than the current situation which is quite the opposite. Reaching level 400 Daken would be something you boast about instead of being something to dread.

Unknown

Cryptobrancus wrote:

I personally crave variety. Most PVE's lately have been repetitions of the same formula dark avengers + goons and sometimes maggia. I would be so happy to see an event that released skrull copies of everyone on the battlefield and I never saw a combination that was just three of the dark avengers again.

There was the simulator event.. Please do it again, devs

Knyghtmare

Personally, I feel everything should be based on the level of your top 3 usable characters (some PvEs only allow certain characters). Why should someone be penalized for doing better with their same level characters than another person who isn't as skilled? In the end, the best players who play the best should win. Overscaling and rubberbanding can take away from that since you can eventually be forced to fight level 400 characters, which aren't only harder to fight, but take far more time to beat.

Unknown

Knyghtmare wrote:

Personally, I feel everything should be based on the level of your top 3 usable characters (some PvEs only allow certain characters).

It's not going to work. Quite the contrary, this is yet another variation of trying to guess a suitable difficulty setting and failing all the way. Let's just look at an example:

a.) Take a level 230 X-Force, a level 141 Daredevil and a level 141 Loki. Sum of levels is 512.
b.) Take a level 141 Lazy Thor, a level 1 Spiderman and a level 1 Hood, all at full covers. Sum of levels is 143.

Which team do you suppose would fare better in PvE? Would you willingly pick the first team, knowing that your encounters will be more than three times as hard as for the second team due to the sum of levels? Would you sell that first team, no matter the effort spent, after realizing it has permanently ruined your account for most future PvE events?

Unknown

Did we ever see a rationale disclosed for scaling? If so it would be interesting to match against reason and practice.

Meanwhile without rationale it's best option to start with a clean state and add only elements that we know to make sense and why.

Unknown

A possible refinement that keeps infinite-turn characters from getting out of control: measure game length by number of actions taken, not just by how many turns pass. So chaining together Classic Magneto blue 5-matches and destroying the enemy team on your first turn would still be counted as many actions taken by the player.

Unknown

pasa_ wrote:

Did we ever see a rationale disclosed for scaling? If so it would be interesting to match against reason and practice.

That's really the best point brought up so far. Why bother giving feedback if there isn't any response at all? For all we know the devs never visit this forum, so any time and effort spent here is for personal entertainment only. Any chance we could get some kind of "chat with the devs" session, maybe once a month for 30 minutes?

Unknown

Moghwyn wrote:

pasa_ wrote:

Did we ever see a rationale disclosed for scaling? If so it would be interesting to match against reason and practice.

That's really the best point brought up so far. Why bother giving feedback if there isn't any response at all? For all we know the devs never visit this forum, so any time and effort spent here is for personal entertainment only. Any chance we could get some kind of "chat with the devs" session, maybe once a month for 30 minutes?

The devs have said that they read every post on the forum, save for the ones that start with "question for the devs" or "hey devs, do this" - any of those kinds of threads.

As to the actual question, it has been stated before:

Demiurge_Will wrote:

Scaling missions individually doesn't do enough to satisfy the design goals of this feature: making our events maximally interesting for the widest possible subset of players and reducing grind. We couldn't do rewards like the 5000 Iso-8 challenge mission reward in this event in a system like that. It would either be too trivial for top-end players to get or they would be the only people that had any hope of getting it.

There have been a couple of other threads where the devs have weighed in before on whether it's accomplishing what they intended it to. In my mind (and I harp on this, whether it's right or wrong; I actually expect Ice to come lay the smack down on me at some point), the actual problem is that they've made PvE competitive and as such need to create a number of meta-mechanics (mechanics outside the core match-3 gameplay) to try and restrict the content so that only <some percentage> of the players come out of an event with all 3 covers.

I feel that personal scaling could still take place outside of the competitive PvE context, so that you would still be handing out a limited number of all the covers, but that your own playstyle and dedication would be the sole determinant for rewards (or if you wanted a community aspect, do either alliance-based or global progression bars, or any other mechanic to force the community to work together against The Man in a community-based cooperative pve.

Unknown

iPulzzz wrote:

Cryptobrancus wrote:

I personally crave variety. Most PVE's lately have been repetitions of the same formula dark avengers + goons and sometimes maggia. I would be so happy to see an event that released skrull copies of everyone on the battlefield and I never saw a combination that was just three of the dark avengers again.

There was the simulator event.. Please do it again, devs

I guess they listened

Knyghtmare

Moghwyn wrote:

Knyghtmare wrote:

Personally, I feel everything should be based on the level of your top 3 usable characters (some PvEs only allow certain characters).

It's not going to work. Quite the contrary, this is yet another variation of trying to guess a suitable difficulty setting and failing all the way. Let's just look at an example:

a.) Take a level 230 X-Force, a level 141 Daredevil and a level 141 Loki. Sum of levels is 512.
b.) Take a level 141 Lazy Thor, a level 1 Spiderman and a level 1 Hood, all at full covers. Sum of levels is 143.

Which team do you suppose would fare better in PvE? Would you willingly pick the first team, knowing that your encounters will be more than three times as hard as for the second team due to the sum of levels? Would you sell that first team, no matter the effort spent, after realizing it has permanently ruined your account for most future PvE events?

Your argument is that either a player leveled up 3 bad players, or that someones team consists of Thor and two number 15s (3 stars start with lv 15). If you look around, do you think anybody team would be this ridiculous. People level up the characters they want to be their top characters. If someone is stupid enough to make either team, they get what they deserve.

Unknown

Knyghtmare wrote:

Your argument is that either a player leveled up 3 bad players, or that someones team consists of Thor and two number 15s (3 stars start with lv 15). If you look around, do you think anybody team would be this ridiculous. People level up the characters they want to be their top characters. If someone is stupid enough to make either team, they get what they deserve.

My argument is that any scaling that vastly benefits "ridiculous" teams and vastly penalizes "stupid" teams in a permanent way must not get implemented. If scaling were based on your proposed numbers and I had the option to field that ridiculous team, I'd do it without blinking an eye. And I'm sure I'd not be the only one.

You might be surprised what people actually do to get around scaling. There are quite a lot of players that don't level their chars because they are afraid they will get penalized for it. When scaling initially hit players even sold part of their roster, to such an extent that IceIX felt compelled to post that this won't help. I'd hesitate to call this stupid, obviously scaling is perceived as such an ultimately bad game mechanic that even seriously maiming yourself is considered an option, on the mere hope it will bring some relief.

For what it's worth, I also intentionally don't level some of my chars, because I'm convinced scaling will make me pay dearly for it. I placed first in about every PvE event the past few months and still my scaling is the lowest within my whole alliance. Maybe I just got lucky repeatedly, I'm not willing to risk it though unless all the details of scaling get disclosed.

Unknown

Cryptobrancus wrote:

iPulzzz wrote:

There was the simulator event.. Please do it again, devs

I guess they listened

Not to the part about how hood it was to have different boss every round -- this time we seem to get stuck with Falcon for all the 3x2.5 days making a huge dent in the fun part.

Unknown

Riggy wrote:

As to the actual question, it has been stated before:

Demiurge_Will wrote:

Scaling missions individually doesn't do enough to satisfy the design goals of this feature: making our events maximally interesting for the widest possible subset of players and reducing grind. We couldn't do rewards like the 5000 Iso-8 challenge mission reward in this event in a system like that. It would either be too trivial for top-end players to get or they would be the only people that had any hope of getting it.

Well, I haven't seen any reduction in grinding. Quite the opposite, the game has become even more time consuming by requiring tedious repetition to manage PvE scaling. I also haven't seen that those level 395 walls serve to make events maximally interesting for the widest possible subset of players, barring some really creative interpretation for "interesting". Most importantly I haven't seen that fabled 5000 Iso-8 challenge mission for a long time now, same as the 8000 Iso-8 progression reward, as far as carrots go they just lasted for a single trip. This really begs the question, why do we still have to endure scaling after all the supposedly positive aspects have been voided?

Unknown

Moghwyn wrote:

Riggy wrote:

As to the actual question, it has been stated before:

Demiurge_Will wrote:

Scaling missions individually doesn't do enough to satisfy the design goals of this feature: making our events maximally interesting for the widest possible subset of players and reducing grind. We couldn't do rewards like the 5000 Iso-8 challenge mission reward in this event in a system like that. It would either be too trivial for top-end players to get or they would be the only people that had any hope of getting it.

Well, I haven't seen any reduction in grinding. Quite the opposite, the game has become even more time consuming by requiring tedious repetition to manage PvE scaling. I also haven't seen that those level 395 walls serve to make events maximally interesting for the widest possible subset of players, barring some really creative interpretation for "interesting". Most importantly I haven't seen that fabled 5000 Iso-8 challenge mission for a long time now, same as the 8000 Iso-8 progression reward, as far as carrots go they just lasted for a single trip. This really begs the question, why do we still have to endure scaling after all the supposedly positive aspects have been voided?

I'm not necessarily defending or promoting this mechanic, just spreading the information that's available. As my sig says, I'm just a random guy here, same as you (presumably) - I have zero insider knowledge or association with D3. I have, however, been known to play devil's advocate just to promote discussion though. So with that in mind...

The people who encounter the scaling aren't the ones who are intended to benefit from the scaling (obvious, but I'm stating it for clarity). A person with a 1* roster will see less effect from scaling than a person with a 3* roster (typically - there are abnormalities for those who boost or over-perform compared to their peers). The rationale behind this is that scaling slows a person down enough that they will not blow the competition away. Rarely in a PvE event do we see true blowouts on the leaderboard. Scaling (in part; rubberbanding as well) helps with that by slowing down the people best able to compete.

And there has been a reduction in grinding. I think the definition for grinding has gotten watered down a bit since the game came out. (I have the perspective of a Disgaea player, so this game has zero grinding by comparison, but that's beside the point. ). The trend towards optimal playtimes has meant that top players (meaning those toppling the leaderboards) play a given event super-hardcore for about an hour or two per day. That is hardly grinding. Compare that to earlier events where the leaders were spending upwards of 4-5 hours per day grinding and building up massive leads (leads that were impossible to overcome - events were essentially decided days before the end).

So I guess I'm not trying to make any real points, just providing more perspective in an attempt to create more discussion. I guess I fail at playing devil's advocate today. Must be because I did the math on my personal scaling and realized I'm going to have to bust out my spidey pretty early if I want to make top 10 this event. =P