Theorycrafters don't want to revel in victory: why 1.1 was the best time in PoE

"

MatrixFactor wrote:

I believe there are three ways of getting content to be reused that were employed in 1.1:

1) RNG-gating. High tier maps were still reasonably rare, sacrifice parts were rare, and mortal fragments were rare as well. To do the new content you had to repeat content. As a result each Atziri run would take 2-3 hours if you grinded the sacrifice fragments yourself. Similarly, powerful uniques like Crown of Eyes, Shavronne's Wrappings and Mjolner had small drop rates but enabled powerful builds. By setting the drop rates low enough, yet rewards high enough, GGG is able to make players redo the same content multiple times without them feeling like they're completely wasting their time. Of course this is the cheapest method for content reusability, and one players see through the easiest.

2) Difficulty-gating. At the same time that Sacrifice and Mortal Fragments were rare, the Atziri fights themselves were challenging on build design and fight mechanics levels. If a player had a 25% success rate to kill Atziri, the 2-3 hours of initial sacrifice fragment grinding would get multiplied by 4x to stretch out the initial content to 8-12 hours. Adding the uber atziri layer meant that everything would be multiplied by the mortal hope drop rate and the probability of success to kill uber atziri. As a result by clever design decisions for the atziri fights and number tweaking GGG was able to generate 200-300 hours of content out of something that would have been only 3 hours of content if it was implemented like Core Malachai. (That's how long I would estimate it took me to figure out how to beat him).

3) Combinatorial optimization-gating. Combinatorics is the study of the number of ways to choose different objects to go together. Each time GGG releases a item or gem, we the players have to consider the other possible skills, passives, supports, and items it will go with. Two good examples of this from the 1.1 era are Mjolner (released in 1.1.2) and Summon Raging Spirit (released in 1.0.6, 1 month before 1.1). When these were first released, people didn't quite know how to use them. Tom94's Rainbownuke build (published 1 month after Mjolner was released) was a revelation, as well as PepprmintButler's HeadBangGang (only published in 1.1.5, 5 months after the gem was released!). Gems and items take a certain amount of work to develop, but it takes us longer to figure out how to make use of them than it does GGG to develop if it isn't obvious how to use them. I believe here the development time to content time ratio is much better than in other cases. Meaning, if it took GGG a week to develop Mjolner, it took the playerbase as a whole about a month to figure out how to use it. This is much better than the 1month:1hour ratio we got from Act 4. RNG-gating makes this combinatorial optimization problem more difficult and as a result makes it take longer until the community can solve how to use new items and gems. The caution is that if theorycrafting takes effort (regrets, leveling time, purchasing items), it should be rewarding if you do it properly. If new items are released that are mediocre, people are discouraged from trying them, and their development time is wasted.

I like this part of the post, but I feel it doesn't quite finish describing the issues.

There is not knowing an amazing skill is amazing, but there is also not knowing a mediocre skill is mediocre. The funny thing is that, if you asked me in the 1.1 era which recently released skill was most OP, I would have said Molten Strike. This was essentially the same error as saying the recently released Hearthstone card Ice Rager is OP, simply because it is strictly superior to a previous card. So it's not just "is SRS strong?" it is also "is Molten Strike strong?" etc.

Really, this all comes down to balance. Not the fake balance that some advocate, which tries to grey out all skills until everything is the same, but real balance, which is about making obviously different things feel and appear as equal as possible. See, fake balance is about squashing variety, making everything the same, while real balance is about promoting variety.

But enough about fake balance. The point is that balance should be about making unequal things - that is, various choices - appear as equal as possible. I say "appear" because, well, they're not. At the end of the day, one particular combination is the best one, there aren't really ties so much as there margins of error caused by uncertainty... and as time goes on, these margins shrink. Thus, balance is a little bit less about making two unequal things equal - their differences need to be highlighted - but instead about concealing from the player/community what makes one superior to the other. Balance is not forever, it naturally decays over time, and what might seem like a close call and good balance today might be an obvious choice with a clear winner a year from now.

However, I think the entropy of balance is not really a third point on how to stretch content, but instead perhaps what you wanted to base the entire post around. After all, "theorycrafters don't want to revel in victory" seems a more directly connected to balance entropy than to stretching out content. Instead, content gating as you listed it impacts balance entropy by lengthening iteration times for direct experimentation (that is, actually playing builds to figure out what is more OP or not). In other words, assuming builds are balanced well enough that the correct answers are not obvious at the theorycrafting stage, the time needed to test it multiple times on the desired content has a huge factor in how quickly overpowered and underpowered things are realized by testers and by the community.

Basically, I think the game remains fun as long as it is fun on the theorycrafting level. I think it loses its lustre once you no longer need to play builds in order to have a pretty good idea of how well they'll perform.

Also, as far as budget-limited PvP goes, I believe the best way to do this would be as follows:
1. Competing players begin a PvE race. The purpose of this race is to give players the required items and levels for the PvP portion.
2. After the race stops, the players have 2 minutes to arrange their equipment before it is locked in stone (you can PvE as hard as you want before going into PvP mode).
3. The remainder of the event is a 1v1 PvP tournament with the race characters created this way.

I feel like there is some not quite true things in your main post. I agree with a lot of it, but a few things are ringing false.

The statement "Tom94's Rainbownuke build (published 1 month after Mjolner was released) was a revelation" is not really something I would agree with.

The power of Mjollnir, and of Shavronne, and other uniques was something that I remember discussing as soon as the unique was known. Yes, the builds might not have been created because people didn't have the item, but the theorycrafting was there and known straight away.

Saying that people didn't know of Mjollnir's power until 1 month after it was released just feels wrong. It might not have been min-maxed, but its potential was well known in advance of it ever being used.


A similar thing for SRS, I remember when it was released making a build that was dual totem SRS. Once more, maybe it wasn't the most min-maxed SRS build like there is now, but the power of SRS was realised very early and saying that it took 5 months to discover is not necessarily true.


I don't disagree that having combinations that are not planned can be powerful, there are new potential builds that come out from time to time that are unexpected, such as using ES regen tanking on the dyness tank build. But the idea that people don't know how to use it until a popular build is made isn't really true, it is rather that people don't make build guides all the time.



Sidenote, people hate RNG-gating a lot

"

Real_Wolf wrote:

I feel like there is some not quite true things in your main post. I agree with a lot of it, but a few things are ringing false.

The statement "Tom94's Rainbownuke build (published 1 month after Mjolner was released) was a revelation" is not really something I would agree with.

The power of Mjollnir, and of Shavronne, and other uniques was something that I remember discussing as soon as the unique was known. Yes, the builds might not have been created because people didn't have the item, but the theorycrafting was there and known straight away.

Saying that people didn't know of Mjollnir's power until 1 month after it was released just feels wrong. It might not have been min-maxed, but its potential was well known in advance of it ever being used.


A similar thing for SRS, I remember when it was released making a build that was dual totem SRS. Once more, maybe it wasn't the most min-maxed SRS build like there is now, but the power of SRS was realised very early and saying that it took 5 months to discover is not necessarily true.


I don't disagree that having combinations that are not planned can be powerful, there are new potential builds that come out from time to time that are unexpected, such as using ES regen tanking on the dyness tank build. But the idea that people don't know how to use it until a popular build is made isn't really true, it is rather that people don't make build guides all the time.



Sidenote, people hate RNG-gating a lot

People hate RNG gating with huge ranges; if the ranges in Path of Exile were toned down abit people wouldn't rage so hard and yell at the game.

Sorry to tell you, sir, but you most likely just don't really matter. GGG stated that the amount of people that reaches maps is actually quite small, the amount of people reaching Atziri is even smaller, the amount of people who have "done everything" (Uber, Core map, L100) is probably just very, very small - and even for those, most of those people probably just try a bunch of different builds. You are just a small minority within a small minority. Doesn't mean your feedback is bad, it is just probably not very relevant to the overall state of PoE.

"

MatrixFactor wrote:

2) Difficulty-gating. At the same time that Sacrifice and Mortal Fragments were rare, the Atziri fights themselves were challenging on build design and fight mechanics levels. If a player had a 25% success rate to kill Atziri, the 2-3 hours of initial sacrifice fragment grinding would get multiplied by 4x to stretch out the initial content to 8-12 hours. Adding the uber atziri layer meant that everything would be multiplied by the mortal hope drop rate and the probability of success to kill uber atziri. As a result by clever design decisions for the atziri fights and number tweaking GGG was able to generate 200-300 hours of content out of something that would have been only 3 hours of content if it was implemented like Core Malachai. (That's how long I would estimate it took me to figure out how to beat him).

Atziri design is good in principle, but for a casual player just incredibly annoying. I can full clear Atziri in less than 10 minutes (which doesn't mean I am very fast) rather easily, but Mortal Hope is just way too freaking rare, so I cannot really attempt to do Uber. At least I have tried, twice (and failed). It is not 200-300 hours of content, it is 200-300 hours of annoyance. And if you fail on her, you have to go again. 200-300 hours for a casual player is several months to a year.

Core Malachai - no idea how you did that in 3 hours, I played quite a bit since 2.0 but still do not have a Core map. Dropped two 80+ maps so far from maps (two more from Uber). Getting to Uber Atziri is still easier than getting to Core Malachai, because of stupid RNG/wealth gating.

"

ScrotieMcB wrote:

However, I think the entropy of balance is not really a third point on how to stretch content, but instead perhaps what you wanted to base the entire post around. After all, "theorycrafters don't want to revel in victory" seems a more directly connected to balance entropy than to stretching out content. Instead, content gating as you listed it impacts balance entropy by lengthening iteration times for direct experimentation (that is, actually playing builds to figure out what is more OP or not). In other words, assuming builds are balanced well enough that the correct answers are not obvious at the theorycrafting stage, the time needed to test it multiple times on the desired content has a huge factor in how quickly overpowered and underpowered things are realized by testers and by the community.

Basically, I think the game remains fun as long as it is fun on the theorycrafting level. I think it loses its lustre once you no longer need to play builds in order to have a pretty good idea of how well they'll perform.

Also, as far as budget-limited PvP goes, I believe the best way to do this would be as follows:
1. Competing players begin a PvE race. The purpose of this race is to give players the required items and levels for the PvP portion.
2. After the race stops, the players have 2 minutes to arrange their equipment before it is locked in stone (you can PvE as hard as you want before going into PvP mode).
3. The remainder of the event is a 1v1 PvP tournament with the race characters created this way.

Well I think it's better to have mechanics that need to be played to be understood. There's more of a risk/reward there. You sink your time into trying out some new mechanic and it turns out great or it turns out terrible. If everything that's released is a reskin of a existing mechanic, it doesn't add a combinatorial challenge because we can apply the solutions we already have from the original versions of those mechanics.

About your PvP suggestion, I think there are two problems that arise by having it start with a PvE race. First, some people don't like to race. Second, a PvE race favors some builds over others. AoE casters specifically. You could argue that this will be balanced that these casters won't do well in PvP, and I agree that totem builds will probably do poorly, but there are other caster builds that would do well. For instance in cut-throat I remember doing really well with freeze pulse+spark (when it shotgunned). In general the problem with low budget stuff is that it hurts attack builds more, that's where I see the balance issue appearing.

"

Real_Wolf wrote:

Saying that people didn't know of Mjollnir's power until 1 month after it was released just feels wrong. It might not have been min-maxed, but its potential was well known in advance of it ever being used.

A similar thing for SRS, I remember when it was released making a build that was dual totem SRS. Once more, maybe it wasn't the most min-maxed SRS build like there is now, but the power of SRS was realised very early and saying that it took 5 months to discover is not necessarily true.
But the idea that people don't know how to use it until a popular build is made isn't really true, it is rather that people don't make build guides all the time.

Well my entire point is about min-maxability. If you can figure out a way to use Mjolner or SRS but you don't min-max it enough to make it OP, that's not relevant to how min-maxable the game is.

I was also one of the people who tried dual totem SRS in 1.1, and going from dual totem to a +3 staff was a completely different game. All of a sudden you could do 77-78 bosses with ease, which you definitely could not do with totems. Also before I posted my SRS calculators people were using all kinds of wrong support gems, curses, passives, and auras for SRS. You couldn't really call SRS solved until 1.1.5.

Your last point about discovery is kind of irrelevant. Sure people can claim they had builds figured out before the first guide was posted, but until the community catches up that combinatorial problem is still "open" for theorycrafters to try and solve. The point is someone needs to tell you that Mjolner is OP if you do x, y, and z. Until then it's still up to you to discover what x, y, and z are.