The leveling and XP curve. What's wrong and the only way to fix it.

[Matt516]

Matt's point is that fixing any individual instance of one source of xp or another - the Endless Paths, the bounties, whatever - won't fix the underlying problem, and it's only going to get worse as we progress beyond this first game.

 

Yup - that's what I keep trying to get across, anyway. They can band-aid it for this game with rebalancing of certain things (Endless Paths and Bounties being the big ones) and it'll work fine.

 

But if the series continues, and continues to use the same leveling curve, one of two things must happen: either the level gap between completionists and non-completionists continues to grow (making proper encounter balance harder and harder, and potentially even forcing them to have stuff like a minimum level when starting PoE 2 or 3)... or sidequest XP will get smaller and smaller until it's a pittance by PoE 3.

 

Honestly, if you use a level cap for each game, and then have a minimum level when starting one of the sequels (with automatic scaling up if you import a character who is lower level), you can sort of partially band-aid indefinitely - though the level gap between the two groups will continue to rise simply by virtue of it being possible for people to start at different levels. It may be that that's what Obsidian intends to do. It'll make encounter balancing harder than it has to be, but I guess it does somewhat avoid the sequence-breaking exploits that are the only real weakness of the exponential level curve. I dunno. If I were Obsidian, I know what I'd do - but I'm an efficiency and balance junkie, so maybe they'd rather just make things harder for themselves/make the game less balanced.

 

EDIT: I guess you could just have the minimum level for PoE 2 and 3 be whatever the level cap was for the previous game. That at least ensures everyone starts on the same foot, and means that each game is roughly the same difficulty to balance since you only have to consider possible level growth throughout that game alone.

 

Obsidian devs, if any of you are reading - please make that happen for PoE 2 and 3 if the quadratic XP system is indeed set in stone (as I suspect it is). It's the only way your game doesn't get harder and harder to balance over time.

 

Huh. Now that I've thought of that, that is an alternate solution that would actually act as an infinite band-aid (instead of a partial band-aid). So.. yup. If they keep quadratic level curve, they should make imported characters in the sequels automatically start at the level cap from the previous game. That's the only way you keep things from spiraling out of control.

Edited April 9, 2015 by Matt516

[Serdan]

Matt, are there actually any problems with my suggestion?

[sparklecat]

I think at this point the best solution may be to apply band-aid solutions for this game, and consider an overhaul with the expansion or sequel.

[Matt516]

Matt, are there actually any problems with my suggestion?

 

Doing away with XP entirely and just handing out levels? Not... reeeaaalllyyy..... But it'd be an even harder sell than changing the leveling curve. And does come with some additional problems (such as no granularity whatsoever - there's a lot more quests than levels, so not all quests would have a level reward. And if you start handing out fractions of levels you're basically just back to XP).

 

It's not a system I'm opposed to, persay... I enjoyed Shadowrun: Dragonfall a ton, and it used a similar system. But that game actually didn't have levels, just Karma (skill points, attribute points, and abilities all rolled into one). So while there's nothing incredibly wrong with it or anything from a conceptual perspective... I don't think it'd work with PoE at all.

[View619]

Cool Matt, I'll take your word for it and stick to looking around source code for curiosity/knowledge purposes. Also, I want to see what the devs plan on doing regarding difficulty and balance in the patches to come; maybe modifications won't even be necessary by the end of it.

Edited April 9, 2015 by View619

[Matt516]

Cool Matt, I'll take your word for it and stick to looking around source code for curiosity/knowledge purposes. Also, I want to see what the devs plan on doing regarding difficulty and balance in the patches to come; maybe modifications won't even be necessary by the end of it.

 

Yeah - as I realized a few posts ago, they can ensure that the game doesn't get progressively harder to balance in the sequels by simply autoleveling all imported characters to whatever the level cap was from the previous game. That would (mostly) do away with my main issue with the quadratic level curve, which is how fundamentally unsustainable it is in the long run.

 

If they do that... I'll be happy. If they don't... I might have to make one character for actually doing all the sidequests and one for actually being challenged in the main quest. Because there's no way they can balance sequels properly for completionists if they let you start out a sequel already higher level than a non-completionist.

Edited April 9, 2015 by Matt516

[Serdan]

 

Matt, are there actually any problems with my suggestion?

 

Doing away with XP entirely and just handing out levels? Not... reeeaaalllyyy..... But it'd be an even harder sell than changing the leveling curve. And does come with some additional problems (such as no granularity whatsoever - there's a lot more quests than levels, so not all quests would have a level reward. And if you start handing out fractions of levels you're basically just back to XP).

 

It's not a system I'm opposed to, persay... I enjoyed Shadowrun: Dragonfall a ton, and it used a similar system. But that game actually didn't have levels, just Karma (skill points, attribute points, and abilities all rolled into one). So while there's nothing incredibly wrong with it or anything from a conceptual perspective... I don't think it'd work with PoE at all.

 

You'd hand out levels based on % content completion tuned to land the players in certain level ranges at certain points of the game (if there are 10 quests and you want the player to progress 5 levels after completing all of them, then you hand out a level for every other quest).

I just completely fail to see how that wouldn't work for PoE.

[Daemonjax]

I found the hex code location to modify the experience per level formula.

 

I'll post simple to follow instructions in a few minutes.

 

EDIT:

 

 

 

 

 

1. Go to your game folder: ...\Pillars of Eternity\PillarsOfEternity_Data\Managed

2. make a backup copy of Assembly-CSharp.dll

3. open then original Assembly-CSharp.dll in your favorite hex editor (I like HxD found here: http://mh-nexus.de/en/downloads.php?product=HxD)

4. Search for the following hex code:

 

6F AC 02 00 0A 12 02 28 AD 02 00 0A 3A B1 FF FF FF DD 0C 00 00 00 08 8C 7E 00 00 1B 6F 1A 00 00 0A DC 2A 00 41 1C 00 00 02 00 00 00 29 00 00 00 59 00 00 00 82 00 00 00 0C 00 00 00 00 00 00 00 32 02 02 17 58 5A 20 F4 01 00 00 5A 2A 00 00 00 32 02 17 59 02 5A 20 

 

5. The hex code immediately following the above should look like:

 

F4 01 00 00 5A 2A 00 00 00 13 30 07 00 14 00 00 00 48 00 00 11 02 6F 12 00 00 2B 0A 06 6F 57 07 00 06 03 28 08 00 00 0A 2A 13 30 07 00 27 00 00 00 00 00 00 00 02 28 04 00 00 0A 39 0C 00 00 00 02 6F 0C 00 00 0A 28 DB 06 00 06 2A 72 B9 1F 00 70 28 42 00 00 0A 72 A1 1F

 

Note1: You're only interested in changing the underlined hex code above.

Note2: Alternatively, you can search for and find the FOURTH instance of F4 01 00 00 (starting from the top of the file).

 

6. Change the F4 01 00 00 (which is a little-endian 32-bit integer for 500) to whatever you want (use a programming calculator -- the one that comes with windows works fine for this)

 

For example, changing it to E8 03 00 00 (which is the integer 1000 in little-endian) will double the xp required per level.

 

 

 

Notes:  

 

1) I use IE mod, which replaces Assembly-CSharp.dll, so this may or may not work for you if not using the IE mod.

2) I'm on Windows, so this may or may not work for you on linux and/or mac.

3) The change isn't retro-active -- it only affects experience needed for your next level.  A full restart would be required if you want to see how this affects gameplay.

Edited April 9, 2015 by Daemonjax

[anameforobsidian]

 

This problem is exacerbated by Od Nua.  I was level 8 or 9 before I finished the first chapter.  Od Nua felt like a whole separate game inside the main thing, and its actually pretty fun as a dungeon runner.  You could rely on heavy handed solutions, or just accept that there's a huge amount of sidequests and some players will be overleveled by the middle.  Hell, it definitely happened in BG2, where I hit max somewhere around underdark, and earlier than that with Watcher's keep.

 

Durlag's Tower did similar things to BG, where I was maxed by the end of the Third chapter.

 

Matt's point is that fixing any individual instance of one source of xp or another - the Endless Paths, the bounties, whatever - won't fix the underlying problem, and it's only going to get worse as we progress beyond this first game.

 

 

eta: as he just said 2 minutes ago!

 

 

On the other hand, if a greater portion of the game is extra content than the critical path, then there will always be severe imbalances in XP.  Players should expect to receive commensurate xp and loot rewards for time spent, that's how rewards work.  The more extra content there is, the harder it is to balance in general without level-scaling (which Baldur's Gate did!).  In that respect, changing the math trends toward making sidequests less meaningful or not really fixing the problem.

 

I thinks it's a mostly intractable problem.  Most of the ways around it aren't pretty.  Make level effectiveness slowly top out ala fallout.  Level scaling like Skyrim.  

[cirdanx]

I have to be honest and don´t "stone" me for it This sound exactly like when BG1 AND 2 came out, there were always people who thought xp gain was to fast (especially in 2) and that it ruined the game. Well i can´t agree.

 

I do agree that it feels off sometimes, being someone who wants to finish everything in a zone i find myself often outleveling content, but i can still turn up the difficultiy if i want to, being max lvl before the end is also common in the old games if you did everything. So again i don´t see a problem.

 

Maybe its a bit to fast. But we, and i expect that everyone here is now an professional DnD player (which i´m not ) have to consider that there are players who get their ass kicked on easy with this xp curve. Obsidian had to find a middle ground. So far, i THINk they have done a good job. Why? Because the super hard stuff will come in form in mods, designed by and for people who like that, just like the Black Isle games, you can´t release a game in this state, 90% of the people wouldn´t play after the first hours, that was true back than and now.

 

THe only thing i would say is, that they should have made more levels and spread the goodies (likes spells) more out...thats my only critic really. (so far)

[Matt516]

On the other hand, if a greater portion of the game is extra content than the critical path, then there will always be severe imbalances in XP.  

 

Yes, but the imbalance in XP isn't what matters - only the imbalance in level matters. With the same exponential formula the BG games used, you could have three times the amount of sidequest XP as story XP and still have no more than 2 levels difference between a total completionist and a completely critical path player. Ever. The reason severe imbalances in XP translates to severe imbalances in level is the quadratic level curve. Exponential simply doesn't have that problem. It allows for severe imbalances in XP to only translate to very slight differences in power. That's why I suggested it.

 

You'd hand out levels based on % content completion tuned to land the players in certain level ranges at certain points of the game (if there are 10 quests and you want the player to progress 5 levels after completing all of them, then you hand out a level for every other quest).

I just completely fail to see how that wouldn't work for PoE.

 

Then you're just replacing XP with % content completion. Looks different, but is for all intents and purposes almost the same thing. With tons of sidequests with differing levels of difficulty, you'd have to weigh each one differently... and then you're basically just back to XP. Not to mention that doesn't solve issue with completionists overleveling unless you just don't give levels for sidequests at all. Which is something you can do with an XP system anyway.

 

Like I said though - it certainly could work. But as I mentioned above it doesn't really offer much beyond an XP system and still runs into a lot of the same challenges. It's also very very different from the IE games, which would piss a lot of people off. And from where I'm sitting I see no chance that it will happen - I prefer to make suggestions I think actually have a chance of being implemented, that's all.

Edited April 9, 2015 by Matt516

[ibanix]

These arguments also make one particular assumption: That one should be able to complete the game without doing any side-quests. I would argue that players who don't do any of the side-quests should find the game harder to complete. I don't really have a reason other than to say that I think it makes sense for characters who do more to be more experienced and therefore better able to handle the end-game.

 

I also think that it would have been better if challenges were scaled to the experience level of the entire party; buffing monsters as party experience increases. 

[scrotiemcb]

I apologize for not reading the entire thread at the time of writing this.

 

I applaud OP for his analysis of the problem. However, in presenting his solution, there is obvious hyperbole in saying there is "literally only one way" to skin the cat. Hint: always at least two.

 

For those who don't know, the XP needed per level is a direct homage to pen-and-paper D&D. Not only is it the same general scheme, it's the exact same numbers.

 

So with this awareness, we can come to one of two conclusions: either

• D&D has a crappy leveling system despite numerous times rewrites spanning decades, or
• the D&D folks came up with an additional XP mechanic to modify the situation

Their additional mechanic is a little... strange. Link. Although giving the precise formula from this table eludes me, the point is that, as your party outlevels certain encounters, the amount of XP gained for that encounter is reduced. Being one level higher could easily mean 20% less XP received, and outleveling it too much could mean a trivial amount of XP, or even none.

 

This creates a very complex system which performs roughly the same mathematical effect as the exponential XP system OP endorses, but also seems to better encourage players to seek out level-appropriate challenges rather than attempting to grind away every last bit of content the game has to offer in a single playthrough. Whether this last bit is a desireable objective or not in a CRPG is an important question, but I feel it's obviously important in a P&P RPG where players have more latitude.

[Ohioastro]

What's missing from this discussion is actual data, and the answer is knowable. If you want the game changed document how much exp there is, rather than simply asserting that there is too much.

 

I think that this won't do anything for the people who find the game too easy, because that's a product of deeper factors. (Speaking of which...I'm also not seeing any objective target for what acceptable difficulty is.

[Daemonjax]

Just change the experience requirement to whatever you want it to be.

 

I posted instructions on the previous page.

[Matt516]

I apologize for not reading the entire thread at the time of writing this.

 

I applaud OP for his analysis of the problem. However, in presenting his solution, there is obvious hyperbole in saying there is "literally only one way" to skin the cat. Hint: always at least two.

 

For those who don't know, the XP needed per level is a direct homage to pen-and-paper D&D. Not only is it the same general scheme, it's the exact same numbers.

 

Read this for a bit more elaboration re: "only way". I was a little hyperbolic in the OP, but given the premises I present in this post (and assuming no radical changes like doing away with XP altogether or adding in a D&D-esque scaling system), an exponential equation is the only way: https://forums.obsidian.net/topic/76649-the-leveling-and-xp-curve-whats-wrong-and-the-only-way-to-fix-it/?p=1648467

 

I should've qualified originally - obviously a completely new mechanic like encounter scaling could help fix the problem. I'm not looking at adding completely new mechanics, just at working within the framework we already have. And within that framework, my statement that an exponential equation is the only solution that achieves the two goals I mentioned in the linked post is correct.

 

But obviously there are all manner of completely alternate systems that could work. An exponential system is the only equation that works within this framework - that of finite XP in the game, each quest having a set amount of XP, etc.

[Matt516]

What's missing from this discussion is actual data, and the answer is knowable. If you want the game changed document how much exp there is, rather than simply asserting that there is too much.

 

I think that this won't do anything for the people who find the game too easy, because that's a product of deeper factors. (Speaking of which...I'm also not seeing any objective target for what acceptable difficulty is.

 

I don't need actual data to know that the equation currently used to calculate the amount of XP for next level is fundamentally incapable of achieving both the design goals I outlined here: https://forums.obsidian.net/topic/76649-the-leveling-and-xp-curve-whats-wrong-and-the-only-way-to-fix-it/?p=1648467 

That's just a consequence (literally) of the laws of logic and mathematics.

 

This is "mathematical possibility" talk, not "specific balance" talk.

Edited April 9, 2015 by Matt516

[Daemonjax]

 

What's missing from this discussion is actual data, and the answer is knowable. If you want the game changed document how much exp there is, rather than simply asserting that there is too much.

 

I think that this won't do anything for the people who find the game too easy, because that's a product of deeper factors. (Speaking of which...I'm also not seeing any objective target for what acceptable difficulty is.

 

I don't need actual data to know that the equation currently used to calculate the amount of XP for next level is fundamentally incapable of achieving both the design goals I outlined here: https://forums.obsidian.net/topic/76649-the-leveling-and-xp-curve-whats-wrong-and-the-only-way-to-fix-it/?p=1648467

That's just a consequence (literally) of the laws of logic and mathematics.

 

This is "mathematical possibility" talk, not "specific balance" talk.

 

 

 

And I want a pony for my birthday.

[anameforobsidian]

By making the curve exponential, you are in essence depowering the side content unless you create a serious power differential with loot.  Maybe players that put in more time should be significantly more powerful.  

 

Also, exponential xp leads to ridiculous situations like needing 1,000,000 xp to reach level twelve (that's killing 50,000 goblins).  It loses serious value as an abstraction.

Edited April 9, 2015 by anameforobsidian

[Ohioastro]

I think that actually knowing how much exp is available is relevant when designing an exp system.

[Ink Blot]

I also think that it would have been better if challenges were scaled to the experience level of the entire party; buffing monsters as party experience increases. 

Have to chime in here with a big "NO, NO, NO". I hate level scaling. I want to be able to enter areas and mop the floor with the mobs sometimes. And I want to be able to enter areas and have my ass thoroughly and completely kicked sometimes. I don't want to enter areas and know that every single time I'm going to be faced with monsters/encounters of exactly my level (or within one or two).

[View619]

I think that actually knowing how much exp is available is relevant when designing an exp system.

 

That's if you're working out specific values. You can work out a formula without knowing the exact number of exp points by making the total some static value x, then applying it.

Edited April 9, 2015 by View619

[Hassat Hunter]

So if 1+1 = 2, the critical path, and side content is +1 = 3... the only way to balance it is rip 3 to 100 instead? Cause logic?

No, I don't see much logic in some of these suggestions at all.

 

Yes, exponential increase is an option. If you seriously want to de-value side-content. Which I don't see has any benefit of being added. Some XP values are off, and trap/lockpick XP is still the worst thing ever but saying stuff like "Make it exponential!" is only going to make working out a proper sollution harder, not easier.

 

There's a reason kids first learn tables from 1 to 10 and only up till 10. Going in and saying they should just after learning 2x2 to go to 3x13 is the exact opposite of making things "easier".

 

EDIT:
Just to clarify a bit more... if all quests in the game give 1000XP and this other one gives 20000XP you can modify the formula for XP required all you want, but you're actually designing a work-around around the bug rather than fixing the bug, and the problem still remains in the game.

 

So, no, I don't see *this* as fix, I see adding exponential formula's for XP as something designed to work around non-intended design. And changing stuff to accomidate bugs is always worse than fixing the reprecencies. Just my 2 cents...

 

EDIT2:

And just to throw about more weird examples (cause I love them so);

Your problem is there's a baby in the bathwater. The foreign object takes much space. Instead of dealing with the foreign object however, you decide to just open the valve and let more water in, and then later on act surprised if the water rises over since no, opening the valve does not make the foreign object (or excessive XP/bug/etc.) dissapear. And you've just soaked the room and gotten all complicated while the answer is so simple and probably staring at your with those baby eyes.

Edited April 9, 2015 by Hassat Hunter

[Matt516]

*snip*

 

Your analogies indicate to me that you don't understand what I'm talking about and are just saying random things in an effort to confuse the problem. That may or may not be true and I'll try not to make assumptions... but I will say this. Please  read what I'm saying, understand it, and counter by pointing out where my math is wrong - or don't. I'm not making statements of opinion here. This: https://forums.obsidian.net/topic/76649-the-leveling-and-xp-curve-whats-wrong-and-the-only-way-to-fix-it/page-3?do=findComment&comment=1648467 is fact. Which is to say - if you agree with the two premises presented in the post above, the conclusion follows mathematically from them. Period.

 

All this talk of de-valuing side content and such has nothing to do with the point I'm making. It's nonsense. Yes, an exponential system would make side content somewhat less relevant to determining your final level in the game - relative to the current system. But in reality, you can adjust the ratios and the coefficients in the exponential equation to make side content as valuable or not valuable as you want. You could come up with coefficients for the exponential equation that created a 10 level difference between the non-completionist and completionist playstyles if you wanted. I'm only proposing the form of the equation - the coefficients are calculated directly from 1) the desired ratio of side XP to story XP and 2) the desired level gap between a completionist and a non-completionist at the end of the game. The importance of side content is entirely tunable with an exponential functional form. And you would know this if you'd actually read what I'm saying. I don't mean to sound rude or dismissive and please forgive me if I do, but that's the truth. Your objections indicate to me that you haven't read and/or understood what I'm actually proposing and are instead knee-jerk reacting to what you think I'm proposing.

 

One of the reasons I'm proposing the exponential system in the first place is that the quadratic (current) system must lead to either side content becoming less valuable or the level gap between those who do and don't complete it growing larger as the series goes on. It must. There's no way around that, it's dictated by the mathematics. The exponential system is in some ways the only way to save side content from irrelevancy. Again, please read the link above. Look at the two premises I put forth. If you disagree with one of those, fine. That means we have different preferences and argument is pointless - we can agree to disagree. If you don't disagree with either of those premises, then know that mathematically an exponential curve is the only way both of them can be satisfied. I cannot stress this enough - I'm not arguing for an exponential curve because I like it, I'm arguing for it because I like these two statements:

 

1) The ratio of optional XP to mandatory XP should be roughly the same no matter where in the game (or eventual series) one is... that is, if in Act I there is 10,000 critical path XP and 10,000 optional XP, and in Act III there is 50,000 critical path XP, there should be about 50,000 optional XP in Act III. If in Act VII (assuming PoE 2) there is 10,000,000 critical path XP, there should be about 10,000,000 optional XP. Specific numbers aren't important, just the concept.

 

2) The level difference between a completionist and non-completionist player should not grow significantly (if at all) with time. That is, if I do every single sidequest and Bobby only does the story quests, and I end up 2 levels ahead of Bobby at the end of PoE, I shouldn't be.. say.. 6 levels ahead of Bobby at the end of PoE 2 if we both follow the same pattern.

 

And because the exponential curve is the only functional form for the leveling curve (that I've seen, anyway) that is capable of satisfying them. The quadratic (current) curve is literally incapable of satisfying either of those two. Fundamentally, mathematically incapable of doing so. That's not my opinion, that's math.

 

Quick addendum: As I've noted in posts since the one I linked, if Obsidian were to autolevel imported characters in PoE 2 and 3 up to the level cap from the previous game, that would also (sort of) solve the problem by allowing them to band-aid it by carefully balancing the quest and side XP. But the issues with the quadratic system would remain.

Edited April 9, 2015 by Matt516

[PsychoBlonde]

 

 

It's a single-player game.  Why does it need to be "balanced"?

 

My only comment on this "problem" is "I don't care, and it doesn't matter".

 

Yet another mindless "the game needs to forcibly prevent people from playing a certain way because I imagine that way as somehow being 'less fun' or 'suboptimal'" complaint.

 

Currently the game "forcibly prevents" players that want to do side quests from enjoying combat because apparently "the optimal" way is to rush main quest, finish the game ASAP and buy the next game.

 

 

I did all the side quests AND quite enjoyed combat.  So it's not the GAME causing this "problem", it's your imagination.