Update to JUPP

Remove all ads!

Now we're getting somewhere. Yes, both would still be at level 1, but I suppose most players would actually care if their characters are "just freshly dinged level a.k.a 1.00" or "almost at the next level a.k.a 1.99", especially if they're trying to get as high as they possibly can.

Terminology again, my bad. Will clarify this part better.

The larger amount of exp my party got is, in fact, exactly caused by the addition of fresh newbies in some key points of the game. Yes, I got a bit less exp during the first half of Chapter One than your six-person party did since my duo leveled up so fast. However, this was more than well compensated in the period from Horde Warrens to at least first half of Chapter Two, which my 2+2 party actually reached at LOWER level than your six-man party did.

The difference is that I selected the parts where I was penalized for exp wisely. Even if you did the first half of Chapter One with average level 3, you'd still not get all that much exp since you only meet weak monsters. However, getting to the inside of the Horde Fortress and especially Ice Temple with a lower-than-usual level party multiplies the exp from the high CR mobs nicely.

Of course it depends. And waiting all the way to the latter parts of Chapter Four to add your last two members is rather extreme. But it will lead to better overall exp total, just as squatting and/or muling would. You will do just fine without it, too.

Besides, I didn't optimize my party so that it would peak at max effectiveness or levels for the normal mode's final battle. If I were to do that, I would have made sure to have all of my party present for the exp feast called Chapter Two.

Noted. Will keep my eyes wide, wide open to double-check the Knight's CR everytime I meet one the next time I am there.

 

You're getting to wrong conclusions. You actually do get more XP after adding a new character to the party, but that's because you've to catch up.

Assume you're in an area fighting only CR8 monsters. You'll get stuck at level 16, no matter if you have a party of 5 or 6. The party of 5 gets stuck with about 120,000 less total XP. If you add a 6th character to the party of 5 you'll continue gaining XP, but only until you've evened out the additional total levels the other party has.

So adding party members later doesn't yield more total levels, you only get XP faster until you reach the levels you'd usually have if you started a party of 6. Therefore there are no key points in the game for adding new party members.

The only reason you might have more total XP is that the XP requirements are higher for higher levels, but this won't yield more party levels.

 

I see your point, but you'd have to be patently stupid to not realize you're not gaining any more exp with the smaller party - which you WILL reach earlier than the full party does. Of course you'd ultimately get stuck at level 16 fighting against CR8 monsters, no matter if you're solo or six-man. If you reach that cut-off point and have the option of adding new party members and still decide NOT to add them - well, it's your mistake, not the fault of the system itself.

You're still looking at it from the perspective of maximizing your party's total levels. True, you won't get more total levels by adding new members along the way. I never contested that statement. Only level-squatting and muling will actually net you extra levels as you're artificially widening the gap between your actual level and the level you have gained the experience for.

The extra experience points available by adding new members, however, grant the possibility of shuffling the party member's levels between them to better suit their needs.

For example, consider the following two party members:
Human Sorcerer(17) - pure caster
Aasimar Cleric(16) - battlecleric

The sorcerer still hasn't reached 9th level spells, while the Cleric doesn't really need that many 8th level spells (the party's OTHER cleric can do that) and gained the 3rd attack per combat round already at level 15. So let's trade one level from the Cleric to the Sorcerer. In other words, introduce the Sorcerer to the party a couple of areas earlier than the Cleric.

Now we have:
Human Sorcerer(18) - pure caster
Aasimar Cleric(15) - battlecleric

Sure, the Cleric is a smidgeon weaker now, but she still retains her 8th level spells and 3 attacks per combat round. However, the Sorcerer can now cast 9th level spells, being much, much stronger than he'd be at level 17. Of course, if you were to do it the other way around - shuffling levels from the Sorcerer to the Cleric - you'd actually lose power in the exchange instead.

Anyhow, thanks for the sparring aid, kmonster. Now I know that I actually need to clarify why anyone could be interested in trying this approach, instead of just assuming that'd be obvious.

 

You always need to be lower than average level to get more XP. But since the whole point of gaining XP is to level up, I find the whole issue counterproductive. Sure, it's neat to have a wizard that can cast level 9 spells before the end, but I can't seem to get excited about it. My party does fine, even with less available spells.
A straight 6 member party from the start probably gets the most XP out of the game anyway. They sure get more total levels than a ragtag group of characters added randomly.
And to diverge slightly from powergaming, I must also remark that a party with such varied levels is asymmetrical and ugly. I usually try my best to keep them at the same levels, by moving the ECL races up front, giving-the-tome-to-the-gnome, etc.

 

That's pretty much the gist of it, yes. You can't really cheese your exp as much as you could in BG2, at least not without it crippling your exp gain rate from there on anyway. That's why I've labeled any attempts of going around this limitation as various degrees of cheese in my guide now. Yes, including the add-members-as-you-go procedure discussed above.

Here's the important thing though - there's already a lot of guides out there that get the things done "just fine". I'm in no way interested in creating yet another "just fine" guide. "Just fine" isn't good enough for me. It's either the most optimal there is or not in my guide, take your pick.

You're wrong on the XP count, I have the proof written above. You're right on the amount of levels, though, as discussed above.

In other words, you should ignore the part of the guide that describes this trick. None of the JUPP parties will assume you do any of it, in any form. Unlike, say, the old Arcane Party's cleric/druid that pretty much stinks unless you do heavy-duty level squatting with her.

Edit: I'm close to done with the first three major chapters (intro, general party advice and the creation of your own UPP), anyone so interested can drop me a PM and I'll email the guide for proof reading. I'm still missing the example parties part.

 

The 2.00 version of the guide is complete, I submitted it to Sorcerer's Place just a while ago. Size of the text file went from 259k to 371k to give you an idea of how much new stuff was added. Besides that, there's not many stones left unturned there - EVERYTHING has been double-checked for sanity and updated accordingly.

Waiting for comments and the inevitable flames.

 

This looks as if you were comparing apples with oranges, a party with many ECL races to a party without ECL.
Since your party consisted of 2 humans, 2 drow and 2 deep gnomes, you should compare it to a party with the same ECLs. Such a party would be level 17-17-16-16-15-15 (the humans 5,000 XP short of reaching level 18, we assume everyone gained the same XP) when stopping to gain XP from standard CR 8 enemies with a total of 6*148,000 = 888,000 XP.
If someone (doesn't matter who) did read the book for 10,000 extra XP after you bought it, you'll stop at 898,000 XP.
And you still get some XP from the tougher chapter 6 monsters and quests after reaching an average party level of 16 ...


The little extra XP gain you got wasn't caused by "timing" btw, only by the fact that you need more XP for levelling up to higher levels.

 

It's up now, I'll have comments on it "soon".

Jukka, did you try this?

 

The comparison was mostly to show that there's a difference between being at a certain average level and having a certain amount of total experience points. While they're related, they're not the same thing.

As an extreme example, 3 guys at level 30 and 3 guys at level 2 would have a hair over 1,3 mil exp with zero ECL and over 1,6 mil exp with +3 ECL (all deep gnomes), while still having "just freshly dinged average level 16".

The point I'm showing is that adding new members along the way does not jeopardize your chances of attaining just about the same total levels, effectively allowing one to "trade levels" between party members to where they give you most bang for the buck.

By the way, I actually misquoted the race on that level 20 "human". He was, in fact, an Aasimar with +1 ECL. No wonder the exp values didn't seem to match. In other words, a human in that spot would have been level 21 instead. Oh well...

Well, considering that I actually did a select few parts of the journey at a HIGHER average level than a normal six-man party would have, I'd like to hear your theory how I ended up with more total exp, then? Other than compensating in some areas by being (shortly) waaaay lower level than I should have been?

Edit: JT, didn't try that. Don't think that has anything to do with the way SR is calculated rather than some internal game timer going nuts when trying to resolve that many simultaneous AoE effects, skipping some calculations on the way. Kinda how the Chain Lightning in Diablo I started to clip and skip monsters when there was too much stuff on the screen simultaneously? I can try that later and report, though.

Edit 2: So I tried it. Cast 9 skulltraps on a spot. You can only put four in a given tile, it seems - i had two groups of four traps and a loner though I clicked on the exact same spot 9 times. Walked into it on an angle so that they'd all go off simultaneously. Two spell resists, 7 successful evasions, no damage.

I upped the ante and wore a hide armor to "disable" evasion. On a group of four traps, two resisted and two damaged me. HOWEVER, I only got two times 10 points of damage from them, which is kinda funky since my Sorcerer planting them is level 23.

To resolve possible counter overflows, I cheated my decoy to level 20 as wizard and planted six traps. Result: Two resists, four times 10 damage.

Finally I planted all fifteen traps I could on a given day in a line and walked through them one at a time. Result: All resisted, even when wearing the hide armor.

Seems like resolving possible Spell Resistance gets dropped from the calculations in case there's too much stuff going on at the same time.

 

The aasimar<>human misquotation actually doesn't seem to affect anything noteworthy.
The party with the same ECL races where everyone gains the same XP would reach average level 16 6*3,000 XP later, at 906,000 XP total, even further away from 720k-867k. Only a very little part of the XP difference in your comparison is caused by adding characters later, nearly everything is caused by the ECL races.

But I agree that your "exaggeration" makes the point you wanted to show (more XP but not more levels) even more clear.

My point was just that this looks as if you were comparing apples with oranges.

Another unimportant question I asked myself when reading this passage: Is it possible to get a party of 6 to "level 17, halfway in exp to 18" during the game ?

My theory includes that your "adding fresh level 1 characters, to coincide with the parts of the game where there's lots of kill exp to be gained" tactic doesn't deserve the word "timing" for having the most XP at the end since if you added the characters AFTER "the parts of the game where there's lots of kill exp to be gained" your total XP at the end would be even higher.
I didn't doubt that partys with character levels far apart have a little more XP at the end, but no special "timing" is required for this, you can easily end with more XP without following your character adding scheme.

 

Jukka, I think the CL in Diablo2 does the same thing, not sure. I do know that Valkyrie (air-to-air unit) in Starcraft are horrible because of exactly that bug -- the battle heats up and suddenly they aren't firing.

Section 2.1 -- Don't be a tease; give your best example party, or at least directions to the section that has it. Otherwise maybe you should rename the document to "Jukka's List of HOF Powergaming Techniques".

Some sections are much wordier than they need to be. For example, you write three long paragraphs on the Rogue, but only imperfectly manage to convey the Essential Rogue Truths:
- Nearly all the skills in IWD2 range from "useless" to "mildly helpful but completely optional".
- Those same skills only need a few (ten or less) points for maximum effectiveness
- Rogues get a boatload of skill points with just one level
- Backstabbing isn't worth the micromanagement

 

Haven't checked THAT exactly, to be honest. Just took a value that could be a reasonable upper bound for the purposes of the example. Read: Never heard anyone being even close to level 18 average at the end of Normal. I have never played a party without any ECL penalties to have a reasonable lower bound on the total exp either. Not that it really matters in either way, as it's not the thing I'm trying to calculate there.

Now that I said it, seems like I need two extra play-throughs - one with exactly zero ECL and one with exactly +3, doing all the side quests and everything else identically on both, going with a full party from the start without any non-standard tricks. Oh the joys of having a Deep Gnome as my Diplomat, can barely wait!

If nothing else, that would FINALLY settle the dispute on to which degree ECL penalties and the dynamic exp counterbalance each other, down to the last digit of precision. Being a scientist at heart, I loathe doing hand-wawy "probably in the right ballpark, possibly not" stuff.

Eh, wouldn't that be going backwards? Certainly you want to get large exp multipliers (aka lower average level) during the passages where there's a lot to multiply, not AFTER it?

Consider three imaginary areas A, B and C. A has an kill exp value of 10,000, B weighs in at 30,000 and C is 20,000 for a party exactly at the "should be" level (read: playing with full party without tricks). Playing through them normally would bring you 60,000 exp. Assume each level above or below the "should be" level gives uniform 20% experience adjustment.

Now, consider doing area A at one higher level (read: with a somewhat smaller party). You'd only get 8,000 exp there. Nothing strange yet. At the very beginning of B you add a new character to end up at one level below standard (averaged, as you'll quickly level up with the newbies), netting you 36,000 exp there. You reach about standard average level by the time you start with part C (this is the hand-wavy part), so you'll get 20,000 there again. Total: 64,000 exp.

Waiting with the new guys all the way to the beginning of C would make you get reduced exp for both A and B, i.e. 8,000 and 24,000. In order to counterbalance this, you'd need to end up at least TWO levels below the expected for the whole part of C to just barely break even with the standard party (8,000 + 24,000 + 20,000 x 1.2 x 1.2 = 60,800) and THREE levels below to beat the record set by the previous example. (32k + 20k x 1.2^3 = 66,560). Keeping such drastically lower average level for any reasonable length of time isn't possible without external means such as level-squatting.

Anyway, ignoring the possible outcomes either way, I like seeing at least somewhat reasonable exp awards for most of the game, instead of them being on a constant downward spiral. Adding new guys at strategical points will bridge me over the parts where I'd end up at too high level and start seeing zero exp awards. Calculating the exact outcome in terms of total exp gained was never the primary goal.

Well, if you actually read the description of the guide at the download page, I consciously shifted the focus from the example parties themselves to the theory behind them as I consider that much more important. Besides, the section before that -- call it 2.0 if you will -- mentions this, and the Table of Contents has the quick pointers even before that.

Edit: Dangit, seems like I wasn't thorough enough in removing the references to the old parties from the 1.01 version. Unnecessarily confusing. *cringe*

 

Your mathematical model doesn't seem to fit well to the problem.
Whether you add a new character before or after the ice temple area (or the other areas you mentioned in the guide) and get more or less fighting XP there doesn't matter for the average party level near the end any more (you'll get 0 XP per kill very often until then), but the total party XP is higher when the levels are further apart.

I didn't make a statement about what makes the game easier or avoids 0 XP awards, only about the XP at the end. Your guide gives the impression to the reader that you made your "timing" (which seems to be very important since you did elaborate it a lot) at character adding to maximize the XP at the end.


2 questions about the HoW section in your guide:

Is the +3 enemies get to saves because of the +6 stat bonuses already included in the +15 to saves or do they get a total of +18 compared to normal mode ?

Which difficulty settings do you use in the HoF<>normal mode comparison ? Normal mode normal difficulty vs HoF mode insane difficulty ?

 

Well yes it does, as any extra exp you have gained there is exp you don't have to gain later to reach to the same exp level plateau. But you're right in that in the long run (read: several chapters later) the extra exp effect becomes rather insignificant, compared to the ECL penalty total I had in that party at least. Well, if the "more exp" teaser only causes this much confusion, better to remove it then I guess.

There isn't actually all that many monsters at CR8 or below in all of Severed Hand. Since you insisted on that "you can't get much exp at level 16 anyway" point so persistently, I kept an extra eye for that. Some Lemures, Demon Knights, Elite Neo Orogs and the weakest type of guards here and there, but a vast majority is still above that. To counter the Slayer Knight myth, I double-checked that too. Yes, every single one I killed was exactly at CR12.

I quoted from the research done by Nobody Important, but had to double-check this and it showed to be wrong. The actual total save bonus is only +10.

The Boring Beetles at the beginning of Chapter 2 have the following saves:
Reflex: 4 base, +0 DEX modifier = 4
Fortitude: 4 base, +2 CON modifier = 6
Will: 1 base, +0 WIS modifier = 1

In HOF mode, the same beetles had Reflex +14, Fortitude +16 and Will +11 as their saving throw bonus. I tested with several different spells to verify.

Likewise, an Aurilite Frosttouch in Ice Temple has Reflex +3, Fortitude +6 and Will +10 during Normal mode, and Reflex +13, Fortitude +16 and Will +20 in HOF. Exactly +10 across the board.

This doesn't rule out the possibility that the actual save bonus is +7 and they get further +3 from the improved stats, but it's irrelevant in this case.

The flat +15 save bonus Nobody Important came up with was most likely caused by the CHA bonus to saving throws since the type of Slayer Knights in his tests had levels as Paladin of Helm. Stupid, right? That would suggest +9 or +10 bonus to CHA, but since I don't know how he managed to measure the stats of a live HOF monster to begin with, that's not much to base further assumptions on.

Normal mode at Normal difficulty vs HOF mode. The in-game difficulty slider has always been locked at max when I play HOF and can't be adjusted. Do you mean you can force it to some other setting in some way?

 

Recently I've read somewhere that it's possible to adjust the difficulty settings in-game even if HoF mode is enabled. I didn't test it myself however, I know it's not possible in IWD1.

How did you find out the monster CRs ?

 

Monster CR is easy - I keep the MONCRATE.2DA file open in Notepad when I play, and cross-reference the amount of exp I got with the average level I'm at to find the monster's CR. For this particular playthrough (not the same as the 2+2+2 party discussed above) I use an extended exp table, so I get an unique, very small but non-zero value for killing stuff all the way to 15++ CR below me.

Edit: For the record, I'm using the "double EXP in Insane Mode" table, available at Sorcerer's Place, extended with lower exp values downwards. I'm doing a variant of the 2+2+2 party, although this time I added the first 2 extras at the beginning of Chapter 3 and a second pair after Eight Chambers. This showed to be anything but a wise choice: My Human Druid / DG tank duo was at average level 16 at the end of chapter 2, but my now six-person party isn't even average level 17 after doing almost everything in Severed Hand. Shooting *too* much ahead with the first two characters, ESPECIALLY if they trudge through all of Chapter two enduring extreme exp penalties all the way (900 exp for the last killed golem there, yay) is just tossing perfectly good opportunities to earn exp into the trash can.

 

I'm not sure if your Monster CR method always yields clear-cut results.

3E allows adjusting the XP you get for monsters of the same CR, it's possible to give double XP if the monster is considered tough for its CR or half XP if it's considered weak for its CR.
Since the XP gain is often doubled every +2 CR you can't be 100 percent sure about the monster's CR, a level 15 party could get the same XP both for a half XP CR12 and a double XP CR8 monster for example if IWD2 uses this option.

In this case you can only get sure about the monster CRs if you edit the whole mooncrate.2da manually, for example by using for each CR a different prime as factor and 2,6,12 or even 60 as factor everywhere in order to avoid rounding in case you only get a fraction of the points in the table.

 

Well, that would be fine and dandy, not to mention fix that annoying "no exp for kills" thingy when doing combat with HUGE hordes of low CR monsters. Alas, Icewind Dale 2 didn't implement any of that.

The only times I've seen a single monster have differing CR values is when there is several different strengths of the creature available in the .CRE files, such as with Illium, Sherincal and Saablic Tan. Even then, I see the other versions only when I see the area with VASTLY different party average levels.

For example, Sherincal has always been CR11 when I've met her at the Ice Temple. However, the DG/Human combo above were at level 14 upon entering the map so the game actually spawned the CR15 version. Almost dislodged my jaw when I saw that "party has gained experience: 6,300" pop up, I was expecting the same 1,400 I was receiving from the golems at that point.

 

That discussion about what EXACTLY was the cause of the seemingly high exp total that example party of mine ended up with made me think of one peculiarity of the dynamic exp calculation system.

The later you add the extra characters, the more your average level drops at that point. This makes the problem of having useless low level characters vanish faster than you'd think, and the old characters will get a fast shot of experience points on top of the pile they already had.

In other words, don't get carried away even if you're aiming for some target level with your first set of characters as soon as possible. Dragging along with a small party at high average level is only hurting your exp gain rate if you could get the same share of exp per character with the extras already included.

For example, while an Ice Golem would be worth 1,125 exp per character for a duo at level 10, a trio with the two characters and a fresh level 1 newbie would get 1,200 (or more!) per character all the way until they're average level 8. Likewise, while keeping a trio at average level 12 intact against Iron Golems sounds tempting with its 900 exp per character, a quartet will get 1,125 exp (or more!) per character all the way until they're average level 10 and even full party would get 900 (or more!) exp per character as long as they're not higher than average level 9.

The extra characters would need to reach levels 6, 7 or 7 in each respective case before you'd end up there - that's anything between 15k (zero ECL at level 6) to 40k (+3 ECL at level 7) gained experience per character.

In summary: Just because you have less characters sharing the exp doesn't mean they'll get more each. It only works like that with quest experience.

---------- Added 16 hours, 54 minutes and 44 seconds later... ----------

New discovery!

TL;DR version: Going duo or trio isn't the fastest way to reach the absolute highest levels. Starting as a duo or trio and adding new members along the way is. I'm ignoring soloers here on purpose as that's a whole another ballgame.

When the goal is to boost the levels of your initial crew as high as possible as fast as possible, we're actually interested in maximizing the amount of exp per character per unit of action, be it completing a quest or killing monsters. Having less characters is always beneficial when considering quest exp as it's always constant, but the kill exp varies greatly. So how do we strike the balance and aqcuire perfection?

Let's do some more maths, shall we?

Some definitions:
Party's current size = N
Quest exp = q (amount of quest exp in this given area)
Kill exp = k = aq (a is a ratio between quest exp and kill exp)

Just carrying on with the current party without modifications would accumulate (q+aq)/N exp per character.

Adding an amount of b new characters would cause the quest exp to be split in more ways, becoming Nq/(N+b). Kill exp would be split similarly, but since adding new members causes the average level of the party to change too, the kill exp gets a multiplier x. Total exp per character would thus become
(q+aqx)*N/(N+b).

In order to get more exp per character with the new setup, the following inequality must thus hold:
N*(q+aqx)/(N+b) > (q+aq)/N

Taking a cross-product and simplifying, we finally end up with:
x > [(1+a)*(N+b)/N - 1] / a

Notice that since we're concentrating on the RATE of acquiring exp, all we really need to know is the ratio between kill and quest exp ( = a ).

The x here denotes simply that "you need to get this many times more exp for the stuff you kill in order to gain as much exp per character as you would have with the unmodified party". Which, coincidentally, gives a pretty decent estimate on the amount of average levels you'd need to go down by adding the extras.

As long as your kill exp constitutes half or more of your total exp, i.e. when a is 1 or above, x doesn't need to exceed 2, i.e. you don't even need to get double the exp for each kill to break even. Depending on your average level and CR of the monsters you're fighting, you'll only need to become between one and three levels lower average with the larger party to reach this. In my experience, this is both easy to reach *and* maintain pretty much all through normal mode if you start with a duo or trio.

 

Moving away from the exp calculations, some other point that could be addressed is party alignment. For example, if I want a Good party, then a possible decoy will look like
A: Deep Gnome Ill(20)-Ro(1)-Sorc(1)-Monk(6)-Pal(2), 6 20 11 18 18 1
If I want an Evil party, then it's more like
B: Deep Gnome Ill(20)-Ro(1)-Sorc(1)-Monk(6)-Bane(1)-something(1), 6 20 11 16 20 1

(something = Ranger if LN for dual-wield, probably Fighter if LE. No Bard possible, though. This sucks, as singing and decoying complement each other quite well).

(those 6 Monk levels give +1 generic AC, increase movement speed, and cause 20% Exp penalty, which can be compensated for by level-squatting the last 6 levels).

The reason that the INT stat is different is that the best INT-boosting item (Xvimian fang) is useless to Good characters. Other interesting sources are*:

• Wizard's ring (+1, just equip it prior to casting 9th-level spell),
• Phaen's staff (HOF) (+3, but two-handed, so it disables Sulo's hook),
• various Robes (useless for a Monk unless patched).

So B had to be given two extra points in INT. (Alternatively, two more points could be taken off STR, and gotten back later via Paladin quest, but a 4 STR character would need constant use of Bull's Strength...)

In endgame with Chimandrae, two Potions of Holy Transference and various quests, those two characters will look like
A: 6 30 11 16 33 1
B: 8 30 11 18 29 1
With Expertise, Dodge, Bands, Crow's Nest, Sunfire, Sulo's Hook, and spells Shield, Barkskin, Haste, War chant of Sith, A gets AC 74 while B gets 72.

So B needs a singing level 11 Bard while A doesn't; but then B is Holy Word-compatible.

On the remainder of the party, there are also some differences:

• a non-Good party won't get either Holy Word or Holy Avengers (sweet 45 SR by dual-wielding them for a Pal(1)-Sorc(29)...)
• a Good-aligned party won't get a Dreadmaster (this sucks). If this party has a Cleric/Druid, this character must then be Neutral Good, and thus can choose between Ilmater, Lathander, Selune, and Oghma. Oghma looks like a good build (9th and 7th level domain spells are excellent but 8th level is useless; all special capabilities for these four domains suck but Identify 1/day frees one Wizard slot for one more Chromatic Orb).