View Issue Details

IDProjectCategoryLast Update
0007206Valley 1Suggestion - Balancing IssuesJul 22, 2012 6:03 pm
ReporterPurlox Assigned ToChris_McElligottPark  
Status resolvedResolutionfixed 
Fixed in Version1.002 
Summary0007206: Specializing your character takes too long
DescriptionGetting enough upgrade stones to specialize your character(s) takes too long from my experience. I have to spend a lot of minutes going from building to building, from stash to stash, but even then it takes a lot of time. I didn't mind at first, but now when I'm about to kill the Overlord, spending tens of minutes just on getting enough upgrade stones when I could be fighting him or trying to get different spells to see if fighting against him would be easier using them.

I think this is mostly cause by the huge cost of upgrade stones needed to make anything other than all-around character. I propose either:

A) Cap the number of upgrade stones needed for an upgrade at 32.
B) Instead of doubling the number of upgrade stones need for an upgrade, increase it by 8. (so it will be 8/16/24/32/40... instead of 8/16/32/64/128)
Tagsgrind, specializing, stones, upgrade
Internal WeightFeature Suggestion

Activities

Chris_McElligottPark

Apr 21, 2012 11:04 am

administrator   ~0022575

You're really not expected to be all the way at caps -- if you're constantly at caps in order to win, then probably I'd suggest taking the difficulty down slightly, to be honest.

TechSY730

Apr 21, 2012 2:22 pm

reporter   ~0022577

Last edited: Apr 21, 2012 2:22 pm

He does have a point, specializing is rather harshly punished by the exponential growth. I'm pretty sure that you said something you learned from AI war is that exponential growth rates are near impossible to balance. However, linear growth seems like it would be too easy (though, with the 10 upgrade cap per character, that might not be an issue)

Maybe polynomial, like something proportional to n^2 or something?

tigersfan

Apr 22, 2012 9:37 am

reporter   ~0022588

I'm with the OP, it does take a REEEAALLLLLY long time to get all 10 upgrades. That said, only going up by 8 each upgrade is probably not long enough.

Terraziel

Apr 22, 2012 9:59 am

reporter   ~0022590

I think instead of lowering the costs I'd rather the benefits increased, if after 5 points in a category the benefits of each point doubled, or whatever, I think that would outweigh the cost, and would probably be horribly unbalanced but you'd get what you pay for.

That said if you are going to lower the cost then I think point 10 costing about 500 upgrade stones is about right, because I don't know about anybody else but I have 1000 upgrade stones just sitting in my inventory, and that's after spending 1000 on upgrading a different character, and I'm only on continent 2.

Either way the costs have to take into account the fact that you can get, well I think the most upgrade stones I've acquired from one stash is 50+, so they are hardly rare.

Purlox

Apr 22, 2012 1:56 pm

reporter   ~0022599

Last edited: Apr 22, 2012 1:57 pm

@tigersfan_: Don't forget some people might want to specialisse more than one character at a time.

@Terraziel: You must be very lucky to get 50+ upgrade stones from one stash and you are certainly playing on a difficulty lower than you should be, because I have around 0 upgrade stones all the time. So having a 1000 upgrade stones in your inventory seems like you don't die much.

tigersfan

Apr 22, 2012 2:06 pm

reporter   ~0022601

Yeah, I'm aware that people might want to specialize more than one character. But, I'm ok with that taking a significant amount of time, as it adds significant benefit.

martyn_van_buren

Apr 22, 2012 6:54 pm

reporter   ~0022604

I don't know about the exact balancing, but it seems valuable to me to have the last two-three upgrades be prohibitively expensive. It seems good if the game works out so that you tend to leave characters with a little bit more you could do to them if you're desperate or really love them. That said, I've never applied more than two-three upgrades; I go through characters pretty fast. Although maybe I wouldn't if I spent more on upgrades.

khadgar

Apr 23, 2012 6:32 pm

reporter   ~0022618

I would suggest an end-weighed upgrade formula.
Instead of 2x, I suggest (0.21y+1)x, where y is the number of times the stat has been upgraded. So, the first time, the amount is increased bv 1.21. The second time, it is increased by 1.42, and the third time it is upgraded it costs 1.63 times more. When you hit the end, the last upgrade is nearly 3 times as much as the previous one. This could be tweaked slightly, but if you want to start with 8 as the first cost, it plays out nicely. Net cost is approximately half as many total stones consumed to upgrade a single stat.
8 10 14 22 41 85 191 472 1264 3654

tigersfan

Apr 23, 2012 6:36 pm

reporter   ~0022619

Hmmm, interesting idea khadgar.

GUDare

Apr 24, 2012 6:18 am

reporter   ~0022634

Last edited: Apr 24, 2012 6:19 am

My personal concern with the cost of upgrading is the wicked level of diminishing returns. By the linear nature of the upgrades in the first place, it's a diminishing return. A quick example.

Starting at ~ 105 % attack, you get roughly 10-11 %/level. This is great for the first 4 upgrades, as you get roughly 40% of a damage increase at the base (not accounting for enchants).

The next 4, if you're trying to specialize, will cost you (8... carry the 2...) 1920 / 120... 16 times as much for a linear 40% more, but comparably a ~ 28 % increase. My numbers for each increase might be slightly off but you get the drift.

I realize those boosts are larger on pre-specialized characters, but the price tag and ratios don't really change much. If anything, that diminishing returns is more obvious on pre-specialized characters.

Add to that... you're right, I try not to die much (if at all) because of the upgrading price. Am I playing at too hard a level? meh, possibly. However, there's a cap on upgrades. There's no reason in my mind I shouldn't be striving to get to that cap and STAY there until I've got enough stones to re-upgrade the next poor sap tapped by the council elders to go off and die in the wilderness. If it was an uncapped upgrade system I'd say you pays your money and you takes your chances. The system as presented seems like you could probably get away with a 1/2 upgraded character in T1 but by T4/5 you're going to want a nearly maxxed lemming.

This adds to the price of perma-death though and really is the heavy inspiration to me doing much if any exploring. I'd rather be doing gathering (and recently was for Continent 2 buildings) but that's another discussion. I'd like the upgrades to continue to be expensive. I'd like them to not be so prohibitive to a specialized character.

A 3/4/3 build currently costs you 288 stones. Not an unreasonable number to hit 10/10 with. Especially if you are tight in your exploring and keep it to specific structures.

However, a comparable 2/7/1 build (a common one for me) costs 1072. If *huge* mp spells are ever released a 2/2/6 build might be common too. These specializations take over 3 times the investment to generate, and that's the primary concern here, at least to me.

Chris_McElligottPark

Apr 24, 2012 8:35 am

administrator   ~0022640

Okay, let me state my design goals for the upgrade stones, which may help to steer the discussion a little bit. I like where a lot of the thinking here is headed, but this may help clarify the goals:

1. Absolutely, this is intended to be something that makes you think twice before dying.

2. This is also meant to be something that you think twice before spending the upgrade stones. If you always have so many stones that you just never have to worry about what you spend, then there's no decision making aside from WHERE you spend it. There should also be the "do I save now or spend now?" choice.

3. The goal is that it is generally non-optimal to go to all 10 upgrades -- seriously. So that if you're topping out at 6-8 upgrades, that's the "best value" most of the time. This feeds back into design goal 2, but it also means that for players who want to have a bit of a "supercharged" character, they can push all the way to 10 upgrades and then REALLY fear the loss of that character.

4. The other goal feeding into point 3 is for the game to have a really solid way of eating excess stones you might have. If you've been playing for a really long time and are super good (Terraziel plays on The Chosen One, by the way, so if he's playing on too easy a difficulty I need to add more difficulty levels), then the worry is getting an excess of stones. However we can't just cap the inventory arbitrarily -- that's lame and transparent. So instead we provide some super expensive options (see point 3) that you can routinely sink your excess stones into if you're playing along and not dying a lot.


Now: all that said, I'm not opposed to changing the formula. It may be that the lower-down levels of this are increasing too fast. And then it may be that the higher-up levels need to increase faster.

Then again, you wind up with problems where it's encouraging you to spread your upgrades out rather than to specialize your upgrades into one category. I'd kind of like there to be some penalty for stacking into one category, but not so severe as we have now or as is being suggested in this thread.

In that regard, my thinking is that we may want to make some changes along these lines:

A. The first few upgrades on a character are cheaper than now (after the first), but ramp up rapidly to pretty darn high costs.

B. The baseline cost increase % is constant across all of the categories rather than being category-specific, but the categories you have upgraded more also get something like 10% more expensive with each upgrade to make that not an obvious choice just to dump all your stones into one stat.

Penumbra

Apr 24, 2012 9:18 am

reporter   ~0022643

I thought the original point of the geometric series for cost was to not need the "arbitrary" limit on upgrades. If each category cost the same amount at all times (like B above) they player would really have the choice of which to do, with the hard limit coming from when future upgrades are prohibitively expensive.

However, the unbounded enchants would lead to players eventually having an invincible character with infinite HP. That would take nearly forever, though.

Chris_McElligottPark

Apr 24, 2012 9:31 am

administrator   ~0022645

The original point of the geometric series is irrelevant, and wasn't suggested by or endorsed by me particularly. ;) By the time that these were implemented, the above were my design goals.

The hard caps are absolutely needed because if it's unbounded then players will save-scum. And plus then you're incentivized to just upgrade repeatedly in all categories as opposed to having to choose between them more.

Penumbra

Apr 24, 2012 9:44 am

reporter   ~0022646

That makes sense. I find it never makes sense right now to ignore health, which is probably why it starts off costing more. And specializing is more expensive in your idea, so that's great. As long as the overall cost "hurts" enough to make your character's death meaningful.

Chris_McElligottPark

Apr 24, 2012 9:49 am

administrator  

Chris_McElligottPark

Apr 24, 2012 9:55 am

administrator   ~0022647

I just uploaded "AVWW Upgrade Stones Math.xlsx" as an attachment to this topic if anyone wants to play around in there. It plays around with the idea as proposed by khadgar, but it adds more exponential gains later in the upgrade process and has a lower first bar to entry.

My logic is that getting back to 3-4 upgrades right now is far too hard. That shouldn't be much of a penalty, that should be more about choice. Then getting to 6-8 stones should be harder to do, and something you feel sorry if you lose it. Getting to 8-10 stones is something that is even harder to do, and usually a waste of stones if you die a lot. But it provides the benefits of being a sink for excess stones as well as a temptation to get players to waste stones and make poor choices (and thus learn from those and adjust -- like "candy" techs in AI War that that players were always wanting me to put in, where it seems tasty but actually rots your teeth).

The one thing that my model doesn't include is the added 10% cost for over-specializing in a single line, so I should probably factor that in. And I guess the overall growth may need to be toned down a bit more anyhow, reading back over GUDare's comments in particular.

Anyhow, I'll play around with it some more and then post my altered formulae. For the record, the changes I made so far are:

- Base costs from 16/8/8 to 8/4/4 for health/mana/attack.
- Cost of tier > 1: ((0.1*(X*X*X)+1))*base, where X is the current tier.

Chris_McElligottPark

Apr 24, 2012 10:07 am

administrator   ~0022648

Okay, I like this formula even better:

- Base costs from 16/8/8 to 8/4/4 for health/mana/attack.
- Cost of upgrade number > 1: (((0.05+(0.05*Y))*(X*X)+1))*base, where X is the current upgrade number and Y is the number of those upgrades that were previously applied to the current category you're upgrading.

If you look at a 10/0/0 build, that's: 1289.6 stones.
If you look at a 0/10/0 or 0/0/10 build, that's: 644.8 stones.

If you look at a 7/0/0 build, that's: 369.2 stones.
If you look at a 0/7/0 or 0/0/7 build, that's: 184.6 stones.

If you look at a 4/0/0 build, that's: 71.6 stones.
If you look at a 0/4/0 or 0/0/4 build, that's: 35.8 stones.

If you diversify more than this, then you get some savings on top of these, too, but it's not huge.

Chris_McElligottPark

Apr 24, 2012 10:17 am

administrator   ~0022649

Another option, which takes some of the teeth out of the middle of the road approach (the 7 stones options) and adds more teeth to the fully-upgraded options:

- Base costs from 16/8/8 to 8/4/4 for health/mana/attack.
- Cost of upgrade number > 1: (((0.03+(0.01*Y))*(X*X*(X/2))+1))*base, where X is the current upgrade number and Y is the number of those upgrades that were previously applied to the current category you're upgrading.

If you look at a 10/0/0 build, that's: 1335.2 stones.
If you look at a 0/10/0 or 0/0/10 build, that's: 667.6 stones.

If you look at a 7/0/0 build, that's: 305.64 stones.
If you look at a 0/7/0 or 0/0/7 build, that's: 152.82 stones.

If you look at a 4/0/0 build, that's: 54.04 stones.
If you look at a 0/4/0 or 0/0/4 build, that's: 27.02 stones.

If you diversify more than this, then you get some savings on top of these, too, but it's not huge.

That's something that would get people back to 4 upgrades really trivially -- goodness that's trivial -- but then gets them to 7 with more expense than in the currently-implemented model in the game.

TechSY730

Apr 24, 2012 10:17 am

reporter   ~0022650

Wow, so it looks like we are going with my earlier suggestion of polynomial growth rate of cost, though it looks like its going to be O(n^3) rather than O(n^2).

Fun stuff. :)

Chris_McElligottPark

Apr 24, 2012 10:19 am

administrator   ~0022651

Well, that's my thinking at the moment. Incidentally, we use to use O(n^3) growth for some of the stats when we had levels and EXP. :)

Penumbra

Apr 24, 2012 11:57 am

reporter   ~0022656

You mentioned wanting to discourage save scumming. What if the Vengeful Ghosts dropped a small % of the upgrade stones used to kill them. You could even increase the total cost a bit by the same amount they drop, only psychologically lessened.

It could actually end up being a larger Upgrade Stone sink if only the last ghost gave stones, and the player died going after them. Are Vengeful Ghosts buffed by the upgrades they had at death?

Chris_McElligottPark

Apr 24, 2012 12:17 pm

administrator   ~0022661

Vengeful ghosts aren't buffed by anything relating to the character that they came from: they take the name and the general appearance of that character, but that's all. I don't think I'd want to get too complex with that, although it's certainly technically feasible. It just seems like getting a refund from the ghost would be giving something back a bit too much.

And when a player dies in a mission instance and goes back to the settlement, their ghost gets lost when the mission instance goes away, incidentally. That's something that could really impact balance if the ghosts are at all involved in this part of it.

Penumbra

Apr 24, 2012 1:04 pm

reporter   ~0022666

You're right. Even if it were implemented as a "small bonus" when recovered, it would be immediately perceived as a "huge loss" when not.

martyn_van_buren

Apr 24, 2012 3:48 pm

reporter   ~0022670

By the way, why is fully upgrading one stat more expensive than evenly doing them all? Not asking for a change, just I'm sure there's a reason for it and it's not obvious to me.

Chris_McElligottPark

Apr 24, 2012 3:51 pm

administrator   ~0022671

Well, the most overpowered characters tend to be the most extreme. So trying to make it so that you at least pay extra for those more extreme builds is a goal for mainly just that reason. It's a fine line, because we want them to be pretty flexible in terms of what you can do, but at the same time not making it so that everyone goes for "all in one category" or "always even across all categories."

GUDare

Apr 25, 2012 1:56 am

reporter   ~0022688

Martyn: If you've got some time to kill, take a character that you've maxxed off hp on that starts with a hp buff, equip Flame Touch or other equivalent touch spell, and walk up to a landspeeder and rip its head off, ignoring everything it does. :)

In general, though, I'd have to say every one of my characters starts off with a 2/2/1 upgrade, simply due to price and the fact that I probably don't have a lot of stones after I just lemming'd the last guy. After that I look to 'tweak' the build, usually dumping a ton into damage boosts. I might toss a second bonus into mp if it's REALLY hurting (IE: even with -50% mp costs I can't fire for 5 seconds continuous), but mostly, yeah, it's damage boosts.

All three abilities taken to the extreme change something drastically, whereas balanced characters are more 'reasonable' to, well, balance for.

HP: Time allowed to continue farming/exploring before a settlement return or trash-mob run is required.

MP: Time allowed to engage a mob at full force before requiring retreat.

Damage: Reduction of time required to engage any mob.

Take any of those to an extreme and it becomes tough to balance... which is kind of the point, they're anti-balancing, thus overpowered. Couple this with a good mix of enchants to cover the 'weak spots' (ie: rediculous amounts of mp reduction) and you don't have to be overly concerned with the weak points.

As an example that's not one I personally use, imagine a high end RiskTaker + -damage legs + 8/1/1 build. You've negated, really, the danger of the risktaker with hp, covered it some more with anti-damage, and can now basically pwn with impunity all the damage you're missing from not beefing attack. With HP +s being so high right now it's a pretty expensive build, but it's viable.

So, knowing that, there definately needs to be an additional 'stop it!' cost to highly unbalanced characters. Right now it approaches the 'Stove is HOT stupid, no touchie' levels, though, instead of being uncomfortable. Part of this is the time price to exploration, part of it is in the intrinsic price of learning HOW to explore effeciently and what buildings to ignore (ie: Maze room, skip!).

However, that last paragraph isn't mechanics or methods, it's biased. I'm a cheesy player and I tend to break things. A lot. :) The rest is just observation on how to do it.

On a side note... if you want some real whack'em fun, grab yourself a damage pre-incentiviced character, slap on + ele damage rt arm, RiskTaker with +ele damage, High speed legs for quick escapes, and Mass cooldown reduction on left arm. Boost damage to comfort levels (3/4 times as the experiment at first) and let 'er rip. This should help anyone who's not sure of why un-balanced characters who DON'T have a cost differential become amazingly overpowered...

... it's also the reason all those dang things pierce, so you can't get someone to 'tank' for you... I am NOT the guy and I die a lot. Thanks for noticing... :)

GUDare

Apr 25, 2012 2:09 am

reporter   ~0022689

Chris: My personal 'feeling', pure bias without all the numbers worked out, would be that a 2x cost for a heavy specialization build vs. the cost of a non-specialized build ( IE: 1/1/1 vs. 0/3/0 is included in 3/4/3 vs. 2/7/1) would feel less heavy-handed, but would help balance the time to power requirement.

There's some (not much, I'm NOT a good platforming guy) skill involved in keeping a lightly hp/mp'd character alive, you forsake a few enchants in place of 'covering weaknesses', but any specialized build would require that in some method.

The current 288 (352 for 4/3/3) stones for a 10/10 pure balance seems fair to me. I'd dislike seeing that pushed higher.

Also I'd like you to consider the 'order' of purchases not heavily affecting the result. I really need to dissassemble the calculation you have above, it doesn't come easily to me, but it seems like there will end up being an 'optimal build path' that comes out of that to minimize stone usage. I'd need to really sit down with it and figure out if you always do HP first or your biggest stat first, but I think it ends up as biggest stat first to reduce the top end, hp next, and leftover last.

Chris_McElligottPark

Apr 25, 2012 11:07 am

administrator   ~0022699

Hmm, that's a good point on order of operations -- but if prior actions affect the costs across the three stats, there's no way to avoid that.

The only way to avoid that sort of problem would be to take a simpler formula that is still based on the current stat being upgraded only, and looks at its history alone when determining the cost of the next upgrade.

You know, for the sake of not giving advanced mathy players a crazy advantage, that's pretty much going to have to be the way we do it -- thanks for pointing that out. I'll need to revise the numbers from above completely.

That means sole goal is to thus make it so that the builds start cheaper and then get polynomially more expensive in each category as you go.

Penumbra

Apr 25, 2012 11:55 am

reporter   ~0022703

Or, you could recalculate the costs based on whatever the optimal path would have been. So, still have the exact same end cost with all the pros and cons as mentioned, but not have the player deal with order.

Penumbra

Apr 25, 2012 12:07 pm

reporter   ~0022704

Actually, I really like this. You would upgrade one stat, and see the cost for a different one _decrease_

That would really help incentivise a diverse build. And show that going all in one direction is more expensive.

Chris_McElligottPark

Apr 25, 2012 12:44 pm

administrator   ~0022705

I can't for the life of me think how a formula for that would work. Unless maybe...


- Base costs from 16/8/8 to 8/4/4 for health/mana/attack.
- Cost of upgrade number > 1: (((0.05+(0.05*Y)-(0.05*Z))*(X*X)+1))*base, where X is the current upgrade number and Y is the number of those upgrades that were previously applied to the current category you're upgrading, and Z is the number of those upgrades that were previously applied to any other category.

Also, it would have to floor it so that (0.05*Y)-(0.05*Z) could never be negative; it would just clamp at 0.

That... might do it.

Penumbra

Apr 25, 2012 1:02 pm

reporter   ~0022706

Health 8 8.4/11.2/18.8/33.6/58/94.4/145.2/212.8/299.6 890


Maybe it should be (0.1*Y)-(0.1*Z)?

Health 8 8.4/12.8/26/52.8/98/166.4/262.8/392/558.8 1586

Penumbra

Apr 25, 2012 1:05 pm

reporter   ~0022707

Mana 4.4/7.2/14.8/29.6/54/90.4/141.2/208.8/295.6 846

looks like it's off by half

Penumbra

Apr 25, 2012 1:10 pm

reporter   ~0022708

(((0.15+(0.15*Y)-(0.15*Z))*(X*X)+1))*base gives:


Health 8 9.2/17.6/40.4/84.8/158/267.2/419.6/622.4/882.8 2502

Mana 4 4.6/8.8/20.2/42.4/79/133.6/209.8/311.2/441.4 1251

Chris_McElligottPark

Apr 25, 2012 1:22 pm

administrator   ~0022709

That seems a reasonable one -- and particularly with health getting so incredibly expensive. It may be that health needs to start at 16 or 12 instead of 8, or something. Not sure. Having more than a handful of health upgrades ought to be bloody expensive. ;)

Penumbra

Apr 25, 2012 1:26 pm

reporter   ~0022712

Wait, my math is wrong... one sec.

Penumbra

Apr 25, 2012 1:41 pm

reporter   ~0022713

Last edited: Apr 25, 2012 2:13 pm

(((0.1+(0.1*Y)-(0.1*Z))*(X*X)+1))*base gives the original series for straight upgrade paths.

Health 8 14.4/29.6/59.2/108/180.8/282.4/417.6/591.2/808 2499.2

Mana 4 4/7.2/14.8/29.6/54/90.4/141.2/208.8/295.6/404 1249.6

Going half and half, 5 mana then 5 damage would give this:

5M/5D 4/7.2/14.8/29.6/54/18.4/23.6/29.6/36.4/44 261.6

Alternating mana and damage give something very different
1M/1D*5 4/5.6/7.6/10.4/14/18.4/23.6/29.6/36.4/44 193.6

I can figure this out, it's actually really fun.

TechSY730

Apr 25, 2012 2:09 pm

reporter   ~0022715

At first, I didn't really like idea of order dependency of upgrades.
But then I realized that doing something like this not only encourages diversifying overall, but diversifying earlier rather than later (aka, level up stats at around the same time, not focus on one then focus on the other)

And as long as the only the cost is path-dependent, and the stats will be the same no matter which order you upgraded them in, I don't have a problem with this sort of model.

Again, numbers may need to be tweaked some, but the overall behavior of the formula I could live with.

MouldyK

Apr 25, 2012 3:01 pm

reporter   ~0022724

Oh geez! Wandering in here is like wandering into my A-Level Maths class.

-draws a U on his paper and walks out- I know how it ends. :P


But the parts I do understand sound nice. Keep it up, brave scientists!

Penumbra

Apr 25, 2012 5:06 pm

reporter   ~0022725

Last edited: Apr 25, 2012 5:42 pm

Got it! The key was to use a separate additive cost per total level to modify the individual cost of each skill level. This scales quite nicely and uses Logs to be a bit kinder in the end.

The bases of each skill needs to change, to 7, 3 and 3 (but the user will not see that) With X being the total level being upgraded to (from 1 to 10), and Y being the new rank of the skill being raised, here is the formula:

((X^2 * Log(X)) + 1) + (((Y^2 * Log(Y)) + 1) * base)

This works for all levels, completely indifferent to order in which the skills are taken.

Some examples.

All Health: 8/17.6/42.3/85.0/147.7/232.1/339.2/470.3/626.3/808 2776.9
All M or D: 4/8.8/21.1/42.5/73.8/116.0/169.6/235.1/313.1/404 1388.4


5H + 5M: 8/17.6/42.3/85.0/147.7/32.0/49.0/74.6/110.1/156.4 723.1
3H+3M+2H+2M: 8/17.6/42.3/13.6/25.0/44.8/116.8/188.1/110.1/156.4 723.1

3D+3M+4H: 4/8.8/21.1/13.6/25.0/44.8/49.4/74.2/115.3/175.4 532.0
3H+3D+1M+1H+2M: 8/17.6/42.3/13.6/25.0/44.8/45.4/133.2/84.9/116.8 532.0

Penumbra

Apr 25, 2012 5:14 pm

reporter   ~0022728

Last edited: Apr 25, 2012 5:17 pm

Also, throwing just a few levels in at the low end is pretty cheap, in particular, 2/2/2 only costs 108!

I attached the spreadsheet I used.

Penumbra

Apr 25, 2012 5:17 pm

reporter  

AVWWLogSkillLevels.xlsx (17,754 bytes)

Chris_McElligottPark

Apr 25, 2012 7:08 pm

administrator   ~0022732

Sweet! Thanks for cracking that -- that looks like really good numbers to me. I won't have time to implement that tonight's release, but I'll try to get that in there tomorrow. That gives other folks a chance to respond and comment anyhow.

Any other thoughts?

Penumbra

Apr 25, 2012 7:19 pm

reporter   ~0022733

Last edited: Apr 25, 2012 7:19 pm

I think this puts 10 Health almost completely out of the realm of possibility and it does greatly incentivize spreading out your upgrades a lot.

I think it also goes well with your goal # 3 above, of making any of the upgrades above, really, 5 or so, not worth it.

3's and 2's, maybe the occasional 4 will probably be the norm.

Chris_McElligottPark

Apr 25, 2012 7:21 pm

administrator   ~0022734

For the 10-health characters, that is basically "dies instantly no matter what," although body enchants can help.

Penumbra

Apr 25, 2012 7:25 pm

reporter   ~0022735

Maybe Health base doesn't have to be 8 anymore. Making the base "5" in the spread sheet reduces the cost of the 10/0/0 to 2082.7, while only reducing the 2/2/2 to 101.

Only 225 for a 5/0/0

Penumbra

Apr 25, 2012 7:26 pm

reporter   ~0022736

Having a really buffed 10 health character is effectively playing the whole game as a JtP mission. So much damage they can dish out. And, on the higher difficulties, the difference between 10 and 100 HP is almost nothing

GUDare

Apr 25, 2012 9:12 pm

reporter   ~0022742

This calculation practically doubles the cost of a 'spread build' 3/4/3 (288 -> 532) that was being used previously.

Is this the desired result? One of the key beginnings of this discussion was the overwhelming price of specialization compared to genericity. This will just make them all whoppin' expensive.

For example, on a fully explored windmill I usually expect to get 40-50 stones out of it. This would cost 12 windmills (roughly) to pull a 10/10 out with a generic build... if you can find 12 of them.

If the goal is to make 'topping off' at 9/10 to be truly expensive, what's the time price for an 8th upgrade character (with a reasonable cushion for death) expected to be? Even for efficient explorers, how does that look right now for an average of 1 16x/4 stashes (or so) and the majority of stashes containing 2x 3xstones? I'm pulling these numbers from my memory and experience so they are probably off, but that runs into roughly 37 stones / 4 stashes.

One of the efficiencies in this game is time vs. power. Your design goal states the expected power, but what's the expected time?

Penumbra

Apr 25, 2012 10:19 pm

reporter   ~0022744

I made the equation with the intent of making level 10 "out of reach" for pratical purposes and a balanced 6 easily within reach at 100 stones, and 200 if you want 4/1/1.

I like having "room" above. Just always picking 10 enchants is boring and predictable. Actually, with the scaling as is, you could probably remove the artificial restriction of 10 levels. That would stop people from feeling arbitrary attachment to getting ten upgrades. You just get as many as you like. This math gets very nasty past 10, so you might get the very, very occasional 11 or 12, but it wouldn't really be an issue. Because of the linear increase in stats, the "bang for your buck" goes down rapidly.

Chris_McElligottPark

Apr 26, 2012 10:18 am

administrator   ~0022754

Generally speaking, my goal with the new math has been to:

1. Have players get 2-4 upgrades a lot more cheaply than now.

2. Have players get 6-8ish upgrades with about the same expense as now but with more flexibility in terms of builds so that it's not really spiking some of those builds so expensively.

3. Have anything above that be crazy expensive, so that most characters aren't maxed out. However... that might indeed lead to people save-scumming quite a bit. It really might be the case that the top end of these expenses needs to not be so high.

Penumbra

Apr 26, 2012 11:02 am

reporter   ~0022755

You can change the base of the two Logs to whatever number you want to incentivize. Separate for total levels and individual skills.

Purlox

Apr 26, 2012 12:09 pm

reporter   ~0022760

I like the equation by Penumbra, but I dislike that it doesn't help reduce the cost of specialization unless you go for 7+ upgrades in certain path and that it increases the cost of non-specialization builds. So I created my own version of the equation:

Y = (2 * X * log(X) + 1) + ((1.2 * Z^2 * log(Z) + 1) * base)

where base is 7/3/3 depanding on what you are going to upgrade, X is the number of upgrades this character will have after this upgrade, Z is number of upgrades in the certain 'path' after this upgrade.

3/4/3 build costs around 260 upgrade stones (slightly less than now)
2/6/2 build costs around 370 upgrade stones (around two thirds of what it costs now)
1/7/2 build costs around 500 upgrade stones (around half of the current cost)
0/0/10 build costs around 1340 upgrade stones (around one sixth of what it costs now)

@x4000: One of the reasons I want to have the cost specialization reduces is, that I mostly never get to the point where I have fully upgraded my main all-around character and I now want a specialized charater for certain types of play only viable with specialized characters or to be able to do certain missions more easily than I could have done with my all-around character. And I can't bring myself to using the upgrade stones on a different character, when my all-around character isn't fully maxed out yet.

Penumbra

Apr 26, 2012 12:49 pm

reporter   ~0022761

Check out the forums, I put the whole thing up as a modifiable spread sheet, and you can change all the values around.

Penumbra

Apr 26, 2012 1:29 pm

reporter  

Chris_McElligottPark

Apr 27, 2012 4:44 pm

administrator   ~0022864

Here's the new design for them:

* Previously, the first health upgrade cost you 16 and then it would go up by a power of 2 from there for each further upgrade (32, 64, 128, etc). Mana and Attack were the same, except starting at 8 (so then 16, 32, 64, 128, etc).
** Now the fee is flat for all upgrades: it's 20 stones for a health upgrade, and 10 for attack or mana.
** After much discussion on this subject, it became clear that being able to apply 10 upgrades to your character is an important avenue for player choice, but not something that should be prohibitively time consuming to do. Even if the upgrades past the 8th were a "bad value" to encourage players to consider conserving stones, some players would push all the way to 10 upgrades no matter what and then be encouraged to save-scum to protect that character.
** Further, not having 10 upgrades on the primary character was creating a disincentive to create a stable of other characters for various purposes.
** Lastly, it was tricky to balance the game because if players used a full 10 upgrades then things could be far to easy; but with no upgrades it could be far too hard. Now we can balance around the general expectation of roughly 10 upgrades being on most characters; and these upgrades being more about choice than they are about a long-term slog through getting thousands of stones.
** The reason for keeping a cost to the upgrades at all is to maintain that sense of loss on death of characters; aside from the vengeful ghosts, upgrades are the one thing that is lost when your character dies. That has seemed a very popular thing (and in fact upgrades were originally added when beta players felt there was not enough of a penalty for death), but we have been aiming to balance it so that the penalty is noticeable but not harsh.

* Previously, each upgrade that you applied in the health, mana, or attack categories would give you a flat bonus every time you used an upgrade. That made sense when the costs of each upgrade went up exponentially. However, now that the costs are flat (not even linear, but literally flat), we chose to instead make the bonuses from each upgrade decay so that they remain balanced.
** Applying 10 health upgrades now gives a maximum 5.57x bonus (so 20% decay) compared to 10x previously (0% decay).
*** With a character of base health 141, that gives the following progression: 141,282,395,485,557,615,661,698,728,752,771,786
** Applying 10 attack upgrades now gives a maximum 1.65x bonus (so 10% decay) compared to 2x previously (0% decay).
*** With a character of base attack 85, that gives the following progression: 85,93,100,107,113,118,123,127,131,135,138,141
** Applying 10 mana upgrades now gives a maximum 2.63x (so 30% decay) compared to 6x previously (0% decay).
*** With a character of base mana 100, that gives the following progression: 100,150,185,210,227,239,247,253,257,260,262,263
** Cumulatively, these changes do help to encourage players to choose characters with base stats that somewhat mirror what they want the final stats to be -- because the changes you can make to a given character are somewhat less extreme, although still really notable. The penalty for diversification is also a lot less now compared to what it was, but the penalty for stacking everything into one stat is now a penalty of effectiveness rather than of cost.

* Mana upgrades have been the least useful of the three kinds of stat upgrades for a while. Part of that is because most of the really high-mana-cost spells that we have planned are not yet in the game. So some of that is just a matter of intent for later stuff.
** However, to address this imbalance in general, we made it so that mana upgrades also simultaneously upgrade mana regen rates. Normally all characters have a flat 83.3 regen rate for mana unless they have had some mana upgrades via upgrade stones; given that upgrade stones are the only way to increase mana regen, that makes this suddenly a lot more interesting.
** Applying 10 mana upgrades now gives a maximum 1.3x bonus to mana regen (so 10% decay on 1 5% boost per increase) compared to 0x previously.
*** So the progression for any character is: 83.3,87.3,91.3,94.3,97.3,100.3,103.3,105.3,107.3,109.3,110.3,112.3.

* The logic for how you find upgrade stones has been changed up somewhat:
** Places that previously dropped 8 upgrade stones now drop 5. Places that dropped 16 now drop 10.
*** Except in the intro mission, where the caches of 8 are now caches of 50, and the caches of 16 are now caches of 100. This lets players explore these mechanics a lot more right in the intro mission.
** Places that previously dropped 4 upgrade stones now drop 5, and those that previously dropped 3 now drop 1.
** Now when you kill a miniboss, you get 5 upgrade stones. When you kill a microboss you get 1.

Penumbra

Apr 27, 2012 4:57 pm

reporter   ~0022869

Wow, that's actually really simple. I was so focused on increasing the cost to lower value, I completely missed the option of just lowering the bonus! Math on that upgrade path is _so_ much simpler..... ;)

Chris_McElligottPark

Apr 27, 2012 4:59 pm

administrator   ~0022871

Yeah, sorry about not using all that work you put into the math -- you did a really awesome job on it! But in the end I think this is simpler all around. I had to change some of my design goals to make it work, but some of those design goals just weren't compatible with one another, as folks pointed out. In the end the thing to emphasize was choice. :)

Penumbra

Apr 27, 2012 5:01 pm

reporter   ~0022872

It was actually a lot of fun to put together. Maybe you can use it for something else sometime. I learned a lot about excel in the process.

Work smarter, not harder. That's what you did ;)

Chris_McElligottPark

Apr 27, 2012 5:03 pm

administrator   ~0022873

Cheers. :)

Purlox

Apr 27, 2012 5:11 pm

reporter   ~0022874

I like the change especialy getting upgrade stones from minibosses and microbosses, but couldn't you find a boss room in non-mission areas or are the bosses inside these areas actualy minibosses?

Chris_McElligottPark

Apr 27, 2012 5:15 pm

administrator   ~0022876

Minibosses do exist outside of missions, yes.

Issue History

Date Modified Username Field Change
Apr 21, 2012 9:37 am Purlox New Issue
Apr 21, 2012 11:04 am Chris_McElligottPark Note Added: 0022575
Apr 21, 2012 2:22 pm TechSY730 Note Added: 0022577
Apr 21, 2012 2:22 pm TechSY730 Note Edited: 0022577
Apr 22, 2012 9:36 am tigersfan Internal Weight => Feature Suggestion
Apr 22, 2012 9:36 am tigersfan Status new => considering
Apr 22, 2012 9:37 am tigersfan Note Added: 0022588
Apr 22, 2012 9:59 am Terraziel Note Added: 0022590
Apr 22, 2012 1:56 pm Purlox Note Added: 0022599
Apr 22, 2012 1:57 pm Purlox Note Edited: 0022599
Apr 22, 2012 2:06 pm tigersfan Note Added: 0022601
Apr 22, 2012 6:54 pm martyn_van_buren Note Added: 0022604
Apr 23, 2012 6:32 pm khadgar Note Added: 0022618
Apr 23, 2012 6:36 pm tigersfan Note Added: 0022619
Apr 24, 2012 6:18 am GUDare Note Added: 0022634
Apr 24, 2012 6:19 am GUDare Note Edited: 0022634
Apr 24, 2012 6:19 am GUDare Note Edited: 0022634
Apr 24, 2012 8:35 am Chris_McElligottPark Note Added: 0022640
Apr 24, 2012 9:18 am Penumbra Note Added: 0022643
Apr 24, 2012 9:31 am Chris_McElligottPark Note Added: 0022645
Apr 24, 2012 9:44 am Penumbra Note Added: 0022646
Apr 24, 2012 9:49 am Chris_McElligottPark File Added: AVWW Upgrade Stones Math.xlsx
Apr 24, 2012 9:55 am Chris_McElligottPark Note Added: 0022647
Apr 24, 2012 10:07 am Chris_McElligottPark Note Added: 0022648
Apr 24, 2012 10:17 am Chris_McElligottPark Note Added: 0022649
Apr 24, 2012 10:17 am TechSY730 Note Added: 0022650
Apr 24, 2012 10:19 am Chris_McElligottPark Note Added: 0022651
Apr 24, 2012 11:57 am Penumbra Note Added: 0022656
Apr 24, 2012 12:17 pm Chris_McElligottPark Note Added: 0022661
Apr 24, 2012 1:04 pm Penumbra Note Added: 0022666
Apr 24, 2012 3:48 pm martyn_van_buren Note Added: 0022670
Apr 24, 2012 3:51 pm Chris_McElligottPark Note Added: 0022671
Apr 25, 2012 1:56 am GUDare Note Added: 0022688
Apr 25, 2012 2:09 am GUDare Note Added: 0022689
Apr 25, 2012 11:07 am Chris_McElligottPark Note Added: 0022699
Apr 25, 2012 11:55 am Penumbra Note Added: 0022703
Apr 25, 2012 12:07 pm Penumbra Note Added: 0022704
Apr 25, 2012 12:44 pm Chris_McElligottPark Note Added: 0022705
Apr 25, 2012 1:02 pm Penumbra Note Added: 0022706
Apr 25, 2012 1:05 pm Penumbra Note Added: 0022707
Apr 25, 2012 1:10 pm Penumbra Note Added: 0022708
Apr 25, 2012 1:22 pm Chris_McElligottPark Note Added: 0022709
Apr 25, 2012 1:26 pm Penumbra Note Added: 0022712
Apr 25, 2012 1:41 pm Penumbra Note Added: 0022713
Apr 25, 2012 2:09 pm TechSY730 Note Added: 0022715
Apr 25, 2012 2:13 pm Penumbra Note Edited: 0022713
Apr 25, 2012 3:01 pm MouldyK Note Added: 0022724
Apr 25, 2012 5:06 pm Penumbra Note Added: 0022725
Apr 25, 2012 5:14 pm Penumbra Note Added: 0022728
Apr 25, 2012 5:17 pm Penumbra File Added: AVWWLogSkillLevels.xlsx
Apr 25, 2012 5:17 pm Penumbra Note Edited: 0022728
Apr 25, 2012 5:42 pm Penumbra Note Edited: 0022725
Apr 25, 2012 7:08 pm Chris_McElligottPark Note Added: 0022732
Apr 25, 2012 7:19 pm Penumbra Note Added: 0022733
Apr 25, 2012 7:19 pm Penumbra Note Edited: 0022733
Apr 25, 2012 7:21 pm Chris_McElligottPark Note Added: 0022734
Apr 25, 2012 7:25 pm Penumbra Note Added: 0022735
Apr 25, 2012 7:26 pm Penumbra Note Added: 0022736
Apr 25, 2012 9:12 pm GUDare Note Added: 0022742
Apr 25, 2012 10:19 pm Penumbra Note Added: 0022744
Apr 26, 2012 10:18 am Chris_McElligottPark Note Added: 0022754
Apr 26, 2012 10:28 am Moonshine Fox Tag Attached: grind
Apr 26, 2012 10:28 am Moonshine Fox Tag Attached: specializing
Apr 26, 2012 10:28 am Moonshine Fox Tag Attached: stones
Apr 26, 2012 10:28 am Moonshine Fox Tag Attached: upgrade
Apr 26, 2012 11:02 am Penumbra Note Added: 0022755
Apr 26, 2012 12:09 pm Purlox Note Added: 0022760
Apr 26, 2012 12:49 pm Penumbra Note Added: 0022761
Apr 26, 2012 1:29 pm Penumbra File Added: AVWWLogSkillLevelsUpdated.xlsx
Apr 27, 2012 4:44 pm Chris_McElligottPark Note Added: 0022864
Apr 27, 2012 4:44 pm Chris_McElligottPark Status considering => resolved
Apr 27, 2012 4:44 pm Chris_McElligottPark Fixed in Version => 1.002
Apr 27, 2012 4:44 pm Chris_McElligottPark Resolution open => fixed
Apr 27, 2012 4:44 pm Chris_McElligottPark Assigned To => Chris_McElligottPark
Apr 27, 2012 4:57 pm Penumbra Note Added: 0022869
Apr 27, 2012 4:59 pm Chris_McElligottPark Note Added: 0022871
Apr 27, 2012 5:01 pm Penumbra Note Added: 0022872
Apr 27, 2012 5:03 pm Chris_McElligottPark Note Added: 0022873
Apr 27, 2012 5:11 pm Purlox Note Added: 0022874
Apr 27, 2012 5:15 pm Chris_McElligottPark Note Added: 0022876