Auto-Level-Up straggling defenders

#1

Hey guys, Lars here. Anthony and I have been planning a feature for an upcoming patch and I’d like to solicit some feedback on it.

Basically, some people have mentioned a problem with units that fall behind in experience, either from not being used in battle, or from being recruited late in the game. The problem is, this kind of gets worse as time goes on - the less you use them, the further they fall behind, and so the less you use them.

We’d like to solve this somehow. Anthony and I have been kicking around some ideas for a while, but I’d like to compare notes real quick before we make a final decision.

First, let’s talk about what the guiding principles are.

1) Units should not all wind up at the same level
2) Straggling units should be able to catch up
3) You should be rewarded for using units in battle
4) Big-party and small-party should both be viable strategies

The original hypothesis was that the exponential XP curve would mostly solve this problem. Since it takes more and more XP to level up, a unit that’s ahead gains progress much slower than a unit “catching up” to them, especially if you’re earning large amounts of XP. Generally this seems to work, but it’s not enough to deal with units that are really far behind.

Also, units that you don’t use in battle still get XP, I think by default it’s just 50%. You can tweak this in the accessibility menu if you disagree, and just make everyone get 100%, of course.

It all kind of falls down on number 4. Recruiting level 1 units early in the game isn’t so bad, but by the time you get to NG+, if you haven’t been recruiting people they take forever to get up to speed, which makes them near useless.

[size=14]Proposed solution[/size]
So, I think the root of the problem is recruiting units at level 1, no matter what. It wasn’t always this way - originally, the algorithm was to recruit units at (lowest level unit of that class) minus one. This wound up being a little weird, though - if you recruited multiple units at once, they’d all cascade down. Still, I feel something like that would be better than what we have now.

So, there’s two angles of attack:
1) Fix recruiting so units are at a “reasonable” level to start with
2) Have some option to “train” straggling units up to some maximum level in exchange for scrap

So, just off the top of my head, a possible solution could be:
1) Freshly recruited units start at 25% of Azra’s level (or the highest level unit of that class)
2) Old units can be buy levels in town with scrap through “training.” Up to a max of 75% of Azra’s level (or highest level unit of that class)

Pegging the formula to Azra, or the highest level class unit, avoids the “cascade” effect from going off the lowest-level unit. I think this could be a good starting point, and Anthony and I will work out the details. Do you guys have anything to add? I want to make sure I’m actually solving the right problem before I make a major change like this.

#2

Could you make it so that units you buy have ~80% of the experience they would have had if you had used them for normal battles up to the store you bought them at?

So, units bought at the Guard’s Hideout would have 80% of the experience for the Normal Gold Stars up to that point.

#3

Laureola - that solution is a bit more complicated. For example, are you talking about a play through on Normal, or Extreme? Have you done any grinding so far? There are a few variables, and ultimately we want to keep this as more of a simple heuristic. My preference is to take the second approach and provide a “Trainer” in towns who can level up your characters to a certain amount. This could be to a percentage of Azra’s level, or potentially to a max level of the Trainer’s skill (so the first town Trainer might only be able to train up to level 5). I lean towards the former for its flexibility, even though the latter makes more sense from a story/world perspective.

#4

I was thinking just the exp for Normal Gold stars up to the store you buy them from, no Advanced or Extreme stars exp.

So, it’s the exp total for if only did each mission once on normal and gotten the Gold star. I was going for if they made to x store they must have been able to get though x areas with stronger monsters, like your max Trainer level idea.

#5

I’d get rid of the training idea - and just have recruits either based on Azra’s level, or maybe directly determined by highest unlocked town or the town you’re recruiting at. The latter would be solved easily simply by adding a parameter in maps/index.xml to the recruit tag - i.e. .

It has some downsides (fixed levels) and some upsides (more modding potential). And the initial recruit levels for stock game could be more or less based on standard linear playthrough, making the recruits about 1-2 levels “behind” that par-of-the-course, which is easy to obtain).

Sure, somebody going for all gold stars as soon as possible will be about level 23-26 near endgame, so for him new recruits level 13 wouldn’t be so great - but then, the challenges he faces aren’t so tough, either.

And the big issue about level 1 recruits is really that they totally suck. This becomes painful once a party with 1 healer discovers dark creeps. Provided that you HAVE the scrap to recruit with, you anyway have to go back and grind somewhere to make them level 5.

Otherwise, the gap in the levels as you progress isn’t that bad - the exponential curve makes sure of it that the longer you play, the closer the gap is in levels. My level 27 party which was used for linear play in normal and NG (and has 54 gold stars, missing only the battles from Hand onwards in NG, has 4x27 (just so) and 2x 24 (almost 25) zerks, 4x26 and 2x24 rangers, 3x25, 2x23 and 1x22 healers, 3x24 and 1x23 knight, 25 Wrenna and 23 Niru - and the weaker zerks and rangers were recruited when I started NG+ - so as you can see, the difference between the characters has dropped from 16 to about 2.5 levels in just 11 levels for the original 16s.

#6

Good points everyone. The reason I like the Trainer idea is that it’s another way to choose to spend your gold to keep the economy in the game a little more interesting. It can seem like an absurd penalty though (why do I have to choose between grinding for gold and grinding for xp?), so the idea of recruiting at 50% of Azra’s level (for example, exact number is likely between 50% and 75%) might be the simpler and kinder option. It also avoids a lot of trouble with balancing prices, adding a new shop interface, looking for exploits, etc., and does a pretty good job of mitigating the level difference. So the Trainer appeals to me on a game world and economy perspective, but simplicity and sanity may just have it be a % of Azra’s current level (or current XP, which would produce a slightly different curve).

#7

I like you idea, but personally I would really tie it to Azra’s EXP and not her level. When at later levels the EXP needed for a level up are increasing more and more, it’s just a pain if, also the exp they got is worth 10 to 15 lower levels, Azra only got 1 or 2 more levels and the new units get no level up at buying at all… (was that sentence ecen unterstandable? -.-

So, i suggest you make it to something between 60% and 75% of Azra’s Exp.
Maybe tie the scrap cost for level up to Azra’s Exp as well and instead of buying levels, you buy Exp for them. If Azra has more exp, the low level units would get more Exp per scrap as well ^^

#8

I’m not sure I see the value in being able to buy levels. If it’s a trivial cost, then the system really doesn’t add anything beyond flavor. If it’s a meaningful cost, a player who’s going for a full team won’t be able to afford it (or, alternatively, will find himself gaining enough levels while grinding for scrap that he doesn’t need to use the system).

Personally, I think the root of the problem is the idea that defenders gain less xp if they’re not in the battle at the end. I’ll admit to finding it moderately irritating to need to summon everyone by the end of combat to avoid them falling behind, and very irritating if a defender dies right near the end of the battle and I don’t react quick enough to get him back in.

#9

@Valint:

You can make all defenders gain 100% xp regardless of whether they’re in battle or not, check out the accessibility options!

#10

It’s the usual problem in RPG games. Give XP to everybody when anybody in the party kills an enemy? Or reward individual actions (and how to make sure that support classes get roughly the same amount of XP as raw damage dealers?)

I went for the simple “no absence penalty” exactly because I hate the ridicule of pausing the game towards the end just to recall a strong defender and summon everybody with the freed PSI. I don’t see any good solution to this, especially one that’d be worth the developer effort behind it.

#11

Yup, we debated this extensively and still didn’t come to a solution that we really like. I specifically want there to be some variation - otherwise any time you hire two people at a time they will be perfectly in sync when leveling up and I feel like that looks and feels weird and artificial. Of course, so does suddenly summoning all of your stragglers.

Vallint has a good point - either pay-to-level is trivial or it’s too expensive to make sense. That’s why I’m leaning towards just getting them at some percentage of Azra’s current status. It’ll simplify the situation, it’s easy to code, and it avoids the oddities of paying to train.

#12

I like the idea of them being scaled to your level, but I have some caveats. It’s dragons and knights that you have the most issue with. Dragons in particular come very late, and this may well contribute to them being seen as useless.

It might be worth giving dragons and knights more of a connection to Azra’s level- 50% say. That way if you hire dragons you’ll quickly have a solid cadre of strong dragons which you can use effectively.

#13

What I think would be an interesting way to implement scaling would be to have different levels of recruits available based off of the number of stars you have. For example if you have 20 stars allow recruiting of defenders that have 30% of Azra’s level at 50 stars allow recruits that have 45% of Azra’s level and so on. The numbers that I am throwing out are not really meaningful at all, they are just to serve as an example.

One other way would be to have the number of kills that class has received affect the recruitable level: For instance I have 10,000 kills using rangers therefore I can hire rangers at X% of Azra’s level.