Jump to content

Questions about maddening and teasing some cool code


Recommended Posts

Teasing some code I've been working on but could someone take a look at this for me and give some feedback.

I've built a stat generator in python that uses the following as input (using a fan favorite for example):

Char Lysithea  
Class Start End
Noble 1 4
Monk 5 9
Mage 10 19
Warlock 20 29
Gremory 30 45

And gives you the chart below. It factors in class change, class base stats, character and class growths. All the goodness. Obviously this doesn't reflect the items you might use (might add that later?)

There are still some bugs to work out. But I'm getting there. A few questions:

  1. What level do people generally end game at in maddening?
  2. Do people generally make the min level targets for class change there?
    • Beginner at 5
    • Intermediate at 10
    • Advanced at 20
    • Master at 30
  3. Is making a tool like this available for public use something people are interested in?
  4. Notice anything off?

 

  class             hp          str        mag        dex        spd         lck        def       res        cha
1 Noble 22 4 11 7 7 4 3 4 5
2 Noble 22 4 11 8 8 4 4 5 6
3 Noble 22 4 11 8 8 4 5 6 6
4 Noble 22 4 11 9 8 5 5 6 6
5 Monk 22 5 12 9 9 5 5 7 6
6 Monk 22 5 13 10 9 5 5 8 6
7 Monk 22 5 14 10 10 6 5 8 7
8 Monk 22 5 15 11 10 6 5 8 7
9 Monk 22 6 16 11 10 7 5 8 7
10 Mage 26 6 18 13 11 8 5 10 7
11 Mage 26 6 19 14 12 8 5 10 7
12 Mage 26 6 20 15 12 8 5 10 7
13 Mage 26 7 21 15 13 8 5 11 8
14 Mage 26 7 21 16 13 8 5 12 9
15 Mage 26 7 21 17 13 8 5 12 10
16 Mage 26 7 21 18 13 8 5 12 11
17 Mage 27 7 22 19 14 8 6 13 12
18 Mage 27 8 23 20 14 8 6 14 12
19 Mage 27 8 24 20 14 9 6 14 12
20 Warlock 31 9 27 21 15 10 12 19 12
21 Warlock 32 9 28 22 15 10 12 19 13
22 Warlock 32 9 29 23 15 10 12 20 13
23 Warlock 32 9 29 24 15 10 12 20 13
24 Warlock 32 9 30 25 16 10 12 20 13
25 Warlock 32 9 31 26 17 10 12 20 13
26 Warlock 33 9 32 26 18 11 12 20 13
27 Warlock 33 9 32 27 18 11 12 20 13
28 Warlock 33 9 33 28 18 11 12 20 14
29 Warlock 33 9 34 29 19 11 12 20 14
30 Gremory 34 9 36 31 19 12 13 20 16
31 Gremory 34 9 37 32 19 12 13 20 16
32 Gremory 35 9 38 32 20 12 13 20 17
33 Gremory 35 9 39 32 20 12 14 21 17
34 Gremory 35 9 40 33 20 12 14 22 17
35 Gremory 35 9 41 34 20 12 14 22 18
36 Gremory 35 9 42 35 21 13 14 22 18
37 Gremory 35 9 42 35 22 13 14 23 18
38 Gremory 36 9 43 36 22 13 14 23 19
39 Gremory 37 9 43 36 23 14 14 23 19
40 Gremory 37 9 44 36 23 14 14 23 20
41 Gremory 37 9 45 37 23 15 14 23 20
42 Gremory 37 9 46 38 24 15 14 23 20
43 Gremory 37 9 46 38 24 16 14 23 20
44 Gremory 37 9 47 38 24 16 14 23 20
45 Gremory 37 9 48 39 25 16 14 23 20

Ideally you'd run the input 100 times and generate statistics on "There's a 90% chance you'll be in this range". Honestly that's a lot closer now that the generation piece is already done. Still big thanks to all who deliver content for Serene Forest 3 Houses.

Edited by Sir_Jamsession
Link to comment
Share on other sites

How is the table you posted generated? Is this the result of one run, doing random level-ups based on the growth rates? I definitely think that averaging over a bunch of runs would be nice, and possibly a deterministic mode where you just treat the numbers as floats and add the fractional growth rates.

Without knowing how exactly you generated the table, it's hard to know if there's anything off, but looks ok at a glance. I'd double check how you're handling the threshold levels where the unit changes class (for example, you post 5 Monk but what actually happens is 5 Noble -> 5 Monk), without looking at the code I can't tell how it's being handled, but seems like an easy bug to miss.

As to your questions:

1) Haven't finished maddening yet, but people seem to say something in the 35~40 range for most units, with perhaps your main ones a bit above? Hard to say, as it depends on how much people grind and do auxiliary battles etc, and it doesn't matter as long as you allow people to choose which level to cap on.

2) This also depends. Ideally you want to advance to a class with higher growths as soon as possible, but there's two common situations that can mess with that:

    2a) Normal fluctuations in gameplay: in my current Maddening run, my Byleth was lvl 4 at the start of the second month mission (red canyon) and lvl 7 by the end of it, as she was one of the few that could survive most of the combats and saw most of the enemy phase action. I could have maybe retried the mission or went with a very stall heavy strategy to make her end at exactly 5, but I think most people won't bother.

   2b) Mastering classes. It seems that with the Maddening reduced XP you can, for the most part master one class before moving on to the next (maybe with the exception of Byleth and the main lord, as they see a lot of combat and have boosted growths), but if you happen to be missing a little class XP  after reaching the lvl for class change, you might end up overshooting a bit on the next mission. After, for the more advanced classes, you might want to master 2 or more classes in the same tier (for example, to get Darting + Death Blow). Again, there will be some normal variation from run to run.

3) I would be interested!

4) See starting comments.

Link to comment
Share on other sites

I can answer some of the questions. This is in context of a BL run which should have the highest levels in general since it has the most chapters. 

1) My highest level was Dimitri at 43. My average level was 37 across all my units. 

2) If you’re prepared beforehand and don’t spread out your resources too far, most units should hit skill level requirements. Certain Advanced classes are probably the most difficult to hit (those with a single A rank), but it can be done. 

3) Sure. I do think it’s probably simpler to just use a fixed-growth system over an RNG system, but running the current system over a large number of runs would still work. 

4) Looks fine to me.

Edited by LegendOfLoog
Link to comment
Share on other sites

On 10/25/2019 at 4:30 PM, LegendOfLoog said:

I can answer some of the questions. This is in context of a BL run which should have the highest levels in general since it has the most chapters. 

1) My highest level was Dimitri at 43. My average level was 37 across all my units. 

2) If you’re prepared beforehand and don’t spread out your resources too far, most units should hit skill level requirements. Certain Advanced classes are probably the most difficult to hit (those with a single A rank), but it can be done. 

3) Sure. I do think it’s probably simpler to just use a fixed-growth system over an RNG system, but running the current system over a large number of runs would still work. 

4) Looks fine to me.

Thanks for the info!

Link to comment
Share on other sites

There's a gimmick I take advantage of in the game, that is getting classes for the stats, like Lysithea getting myrmidon, mercenary and mortal savant to raise strength and defense, which she doesn't actually grow on her own.

With axe users, getting armor knight for some extra defense is a nice touch early on, which usually takes very little effort.

Link to comment
Share on other sites

At the start of my Silver Snow endgame my Byleth was a level 43 bow knight. A lot of characters ranged from 38-42. One thing I did was abuse the garden for speed carrots and used the kitchen every week in part 2. This made it possible to have a handful of units double most enemies. Just don't ever expect to double a flier or sword based enemy class. They require massive investment to not even get doubled by them.

https://imgur.com/a/1Mb8lTu

I don't know how your program does promotion but changed classes quite a bit on my Byleth. A very convoluted path because difficulty in training bow rank.

Just note that I have +4 speed +3 dex/def on these units from the kitchen. My Byleth actually has 33 speed.

Link to comment
Share on other sites

I'd be interested in occasionally using a tool like this. Here's some things it'd be nice if it could do:

-interpret starts and ends at the same level as class changing to adjust base stats without leveling up in the class. Ex: Swordmaster Start 34 End 34 for Lysithea to change her Str to 17.
-show average stats
-(following the above) show decimals just to tell if it'd have been likelier to get a low-growth stat up by then
-include enemy class growths
-have class stat boosts as an aside in parentheses

I think sticking to average stats would be best since those take more work to generate for this game than previous ones, but if you also have a random mode that could be fun to check. You probably already noticed (since the sample table seems to show it), but if there's a random mode, it should consider non-Byleth faculty/knights have a minimum of 0 stats per level up and the rest have a minimum of 2. An average mode can ignore this since everyone's growth totals are over 200 and the decimal places keep carrying over. 

Link to comment
Share on other sites

1. In Crimson Flower the units that I actually used:

  • physical units: level 37-39
  • magic users: level 37-41
  • dancer: level 44

2. Usually my units can be promoted asap, but there are three major cases in which the class change does not happen at the first possible level:

  • 2-3 level-ups in one map (happens about 5% of the time I'd say)
  • advancing to advanced classes with A-rank requirement (Swordmaster, Warrior, Sniper, Grappler, Warlock, Dark Bishop, Bishop) and the master classes which are harder to transition into (Great Knight, Bow Knight)
  • training a unit in more than one skill that they have no proficiency in

 

A tool like this would surely be interesting though I doubt I'd actually use it before infernal difficulty. (and I'm not even sure if I'd actually want to play on that difficulty lol)

 

Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

 Share

  • Recently Browsing   0 members

    • No registered users viewing this page.
×
×
  • Create New...