RPG balancing (split from Amber's Beta thread)

Games with combat, inventory, crafting and more beside a story and dating/life sim gameplay
Post Reply
Madance
Druid
Posts: 301
Joined: Tue May 15, 2012 9:41 pm

RPG balancing (split from Amber's Beta thread)

Post by Madance » Sun Apr 23, 2017 4:50 am

jack1974 wrote:Yes definitely! Though in this last update, reducing daily cap to 5 and slightly increasing crafting time, the young age should be almost back to where it was. I'd say people crafting madly could end around 150-160 skill which is OK.
In any case, if you were able to win in the old version, you'll be for sure in the new one. As for playing the game more than once in crafting mode... ahem! Now that I think about it though, even RPGs are rarely replayed from scratch (though I know people who played SOTW two or three times, but I think they're mad :lol:).
I played SOTW 6 times in a row, and no, NO VISUAL NOVEL MODE WAS USED. :twisted:

...I really love that game. :oops:

User avatar
jack1974
Pack leader
Posts: 13625
Joined: Thu Jun 16, 2005 4:43 pm

Re: Amber Full Beta 0.8.8

Post by jack1974 » Sun Apr 23, 2017 6:38 am

Haha thanks. Anyway, that proves my point: at least with a RPG (as long as I don't make the mistake of having randomized missions again like in QoT, and that WON'T happen!) you're sure that players will test it :)

Troyen
Elder Druid
Posts: 857
Joined: Fri May 06, 2011 2:23 am

Re: Amber Full Beta 0.8.8

Post by Troyen » Sun Apr 23, 2017 7:54 am

Madance wrote:
jack1974 wrote:Yes definitely! Though in this last update, reducing daily cap to 5 and slightly increasing crafting time, the young age should be almost back to where it was. I'd say people crafting madly could end around 150-160 skill which is OK.
In any case, if you were able to win in the old version, you'll be for sure in the new one. As for playing the game more than once in crafting mode... ahem! Now that I think about it though, even RPGs are rarely replayed from scratch (though I know people who played SOTW two or three times, but I think they're mad :lol:).
I played SOTW 6 times in a row, and no, NO VISUAL NOVEL MODE WAS USED. :twisted:

...I really love that game. :oops:
I had to replay Act 1 twelve times during the beta on nightmare. I still haven't recovered. :x

User avatar
jack1974
Pack leader
Posts: 13625
Joined: Thu Jun 16, 2005 4:43 pm

Re: Amber Full Beta 0.8.8

Post by jack1974 » Sun Apr 23, 2017 8:04 am

That was another mistake. I mean, of course it's not like I expect to not do any tweaks, but all those updates/balancing etc (especially on something like act1 of a big game, where players wouldn't be able to see the whole picture), were really worth it? I don't think so. There needs to be a line drawn between "improving a game substantially" and tweaking/doing updates just for the sake of it :lol:

Troyen
Elder Druid
Posts: 857
Joined: Fri May 06, 2011 2:23 am

Re: Amber Full Beta 0.8.8

Post by Troyen » Sun Apr 23, 2017 11:30 pm

For an rpg, I've noticed many of the MMOs these days focus on skills and variety of skillplay first, then specific numbers on mob fights/damage. I think we were trying to do everything at the same time, which required rechecking every fight to see if it was beatable (or way too easy), and also re-evaluating the experience earned to ensure we could be at a reasonable level for the major story fights.

Also, we're black box testing where we have to do a lot of playing from scratch which is good for the "actual player" experience, but inefficient when trying to evaluate tuning or specific encounters. I don't know if you have private testing tools, but for encounter tuning, etc., it might be more useful to have the ability to start in the middle of the game (fresh, not via save file) so you can say "I want to try out the Deceit fight" and then assemble your party at the pre-defined intended level without having to play the first three acts again. The pacing is of course the trickiest part to nail down because there's a huge gap between 100% completionists and casual players and any feedback you get has to be evaluated against the type of player and the intended difficulty of the target difficulty level. I'm not sure what a good solution is there, outside of having rough checkpoints (e.g. level 6 at the end of act 1, level 12 at the end of act 2) and a table of estimated number of encounters to finish an act, and then compare that against metrics you collect in players' save files.

I'd need to think more about something like simulation testing. But maybe in the future it would be useful in all your games for the beta to add additional analytics to the save files like what order recipes were learned, how much studying/gathering/crafting someone did (in the young vs adult age), snapshot of the player's skills and level at various story checkpoints (say at the beginning or end of each romance/main plot scene), so you'd have more data available to look at when players finish the beta and upload their save files. Hopefully some of that would lead to less guessing or more targeted fixes - e.g. if the young age progression looks good, but between the 8th and 10th romance scenes the skill/level progression falls behind, maybe that means tweak the Tier 4 and 5 recipe benefits.

Course, this won't help this game now, but may help when designing a future game, as long as thoughts on the testing are included early enough in the game's design.

User avatar
jack1974
Pack leader
Posts: 13625
Joined: Thu Jun 16, 2005 4:43 pm

RPG balancing (split from Amber's Beta thread)

Post by jack1974 » Mon Apr 24, 2017 6:40 am

Yes, the future RPG framework had even "automated testing". Of course using computer AI, so inferior to human brain (usually! :lol:) but you could already see if a battle was too uneven or not. I tried to search the forums since I remember Anima posted a screenshot but I can't find it. Anyway, I saw it personally running! :)
Basically you could run 1000 automated battle so you could see the odds of team A vs team B of winning/losing. Very useful since I could use the tools to setup a "standard party" and see how many chances the AI had to win that specific battle. Then of course depends on the actual RPG rules, as said the AI is not perfect: but (unless there are some very unbalanced/random skills) the results are reliable, I could always say if boss encounter happening at end of chapter 1 was more difficult or less.
This sort of automated testing isn't possible with Loren's framework for example and it's a big step forward for sure :)
Maybe I split this discussion in another topic since it's not really about Amber beta :mrgreen:

Post Reply