How do you go about database work?

[vanillacoke]

Well I'm aiming to make a 40+ hour game with lots of loot, weapons, armour and the such.

Combining this with character and class stats, skills, monster stats - it seems that this is going to take a lot of work and some trial and error. So I've been wondering, is there a specific way you go around the database work. I'd especially love to hear from RPG Maker veterans.

 

[Soma]

Mm..I normally just hack away at the thing without a set plan in mind. I pick something I'm interested in at that time, like items for example and move onto the next tab whenever I feel bored~ :3

 

[Espon]

Add it as you go.  I find it much easier to balance.  If I have a general idea how how many of a certain item I want I'll just leave the stats blank or at a rough estimate until I'm actually at that point in the game in development so I can properly test it.

Prep the database before you start adding things.  For example, I would reserve like 30 spaces for body armor, and even then I'll also have it split up light and heavy.  That way later on your inventory will look organized after you own a lot of items instead of having your Mega Potions buried under a bunch of junk.  Also make sure you leave plenty of room.  Better to have 5 empty spots in your database instead of realizing "Oh shoot I can't add anymore swords because the gophers are in the way."

I also try to do the same with enemies by trying to keep all enemies of a particular dungeon or area together.  Just makes it easier to balance stats and when creating troops.

 

[vanillacoke]

Thanks for the feedback guys.

I already started by sorting my swords and stuff with no stats.

I'm thinking of organising my monsters into groups also, like lvl 1-5 (lvl 5 boss) lvl 6-10 (lvl 10 boss) and so on. Earlier when I was working on it, it started to overwhelm me. But now I'm at work and thought of it that way. So it's nice to see someone uses that method.

Do you just use test battles to check it's all in order or would you recommend having a playthrough?

 

[kerbonklin]

You definitely need to do playthroughs as you go, to check how much EXP/Gold the player would get from battles, treasures, buying things, how difficult battles are, any kind of grindings that may need to be lessened, etc.

When doing tests for specific parts, have an NPC ready with all needed variable/switch setups, whatever items you might need, level boosts and gold, etc. so you can use Set as Starting Position (player) and then bring the NPC there. Just make sure you erase it afterwards.

You can pretty much do that for every dungeon or two that you finish.

 

[Kalithro]

I plan to build my database up in Excel and calculate everything ahead of time. Hitting a hotkey and waiting less then a second for results is faster than doing the play testing. Will do things such as static effects as a final tweak once the data has been populated in RPG Maker.

Then I plan on creating an option to test single variables vs many. X player X weapon/armor X skills Xclass vs Xenemy, go level 1 to 99, view damage chart... when does player die? When does Enemy die? Etc...

I like to plan things out.

 

[kerbonklin]

Trust me, even though you can have things like fancy charts, those don't show the practical side of things, which can be VERY different in video games. Especially since luck is always a factor, and players may be either too unfortunate or too lucky.

You know what they say, in a predetermined world, miracles can happen~

 

[vanillacoke]

I have to say I'm on Kerbonklin's side of things. What better way to test the database than to do it manually. I'm making skills based on agility and luck stats and it seems more sensible to try them manually.

I'm sure using Excel has it's benefits timewise, however, hard work and manual labour may give the best results. Excel doesn't know the game engine quite like the software itself does.

I'd be interested to hear more on using excel, or indeed, any other ideas.

Thanks guys!

 

[Espon]

I plan to build my database up in Excel and calculate everything ahead of time. Hitting a hotkey and waiting less then a second for results is faster than doing the play testing. Will do things such as static effects as a final tweak once the data has been populated in RPG Maker.

Then I plan on creating an option to test single variables vs many. X player X weapon/armor X skills Xclass vs Xenemy, go level 1 to 99, view damage chart... when does player die? When does Enemy die? Etc...

I like to plan things out.

Unless your game is completely free of RNG and the player has no choice in what they do, a simulation will never been as good as actual testing.

Through testing, you might discover a tactic that completely throws the game out of balance, something you won't even catch through a simulation.

RNG can be particularly tricky because even though on average everything looks fine, there might end up being incredibly frustrating scenarios that might occur.  Players usually don't like it when a series of random unfortunate events gives them a gameover.

 

[Andar]

I plan to build my database up in Excel and calculate everything ahead of time. Hitting a hotkey and waiting less then a second for results is faster than doing the play testing. Will do things such as static effects as a final tweak once the data has been populated in RPG Maker.

Then I plan on creating an option to test single variables vs many. X player X weapon/armor X skills Xclass vs Xenemy, go level 1 to 99, view damage chart... when does player die? When does Enemy die? Etc...

I like to plan things out.

Under one additional aspect I would agree to this plan:

As others pointed out, you'll still need a lot of playtest to check out the random elements, but the excel sheet helps in getting everything into the same balancing range.

If you do not use such a plan for your balancing but set the starting values before playtesting more or less at random, then you'll need double the playtesting to get everything balanced.

So no, the planning alone will never do in balancing, but it will help and reduce the time needed to balance the game.

 

[Kalithro]

[SIZE=medium]Organization and math will lead to a well balanced game. a.atk + a.luk for a player with 6 attack and 12 luck will always be 18 dmg. If you set your actor to have 1-99 levels and set you attack from 6 to 45 and luck from 12 to 50, then at level 50 you will do 57 +/- 1 dmg. If you change the formula to (a.atk * 2) + (1.2 * a.luk) / b.def ... At level one vs a slime with 4 defense you will do 15 dmg. At level 50 you will do 61 +/- 1 dmg. I can take this info and add in weapons then hit one button and test all of my weapons. [/SIZE]

[SIZE=medium]Doom Slayer Sword of 100 attack for a level 1 attacking a slime now does 223 dmg. Probably don't want this item in early game. Testing vs all weapons lets me know that a weapon of no more than X value will get me to Y number of attacks before a slime dies. I may not want to 1 hit kill a slime. Maybe for my game I want 3 hit.[/SIZE]

[SIZE=medium]Let's say I make a skill named Bash. I give it (a.atk + a.luk) * (1.2 * a.level) for the same player. At level 1 he does around 22 dmg (nice improvement over 15-18 dmg). At 50 he does 3420 dmg. Wow that isn't even close to 57 or 61. I need to balance this one.[/SIZE]

[SIZE=medium]With what I'm doing, I am able to set my constants and then test a range of variables. Like level 1-99 for my Main Actor, Class Fighter, Skill Attack, vs Mr. Slime or Dragon with nothing equipped. Sure things are going to change in game... But developers control when and what happens.  [/SIZE]

[SIZE=12pt]Sure you will have to do testing in the game, by no means was I rulling that out.  I find this approach much better then add skill test.  Add class test.  Add weapon test.  Tired of testing, add 5 classes, 2 skills, and 9 weapons; test.  Back to the first class test, did anything break?  This to me is simply too much extra effort back tracking.[/SIZE]

[SIZE=12pt]Luck in games is still math.  0.00001% drop rate is still 0.00001%  drop rate.  If I make Doom Slayer drop off a slime at a 0.00001% and someone gets it, thuse making the game way unbalanced; then it my stupid falut for setting that drop on a slime in the first place.[/SIZE]

[SIZE=12pt]Like I said, I like planning things out.  I plan on sharing what I come up with when I'm done.  Better shown than written, I guess.[/SIZE]

 

[Engr. Adiktuzmiko]

Luck in games is still math.  0.00001% drop rate is still 0.00001%  drop rate.  If I make Doom Slayer drop off a slime at a 0.00001% and someone gets it, thus making the game way unbalanced; then it my stupid fault for setting that drop on a slime in the first place.

I'm gonna hunt slimes in your game... wahahahahaha!

Anyway, GOOD planning will help in reducing the possibilities of faulty tests and reconfigurations...

 

[vanillacoke]

It is good to see a variety of opinions,and there isn't really any right or wrong answers.

I'd be interested to see video comparing the two methods and seeing if how they both compare.

 

[Ksi]

Make sure you leave a lot of white space between different things where you might want to later make new items/weapons/armour/spells.

I usually start with a list of things I'm sure I want. I then leave about 10 spaces or more between the healing and damage/misc items so that if I want to add something later on I don't need to worry about reordering the items or redoing all the treasures. This way all items are in groups. Usually something like:

Healing Items

Damage Items

Quest Items/Other Items

Weapons

Armour

In regards to weapons and armour, I like to keep them in groups as well, again with space between them just in case. So, Swords are all together, then Bows, daggers, spears, etc. That makes it easier for me when giving out items, too - this is the sword list, this is the bows...

And with armour I like to keep them again in groups. Light Helms, Heavy Helms, Light Shields, Heavy Sheilds, Light Body, Heavy Body, Accessories: Magic, Accessories: Special...

again with enough space for extra and revisions.

Why? Because I've made many a game where the inventory was mixed up and until we can get a sort feature up in this place, it's best to keep the items that are going to be used all at the top of the database for easy-access. The players will thank you for it, especially when it comes time to look through a huge inventory for that one sword they want to sell, or for an antidote.

 

[Zechnophobe]

Good math is important for sure, but also consider your border cases. If someone DOES get really lucky with something (or unlucky!) How will that impact your game?

As an experienced GameMaster, I know that the worst breaks in gameplay happen when the villains get lucky. Oh the dragon just crit the party and rolled max damage... uh, everyone is basically dead. That's way more damaging to the fun of the gamer than 'oh you sailed through that one fight easier than normal.'

For progressive database work, I feel like making a skeleton first is the best idea (though not one I've adopted yet). If you *know* you want to have 10 swords, for example, fill them into the db, pad a little, then start your 'axes' or 'staffs' or whatever. If you do the planning right you'll get where you need to be and can just fill in what needs filling in.

 

[Solo]

When I was creating my game, I made everything as I went along. I did not want or need the stress of constraints to inhibit my creativity. There was no way I could have said "I want 10 swords" and limit myself to that. Besides, late in development, I added tons of things I couldn't possibly have foreseen in the initial stages of development, that the game wouldn't be nearly as good without. And I would never consider reordering things in the database and having to totally redo treasure, encounters, etc... not after I completely ruined an older project like that. It's not worth it.

Some people may think my way is unorganized, but it worked for me.

 

[Espon]

The organization is more for the player, but it also helps you locate things quickly if you need to change something.  It's annoying trying to find an item and your inventory looks something like:

Spoiler

PotionLump of Something

X-Potion

Dungeon 2 Key

Ether

Insect Wing

Questionable Pie

Hi-Potion

Elixir

Phoenix Down

Dungeon 1 Key

Magical Rock

STR Fruit

etc.

when it would be better if it looked more like:

Spoiler

PotionHi-Potion

X-Potion

<Potion Placeholder>

<blank>

<blank>

Ether

<Ether 2 Placeholder>

<Ether 3 Placeholder>

<Ether 4 Placeholder>

<blank>

<blank>

etc.

Plenty of room to add things if you need to and it just makes it easier for the player to locate stuff.

 

[T.Bit]

Keep an Excel spreadsheet of all the stuff. It's easier to edit and sort the different armor, weapons, items, skills, etc. You can have stuff in any old order in the database and periodically update it in accordance to the spreadsheet.

 

[Chaos Avian]

I agree with organization. Organization is a MUST. Not just for the player, but for the developer as well. I used to do databasing in any order and for Skills especially the database becomes a waking hell.

For skills I have general categories, i.e. Healing, Magic, Damaging, Status, Special, and maybe character/ boss unique skills. Items are usually Healing, Damaging, Parameter increases, Event and Enemy Drops. I even have an ExCel spreadsheet on one of my databases. For skill balancing I do minimal math at that and tend to leave the variance at 20 except for very specific skills and whatnot.

 

[Zechnophobe]

The difficult part of doing this in RPG maker, is that you can't always move things around in the DB without having to make changes elsewhere. Ultimately, once something is in there, it's a pain to move. For people doing certain types of events and scripts (where you refer to an item/skill/actor by ID) you may have to do a lot of updates just to move 'dark sword of doom' to a different slot. This is why I strongly recommend doing that as little as possible.