Threads that make us Laugh, Cry, or Both, Continued

[Crissa]

[counturl=131]Warning: Severely not work safe.[/counturl]

Anyhow, it's a meme that's been around from the oldest cave art.

Also, that's a terrible method of playtesting. It's missing the whole first series of math that should be done - before players ever see the dice.

No amount of 'see the daylight' playtesting tells you anything about the math behind the system. This sort of playtesting shows you whether the interface of the game between the player and the DM was working.

...And this is only relevant if the DM and/or player have never seen the printed material before.

-Crissa

[Count Arioch the 28th]

1198124312[/unixtime]]Or things you never wanted to know: How to fuck a scorpion.

Well, the male scorpion breaks off his wang and sticks it in the ground. Then he grabs a female by the claws, and drags the female over his severed wang for insemination.

Seriously. I'm not making that up, I'm not that creative.

[Count Arioch the 28th]

1198103528[/unixtime]]

And, of course, the fluff assumption thats going to be built into the dwarven race (for no apparent reason): they were the slaves of giants, that eventually revolted and gained true freedom and shit.

They must have stolen that idea from my shitty homebrew. In my world, gnomes and dwarves were servitor races created by ancient jotun wizards, who broke free after a bloody revolution.

They developed the dodge bonus to AC because the dwarves and gnomes that weren't able to avoid getting stepped on usually didn't survive long enough to produce offspring.

[Username17]

Good to know that David Noonan doesn't know fuck all about what he is doing.

OK, let's say that you wanted to test D&D:

First, define your test sample. That's where you make a sample character of each class at each level. With 8 classes and 30 levels (apparently), that's 240 sample characters. Time yourself, because if this takes a very long time you'll want to go back and simplify. But we're writing the game right now, so some time will be spent discussing how things work or what options the sample tiefling warlock should have and such. You will want a like number of monsters. I suggest 12 monsters level 1-10, 8 monsters level 11-20, and 4 monsters level 21-30. This is because people will play more at low levels and people who are invested enough in the game to play to higher levels will be more likely to buy a Monster Manual II which has 4 monsters level 1-10 and 12 monsters level 21-30. In any case, you now have 480 opponents (since you can mirror match PCs as monsters).

Second, run through completely theoretical sample combats of one-on-one of characters vs. monsters and characters of other classes. You don't need other people at the table for this, just crank the damn numbers. Theoretical sample combats should initially be defined with "perfect tactics and position" meaning that noone gets hit by attacks of opportunity and everyone gets their attacks off. You want every character at every level to face every monster of their level (except themselves, as that's pointless except to demonstrate that they can hurt each other at all). That's 3600 one-on-ones you have to do, so get cracking. Fortunately, you can part this uphowever you feel like because it's all theoretical. Otherwise it'll take one tester 2 or 3 work weeks to crank through. Better yet, do this redundantly so you get ten or more sets of data and still finish in a week. At the end you should have identified a number of strong and weak abilities, classes, and monsters. Go back and revise for a week. Do it again.
Test time: 4 Weeks

Next, Group Combat theoreticals. Again, perfect positioning. And while you have now entered the realm of implausibly large numbers, you've also already eliminated most "problem" units. So what you'll want is to define some sample enemy groups. Probably 10 or so flag ship enemy monster groups. This is going to take a while, because you'll want to test each against possible arrangement of 4 different PC classes at each level. That's 504,000 combats, and each one will take a single person a few minutes to crank through. So you'll cheat and limit things to a more manageable level by simply assigning each battle a random monster assignment and testing each party compositon ar each level - reducing the task load to just 50,400. It should elucidate a number of over and under-synergies and demonstrate the structural limitations of defensive classes and such. Where your 30 man testing group gave you 10 data sets in a week, for this test they'll only generate one complete run through in a week and a half. Do it twice. Revise for a week, and do it again.
Total Playtest Time: 9 weeks

Next, Positional Tests Now you want to test maneuverability and position. This means that you'll want to divide your playtest group into six groups of five and have them run through these same combats in actual battlefields. Have each group play random assignments of monsters in standardized "close" and "open" battlefield conditions at level 4, 8, 12, 16, 20, 24, and 28. Remember to the have the higher level "close" battlefields filled with crazy crape pillars of flame and walls of force. Also remember to have the higher level "open" battlefields be big. You'll want to test each party configuration at each sample level, which puts you in for 11,760 test battles. This will take an unacceptably long time because you can probably only get in 8 such battle sims a day per group, and you have just 6 groups. Since you don't have 245 days to let this incubate, you'll have to cheat again. This time we'll do it by having each class represented in each level, with a random assignment so that each play group uses each class at each level. That's much easier, gives us just 14 battles per team, which means that we can crank through a development cycle with dodedecaredundancy every two days. We should do it 5 times, putting us out a total of 3 weeks including rewrites.
Total Time: 12 weeks

Next up: Combo Testing! We need to stress test this thing. People can find combos and whatever, but the people who have been playing with the sample classes aren't going to be all that good at some of it because while they have a good feeling for how good some of the abilities are in real play, they'll be inherently resistent to finding "broken" uses of abilities which are outside the character types "role" - your core group will find it difficult to find Cleric Archer loops and such. So it's going to want to be outsourced to crazy guys who break stuff. Like the Optimization Board would be if they didn't keep hitting people with the ban stick every time they raised a fuss there. Anyway, it's D&D, so people who are genuninely good at stress testing mechanics will do it. Every time someone finds Phoenix Duplication or Balor Mining or More Wishes or whatever, go back and revise.

And yeah, Now you can write up some initial galley proofs and submit them to outsourced playtest groups for legibility tests and campaign reports. You'll want to do some more revision after that.

So yeah, that's my 5 month proposal to systematically and rigorously test the D&D rules for 4e. If I had a May release date from right now, that is what I would do and I would have a quality product shipping out the door in time to do that.

-Username17

[Koumei]

That sounds like effort, planning and having a game ready to test though. I still wouldn't be surprised if they turned around and said "Okay, so we haven't really got anything. But we have decided to make the main die a d20 though!"

Is "New Edition" going to have that much testing? Wait, that's what we're all for, right? You and Keith argue until you get the system worked out, then the rest of us cover hundreds and thousands of playtest encounters?

[Username17]

Is "New Edition" going to have that much testing? Wait, that's what we're all for, right? You and Keith argue until you get the system worked out, then the rest of us cover hundreds and thousands of playtest encounters?

I wish.

Yeah, it would be stellar if I had an in-house full-time writing staff and 30 testers working 30 hours per week on contract for 1 dollars an hour. That would be fabulous. But realistically, that ain't going to happen.

Instead what we'll be doing is raising the resolution even further and creating subsystems which are functional in and of themselves and then plug-and-playing characters and monsters on top of that. So we're going to test nmerical ranges and then we're going to exhibit creative discipline where we don't write things which fall outside of those numeric ranges under any circumstances.

And then we're going to have people like us comb it over to make sure that there aren't combos that allow people to leave the specified numeric ranges. And then we're going to cross our fingers and hope. Which is sad, but it's what happens when you don't have the time or the manpower to do things "right."

---

World testing is easy. You write up the world and then you hold up all the monsters and abilities and check to see if any of them destabilize the world as presented. You can also grab a check sheet like this one:

• How do the individuals or interest groups who are in power stay in power? What forces are trying to overthrow them and why haven't they already succeeded?
  
  
• How stable is the world? How quickly will it change into something unrecognizable?
  
  
• Is the presented society consistent with characters existing who have the powers that the game says that they will get?
  
  
• If the end of the world is a serious possibility, why hasn't it already happened?

Pick up all the character and monster abilities and consider them in light of those simple questions. You'll probably have to toss or modify a lot of world and ability text.

-Username17

[SunTzuWarmaster]

As long as we are talking about playtesting:

Frank, have you seen my quasi-optimized Samurai?
http://bb.bbboy.net/thegamingden-viewth ... br][br]The key was 19 attacks/round with 31 attacks in a burst if desired, playing on the concept of:
Combat Reflexes (granted as a class ability)
Up to AoO # of attacks as immediate action (class ability)
High Dex (who wouldn't?)
TWF (3 extra attacks, 3 extra AoOs, who wouldn't?)

I have not playtested this yet (and the character probably won't make it to level 11), but this is... powerful.

[JonSetanta]

So, pretty much a playtest of:
Optimise
Break
Tweak
Rebreak

[fbmf]

[The Great Fence Builder Speaks]
The second thread in the series has gotten too long, folks, but feel free to start a new one.
[/TGFBS]