Tim Cain at Reboot Develop 2017 - Building a Better RPG: Seven Mistakes to Avoid

[TimCain]

 

 

But you could have solved the random encounters by giving them a chance to appear according to luck number. Instead of them starting to appear at 7 luck, you'd have 70% chance of a random encounter appearing & 60% chance at 6 luck etc.

 

We actually tried that at first. But with all of the back and forth people did on the world map, no one noticed a difference between a 6 or 7 Luck. You got all of the encounters if you kept playing long enough. So we switched back to thresholds.

 

This is why I said in my talk that critical hits should adjust damage and not chance. People notice when their critical hit damage goes up 10%, but they don't notice when they critically hit 10% more frequently. Mathematically they are the same in terms of DPS, but psychologically they are completely different.

[Azdeus]

I will admit that I don't get how having geometric shapes would help make any situation like that better, you would still end up with thresholds wouldn't you? It would just mean that you'd have to have "great" luck instead of 7 to get those random encounters.

 

On the crit chance vs crit damage, after raiding in WoW where they had both, I'd have to say I really never thought about the increased crit damage as much as I noticed the increase in crit chance. Going from 10% to 20% crit chance was really noticeable, but my strikes going from ~5k to 5.5k wasn't really worth noticing.

I'm weird I guess.

[Quillon]

So far:

 

Shapes accomplish less and clearer thresholds. / Which you can accomplish with numbers 1 to 5 and call them YUCKY-OK-GOOD... without the shapes.

 

They have built-in constraints. / And potential synergy combinations between different shapes & groupings, might be fun to play around.

 

And apparently humans understand shapes with more ease than numbers. / We'll see if it'll be true for this case.

 

 

 

I will admit that I don't get how having geometric shapes would help make any situation like that better, you would still end up with thresholds wouldn't you? It would just mean that you'd have to have "great" 

On the crit chance vs crit damage, after raiding in WoW where they had both, I'd have to say I really never thought about the increased crit damage as much as I noticed the increase in crit chance. Going from 10% to 20% crit chance was really noticeable, but my strikes going from ~5k to 5.5k wasn't really worth noticing.

I'm weird I guess.

 

I agree, I played Age of Conan back in the day, how often I crit was more noticeable than how much I crit for. I was happy if I critically hit for 200% damage, doesn't matter if its not 250%, its still a lot bigger than my normal hits.

 

It could depend on the type of game tho, might be different for a [turn-based]game where you have time to read how much you hit for on every attack.

[Gromnir]

am still complete missing the fixation with shapes.  gonna start with the recognition that no, not all people is visual learners. any teacher is gonna recognize the mistake o' assuming people all acquire and process information the same way.  the fundamental assumption o' the superiority o' visual learning is in error. visual will be best for many people.  not all.  probable not even most.

 

use fallout example of special also is curious.  if the gameplay difference between a special three intelligence and a four intelligence is more significant than the difference 'tween four and five, such an observation ain't a recognition of the fail problem with numbers but is rather a problem with how the developer used the numbers.  simple solution is to genuine provide the player equal value for every point invested in an attribute.  such is even more clear than shapes, no?

 

numbers is gonna be essential to virtual any crpg mechanics.  whether you present the player with shapes or colors or sounds is gonna simple be a way of expressing the numbers which is actual working behind the scene, so use numbers from the start seems most logical.  

 

temperature increase may be expressed with a volume change in a tube and color gradients and perhaps a steam whistle sound when ______ gets hot enough.  whatever.  no matter how you wanna express temperature, a chemist or baker or even a chef is gonna want those thermometer readings with exactitude.  fry chicken or bake a cake succesful can be achieved through trial and error, by response to visual, auditory and olfactory cues. why? certain chemical changes happen precise and predictable.  go ahead and express exact volume or weight of ingredients w/o numbers.  express temperatures w/o numbers.  complicates rather than simplifies.  

 

a crpg, combining conditions and qualities and circumstances is gonna make predicting success or failure of various actions at least as complex as is frying chicken. have us try and predict success or failure with lucky charms cereal shapes and colors does not strike us as particular intuitive. wanna add shapes and colors and sounds to the numbers for a more visceral experience and for those situations when exact is unnecessary?  sure.  is extra work, but am seeing the value of adding learning modalities (gonna need ask hurl if such is the correct nomenclature) but the numbers is the essential.  

 

dunno. am still missing something.  

 

HA! Good Fun!

[Bartimaeus]

This is why I said in my talk that critical hits should adjust damage and not chance. People notice when their critical hit damage goes up 10%, but they don't notice when they critically hit 10% more frequently. Mathematically they are the same in terms of DPS, but psychologically they are completely different.

 

Hm. Yeah, I'm just not totally sure how true that sentiment actually is - at least, on the scale you're implying. For very small number increases, it's going to be hard to notice a difference either way...but personally, I prefer the greater chance rather than greater damage: though mathematically it all evens out in the end, on a more micro-scale (e.g. a single combat encounter), I prefer a more consistent and predictable output, which means I prefer the greater chance to score a critical hit rather than the greater damage.

[Gizmo]

This is why I said in my talk that critical hits should adjust damage and not chance. People notice when their critical hit damage goes up 10%, but they don't notice when they critically hit 10% more frequently. Mathematically they are the same in terms of DPS, but psychologically they are completely different.

This reminds me of that certain types of fix-it software that once installed, spares no time in pointing out all of the many things it's detected, and/or fixed (or "fixed"), lest the customer doubt the product, and decide to uninstall it, or return it for a refund.

 

Regardless of the internal mechanics that Fallout had, I always interpreted a head shot that did shockingly little damage, as a grazed cheek or a knick to the ear; or some other technically valid hit, but not life threatening. Critical hits were the serious injuries to vital areas, and they were harder to pull off; and thus relatively rare.  My gut says to give people more credit... but I cannot disagree —the majority probably doesn't notice the 10%... But I would certainly notice (and not appreciate) for critical hits to be re-imagined as simply a measure of added severity rather than an exceptional shot.  The difference is subtle, but if critical hits predictably happen on a schedule, or come of a charged-up ability bar, then they become just another special attack, rather than an unexpected boon of circumstance, or weighted by exceptional character skill.  

Edited May 20, 2017 by Gizmo

[Infinitron]

am still complete missing the fixation with shapes.  

 

Don't think about it in terms of reducing complexity or increasing the ease of gaining a true understanding of the underlying mechanics. Think about it in terms of SPEED. Getting through character creation as quickly as possible. Shapes allow you to select the value of several stats with a single click. CLICK and off we go into the game.

 

I believe the basic motive here is that visually overwhelming, time-intensive character creation sequences scare people away from RPGs. It doesn't really matter so much how "difficult" the underlying mechanics are. The real problem is that you have to spend a significant amount of time comprehending and then manipulating a wall of numbers before you get to play the game, regardless of the true difficulty/complexity of the system they represent (although of course those things are correlated to an extent).

Edited May 20, 2017 by Infinitron

[Gromnir]

 

am still complete missing the fixation with shapes.  

 

Don't think about it in terms of reducing complexity or increasing the ease of gaining a true understanding of the underlying mechanics. Think about it in terms of SPEED. Getting through character creation as quickly as possible. Shapes allow you to select the value of several stats with a single click. CLICK and off we go into the game.

 

I believe the basic motive here is that visually overwhelming, time-intensive character creation sequences scare people away from RPGs. It doesn't really matter so much how "difficult" the underlying mechanics are. The real problem is that you have to spend a significant amount of time comprehending and then manipulating a wall of numbers before you get to play the game, regardless of the true difficulty/complexity of the system they represent (although of course those things are correlated to an extent).

 

"yeah, the learning curves should be flattened even in highly complex crpgs.  there is no genuine advantage to be dumping the entirety o' a new rule system 'pon a game purchaser.  if you got levels in your game, to be demanding a player make the mostest important and numerous character development choices at level 1 is bass ackwards. is axiomatic that the point o' the game where we will know the least 'bout game mechanics is at the beginning.  life may be unfair, but gamess don't need be.  developers should make more significant character generation choices occur later in the game rather than early. sure, there is gonna be fundamental choices made during the intro portion o' a title which will have major impact 'pon future gameplay.  race o' your character is likely gonna be decided at the start o' the game, and such a choice will and should be meaningful.  is gonna be multiple fundamental choices made early.  fine.  is all the more reason to avoid demanding more than necessary o' the starting player.

 

"am actual (kinda) liking tim's triangles as a starting point.  we want many character development choices, but we do not need all, most, or many such choices at the beginning o' the game.  is any number o' games Gromnir has played, games we liked, which had us restart after investing dozens o' hours into the title 'cause we discovered the character we believed we were building did not match expectations.  only way to fix fundamental flaws were to start over from the beginning.  the desire to restart and "get it right," is a common issue for us, and having read these boards for decades, we know we ain't alone.  is no reason the player cannot be confronted with more character generation/development choices after they gots a better understanding o' the game mechanics. have us make choices after we get comprehension o' how the game actual works makes far more sense than demanding the uninitiated and ignorant beginner decide the most important character development and generation choices w/i the first 20 minutes o' game experience."

 

repeating what we said at start o' thread.

 

we get the value o' a simple start, but as some kinda general rule o' game development for replacing numbers with shapes?  uh, no.

 

http://forums.obsidian.net/topic/92420-tim-cain-at-reboot-develop-2017-building-a-better-rpg-seven-mistakes-to-avoid/?p=1908891

 

HA! Good Fun!

Edited May 20, 2017 by Gromnir

[hilfazer]

On the crit chance vs crit damage, after raiding in WoW where they had both, I'd have to say I really never thought about the increased crit damage as much as I noticed the increase in crit chance. Going from 10% to 20% crit chance was really noticeable, but my strikes going from ~5k to 5.5k wasn't really worth noticing.

I'm weird I guess.

It would be 11%, not 20%. 10% * 1.1 = 11%. You are not weird, almost everyone gets it this way. That's because math is hard (if you are not MaxQuest).

 

https://en.wiktionary.org/wiki/percentage_point

[Azdeus]

It's old braindamage from playing WoW, noone spoke in mathematically correct terminology. A 10% increase when you speak WoW isn't going from 10 to 11 percent, it's going up to 20% edit; or to 110% critdamage.

Edited May 21, 2017 by Azdeus

[Bartimaeus]

It will differ from game system to game system, because sometimes a "10% increase" means adding 10% to your critical chance (so if your base chance was 5%, it would then be 15% - examples of this would be in Diablo, and as Azdeus just said, World of Warcraft, apparently), and sometimes it means to multiply your base chance (so 5% * 1.1 = 5.5%). Sometimes, games aren't totally clear about what they mean at first glance.

[majestic]

On the crit chance vs crit damage, after raiding in WoW where they had both, I'd have to say I really never thought about the increased crit damage as much as I noticed the increase in crit chance. Going from 10% to 20% crit chance was really noticeable, but my strikes going from ~5k to 5.5k wasn't really worth noticing.

I'm weird I guess.

 

That's not weird, that's the way class abilities and gear synergizes with critical strike/heal chance (or other stats). WoW and a lot of other game systems (not necessarily only in MMORPGs) activate special abilities, buffs or debuffs on critical hits. When you have a case of increased critical strike chance being demonstrably superior to pure increased critical effect on the same scale then you're going to notice it much more.

 

Especially if the game in question gives you interesting feedback about it. If you have a cool ability that can only be used every third critical strike you'd probably notice more crits rather than larger ones in a much more satisfying manner.

 

If you don't and all crits really are is bigger numbers on the screen a few of them being much bigger will be much more memorable than a few more being smaller.

[Infinitron]

[kirottu]

I guess you could say Infinitron rebooted this thread.

[injurai]

Just watched this. I really like all the points he had to make. He made a salient point on players experiencing things subconsciously. It's those details that make the world alive. Even in total non-rpgs I think this matters a lot. I think GTA is a great example of accomplishing this at a city wide scale. The Witcher 3 lacked this in Novigrad.

 

Additionally the geometric approach to attributes could be really taken to some complex extremes. Embedding certain traits within the geometry. Perhaps using area sweeping metrics. In his example of one triangle and one inverse triangle, you could make a partition that keeps the area on either side of the division equal. Otherwise this approach would become imbalanced. I can think of some really clever tricks using sparse matrices, overlay graphs, and other things to get an user facing organic feel to stats that is supplemented by a vast and complex set of underlying layers. Using simulated annealing to generate the base balance profiles, etc.

[Infinitron]

Now on YouTube:

 

Edited June 8, 2017 by Infinitron

[Quillon]

And this one was after reboot:

 

[Gizmo]

Late to the party; I just watched Tim Cain's 7 Mistakes talk. I liked it (the whole thing), though I'm not too keen on some of his mentioned "mistakes" as being mistakes,  but I won't doubt (and certainly won't dispute) his change in design preference, or his observations on player reactions to game systems.  (That'd be silly of me )

*But I'd rather the numbers, no guaranteed hits, and actual random random; (indefinite miss streaks and all).

 

 

I wonder if anyone (or how many dozens) mentioned to him that the Witcher sequel asked those many questions because he didn't have the previous two saved games that store the answers to them... The game seeks to match the history of the Player's Geralt, by the way that the previous games concluded; what decisions were made.   Answering at random (instead of with a save), just chooses one of the past options—and its effects.

 

**I recommend playing Witcher 1 though, it's very different from W2 & W3; and can play closer to NwN (not surprisingly)...but with better combat. 

There are a several sad parallels between W1 and what happened with its sequels, as with Fallout and its own most recent sequels. 

(Like Fallout, the first Witcher is my favorite, and most preferred of its series.)

Edited October 9, 2017 by Gizmo

[Jozape]

Late to the party; I just watched Tim Cain's 7 Mistakes talk. I liked it (the whole thing), though I'm not too keen on some of his mentioned "mistakes" as being mistakes,  but I won't doubt (and certainly won't dispute) his change in design preference, and his observations on player reactions to game systems.  (That'd be silly of me )

*But I'd rather the numbers, no guaranteed hits, and actual random random; (indefinite miss streaks and all).

 

Ignorance of probability is one of the worst scourges to have to deal with today. Sad to see such respected designers as Tim Cain change their game design to allow people to remain ignorant. Hopefully he changes his mind...