**kmktr**

Probability, in games, is a heck of a lot of fun.

I used to play face-to-face, dice-randomized RPGs quite a lot, starting -- wow, about 23 years ago. (Now I feel really, really old! I'm sure I'm not the only one 'thac0' has meaning for, tho'! ^_~)

After a while, an instinct for probability developed naturally. Base 10 is easy for many people to grasp, so calculating the chance of a successful roll on a d10 (10-sided die) or d100 (two 10-sided dice, one for the tens and one for the ones) in any given situation becomes second nature after a while. All of us use it to some extent, as percentages in our mundane lives.

Probability for something like d4, d6 (which is fairly well-known to anyone who gambles with standard dice), d8, d12, d20 and so forth is different. Quick, which is better, a weapon that does 1d10 or one that does 2d4? All other attributes being equal, attack speed, accuracy, thac0 and so forth, I'd take the 2d4 weapon.

Here's why:

The single 10-sided die of damage does give you the chance to roll a 10 as the result 10% of the time, but it also gives you the chance for a 1 10% of the time as well. The 2 4-sided dice of damage gives you 8, the maximum damage, 12.5% of the time, but the least damage you can do is 2, which you also can get 12.5% of the time. But, it's not really quite so straightforward. In order to get a 2 as a result on the 2d4, you must first roll a 1 on the first die. The chance of this is one in four, or 25%. You must then roll a 1 on the second die roll, which is also a one in four chance. Ultimately, the mean average chance of rolling anything from a 1 to a 10 on the d10 is 10%, and the mean average of rolling anything from a 2 to an 8 on the 2d4 is 12.5% -- but, in a way, you have to be 'unlucky' twice

That's a very simple probability model, as the only variable is the damage per die, and the number of dice rolled per attempt. Consider: the more interesting games (at least for face-to-face, die-randomized RPGs back in the day) had more complex probability matrices. What if the refining matrix for ROSE, with its guaranteed ability to refine to a (3), uses a two (or more) variable formula for success? It seems as if it really must, since the majority of people are sensing a difference in refines and drill successes, now. The GMs have stated that there has been no change to the success (since the 'guaranteed to three' refining change) and there's no reason to believe they'd not be truthful to us. Therefore, the only change is the number of attempts to refine items/add slots both by individuals, and across the population of the game.

Some people used to 'rest' an item after gaining a refine level. Perhaps, after a success, the next refine success chance drops by 1%. What if this variable is 1% per current refine level? Your chance to fail has gone from 1%, when you start, to 5% trying from (4) to (5). Now, what if this is a cumulative effect? That is base 1% failure rate, add 2% (going from 1 - 2), plus 3% (2 - 3) -- under the new refine system, these chances to fail might be 'masked' until you go to (3 - 4) picking up another 4% handicap, then 5% (4 - 5) and so on. Going from a 4 to a 5 now has a whopping 15% chance to fail! 5 - 6 can fail 21% of the time, 28% would be the failure going 6 - 7, 7 - 8 is 36% and 8 - 9 is 45% likely to fail. That's a significant jump in failure rate per level. Now, what if you 'rest' the item for 10 minutes/percent (or current item refine level)? If you wait 50 minutes (in the simple 1% per level model), the chance to fail has reverted back to 1% -- assuming that 'resting' an item actually has any sort of an effect. (ALL of this is just supposition -- a what if it's this way? -- on my part. I have no real idea of the variables at play in refining success/failure rates.)

Ah, but what if there's a multiple variable system in play? What if the number of minutes in battle between refine attempts, or the level of the character attempting to refine, or the number rendered by your Charm divided by your highest stat somehow comes into play, too? The chance to succeed/fail has become a completely unknown variable, and only very tedious, zulie-expensive, well-documented, rigidly controlled attempts (scientific method, anyone?) at refining will reveal a winning combination of events to reduce the failure rate of high refine attempts to 'acceptable' levels -- assuming that there even

I used to play-test a fair number of face-to-face RPG systems, back in the day (I used to work in a gaming store). One of the things the designers were always looking for was 'game balance'. Was the effort-versus-reward ratio too high (was it too frustrating to 'earn' a reward in the game -- or was the reward too puny for the effort and game resources expended to earn it)? Was the ratio too low and thus unsastifying? Each of us who play-tested, at one point in our favorite games, had had a 'godly' character. Funny thing -- godly characters are a heck of a lot of fun -- for about twenty minutes. With no challenge and no ability to grow in any sort of measurable way "you overkilled Gormgoth the Mighty, in one shot, by a thousand points, and gained 100 levels! Hernstance the Invincible is reduced to atomic dust with your backswing (of the same attack, mind you), because you overkilled him by a million points!" they quickly become boring.

It's long been my belief that we have certain extremely juvenile individuals (no one on this forum, I'm certain!) to thank for the fact the beta closed when it did, before the ROSE team was able to do more than get a general sense of where this effort-to-reward balance point makes the most number of players happy. People get discouraged if it's too hard. Others get bored when things are too easy. I think this is one of the game refinements that Leonis and his team are paying attention to. But, the research for something like this takes a while. Changes that are too sudden or too drastic tend to make people unhappy. I wouldn't be at all surprised if there is one person on the GM team who is paying particular attention whenever refines (and things like drill successes) are brought up on the forums or in 1 on 1 questions and comments, and is keeping track. If too many people, who otherwise don't complain about the random aspects of the game that are deemed to be 'balanced' by the GMs, are found to be complaining, I am certain the GMs pay attention.

My 'gut' feeling about refine and drill successes and failures changing with the new drop system? I wonder if it might not be a hidden 'two (or more)-factor' matrix. People have discussed, sometimes with a bit of heat, if the success and failure of people refining around you matters to your refine chances. Technically, perhaps it shouldn't. But, I recall mobs of people around both the refine NPC in Breezy Hills (Mairaith?), and Crune in Junon Polis, with their item and materials all loaded up, fingers itching to hit the 'go' bar, waiting for someone else to get that gaudy 'failure' thinking that trying a tricky refine right after someone else's failure gave one a better chance for success. Well, with so many more refining materials, and drills dropping and being used in game, perhaps that is a more perceptible part of the variable now. If 100 people throughout the channel are trying to drill at the same time, perhaps only a handful really have a shot at a success.

It could well be that there was no change at all to the refine or drill system accompanying the change to the drop system. It could be this change is the result of that many more refine and drill attempts being made -- sort of pointing to the fact there is more than just a simple, straight percent chance to succeed/fail to refine or add a slot to an item. It could also be that more people, with more readily available refines and drills, are more willing to risk refining higher, or drilling more equipment. More attempts per person = more failures per person, too. (And, of course, luck *does* play a role in everything, no matter how supposedly perfectly 'random' things are supposed to be. I, for one, have rotten luck, so I'm not looking forward to trying to drill equipment when I finally get there!)

I used to play face-to-face, dice-randomized RPGs quite a lot, starting -- wow, about 23 years ago. (Now I feel really, really old! I'm sure I'm not the only one 'thac0' has meaning for, tho'! ^_~)

After a while, an instinct for probability developed naturally. Base 10 is easy for many people to grasp, so calculating the chance of a successful roll on a d10 (10-sided die) or d100 (two 10-sided dice, one for the tens and one for the ones) in any given situation becomes second nature after a while. All of us use it to some extent, as percentages in our mundane lives.

Probability for something like d4, d6 (which is fairly well-known to anyone who gambles with standard dice), d8, d12, d20 and so forth is different. Quick, which is better, a weapon that does 1d10 or one that does 2d4? All other attributes being equal, attack speed, accuracy, thac0 and so forth, I'd take the 2d4 weapon.

Here's why:

The single 10-sided die of damage does give you the chance to roll a 10 as the result 10% of the time, but it also gives you the chance for a 1 10% of the time as well. The 2 4-sided dice of damage gives you 8, the maximum damage, 12.5% of the time, but the least damage you can do is 2, which you also can get 12.5% of the time. But, it's not really quite so straightforward. In order to get a 2 as a result on the 2d4, you must first roll a 1 on the first die. The chance of this is one in four, or 25%. You must then roll a 1 on the second die roll, which is also a one in four chance. Ultimately, the mean average chance of rolling anything from a 1 to a 10 on the d10 is 10%, and the mean average of rolling anything from a 2 to an 8 on the 2d4 is 12.5% -- but, in a way, you have to be 'unlucky' twice

*in a row*in the 2d4 die roll to get the worst result. It's a very subtle difference, but over time, the 2d4 seems to render more stable, and higher average damage than the d10.That's a very simple probability model, as the only variable is the damage per die, and the number of dice rolled per attempt. Consider: the more interesting games (at least for face-to-face, die-randomized RPGs back in the day) had more complex probability matrices. What if the refining matrix for ROSE, with its guaranteed ability to refine to a (3), uses a two (or more) variable formula for success? It seems as if it really must, since the majority of people are sensing a difference in refines and drill successes, now. The GMs have stated that there has been no change to the success (since the 'guaranteed to three' refining change) and there's no reason to believe they'd not be truthful to us. Therefore, the only change is the number of attempts to refine items/add slots both by individuals, and across the population of the game.

Some people used to 'rest' an item after gaining a refine level. Perhaps, after a success, the next refine success chance drops by 1%. What if this variable is 1% per current refine level? Your chance to fail has gone from 1%, when you start, to 5% trying from (4) to (5). Now, what if this is a cumulative effect? That is base 1% failure rate, add 2% (going from 1 - 2), plus 3% (2 - 3) -- under the new refine system, these chances to fail might be 'masked' until you go to (3 - 4) picking up another 4% handicap, then 5% (4 - 5) and so on. Going from a 4 to a 5 now has a whopping 15% chance to fail! 5 - 6 can fail 21% of the time, 28% would be the failure going 6 - 7, 7 - 8 is 36% and 8 - 9 is 45% likely to fail. That's a significant jump in failure rate per level. Now, what if you 'rest' the item for 10 minutes/percent (or current item refine level)? If you wait 50 minutes (in the simple 1% per level model), the chance to fail has reverted back to 1% -- assuming that 'resting' an item actually has any sort of an effect. (ALL of this is just supposition -- a what if it's this way? -- on my part. I have no real idea of the variables at play in refining success/failure rates.)

Ah, but what if there's a multiple variable system in play? What if the number of minutes in battle between refine attempts, or the level of the character attempting to refine, or the number rendered by your Charm divided by your highest stat somehow comes into play, too? The chance to succeed/fail has become a completely unknown variable, and only very tedious, zulie-expensive, well-documented, rigidly controlled attempts (scientific method, anyone?) at refining will reveal a winning combination of events to reduce the failure rate of high refine attempts to 'acceptable' levels -- assuming that there even

*is*any sort of action players can take to change one of the variables.I used to play-test a fair number of face-to-face RPG systems, back in the day (I used to work in a gaming store). One of the things the designers were always looking for was 'game balance'. Was the effort-versus-reward ratio too high (was it too frustrating to 'earn' a reward in the game -- or was the reward too puny for the effort and game resources expended to earn it)? Was the ratio too low and thus unsastifying? Each of us who play-tested, at one point in our favorite games, had had a 'godly' character. Funny thing -- godly characters are a heck of a lot of fun -- for about twenty minutes. With no challenge and no ability to grow in any sort of measurable way "you overkilled Gormgoth the Mighty, in one shot, by a thousand points, and gained 100 levels! Hernstance the Invincible is reduced to atomic dust with your backswing (of the same attack, mind you), because you overkilled him by a million points!" they quickly become boring.

It's long been my belief that we have certain extremely juvenile individuals (no one on this forum, I'm certain!) to thank for the fact the beta closed when it did, before the ROSE team was able to do more than get a general sense of where this effort-to-reward balance point makes the most number of players happy. People get discouraged if it's too hard. Others get bored when things are too easy. I think this is one of the game refinements that Leonis and his team are paying attention to. But, the research for something like this takes a while. Changes that are too sudden or too drastic tend to make people unhappy. I wouldn't be at all surprised if there is one person on the GM team who is paying particular attention whenever refines (and things like drill successes) are brought up on the forums or in 1 on 1 questions and comments, and is keeping track. If too many people, who otherwise don't complain about the random aspects of the game that are deemed to be 'balanced' by the GMs, are found to be complaining, I am certain the GMs pay attention.

My 'gut' feeling about refine and drill successes and failures changing with the new drop system? I wonder if it might not be a hidden 'two (or more)-factor' matrix. People have discussed, sometimes with a bit of heat, if the success and failure of people refining around you matters to your refine chances. Technically, perhaps it shouldn't. But, I recall mobs of people around both the refine NPC in Breezy Hills (Mairaith?), and Crune in Junon Polis, with their item and materials all loaded up, fingers itching to hit the 'go' bar, waiting for someone else to get that gaudy 'failure' thinking that trying a tricky refine right after someone else's failure gave one a better chance for success. Well, with so many more refining materials, and drills dropping and being used in game, perhaps that is a more perceptible part of the variable now. If 100 people throughout the channel are trying to drill at the same time, perhaps only a handful really have a shot at a success.

It could well be that there was no change at all to the refine or drill system accompanying the change to the drop system. It could be this change is the result of that many more refine and drill attempts being made -- sort of pointing to the fact there is more than just a simple, straight percent chance to succeed/fail to refine or add a slot to an item. It could also be that more people, with more readily available refines and drills, are more willing to risk refining higher, or drilling more equipment. More attempts per person = more failures per person, too. (And, of course, luck *does* play a role in everything, no matter how supposedly perfectly 'random' things are supposed to be. I, for one, have rotten luck, so I'm not looking forward to trying to drill equipment when I finally get there!)

## no subject

Date: 2007-06-30 06:33 pm (UTC)tex-chan.livejournal.com