Posted: Thu Nov 08, 2007 3:48 am
Aluman covered the statistical analysis critique pretty well from the advanced statistics perspective, but for the not-so-statistically inclined, I have a pretty simple method for doing the analysis of cyber armor vs dodge. The only math involved here is addition and division. You can do this by hand, but it's quicker and easier to use a spreadsheet.
How many possible outcomes are there for an attack vs dodge? Exactly 400! 20 equally-possible attack results vs 20 equally-possible dodge results. We're going to account for all 400 possible outcomes. Trust me, it's a piece of cake.
Take an excel spreadsheet. You'll be making a 21 by 21 table. For convenience, set your column width to 2. (Hit Control-A to highlight everything, right click, select Column Width, and enter 2). This will allow you to see the whole table without scrolling side-to-side.
In the far left column, put all 20 values the attacker could get on his roll to strike (he has a +1, so it goes 2, 3, 4... 20, 21). Leave the bottom cell (The 21st one) empty. Then go to the bottom row. Leaving the left-most cell empty (Same cell you just left empty), fill the rest of the row up with the 20 values the defender could get on his rolls to dodge (he has +3 to dodge, so 4, 5, 6... 22, 23).
Now you have an empty field to fill that covers every possible combination of rolls (400 cells). Pick any cell in there. This is one possible instance out of 400. You can look to the left of it and see the attacker's roll. You look to the bottom and you see the defender's roll. If the attack fails, put a 1 in that cell. Repeat that for every cell in the table. Here's a shortcut: Put a 1 in for all tying values. These cells will form a diagonal line going across the table. Everything to one side of the line will be 1's, and everything on the other side will be blank. Remember that a natural 20 beats everything but a natural 20, and any modified roll under a 5 is a miss. That can change things up a little at either end of your triangle of 1's. Count up the 1's and divide by 400. That's the probability that a dodge will work.
I get 245 instances out of 400 where the attacker misses either due to a bad roll (60 instances) or due to a successful dodge (All the rest). In two instances, an attacking natural 20 beats out a modified 21 and 22.
Now the odds that the cyber armor will work with no dodge involved is pretty simple. You take the natural value at which the attacker's roll, with bonuses, equals 16 (15, in this case). Divide that natural roll by 20 and you have your probability of an un-dodged attack either missing or hitting the armor. I get 75%.
The question remains, how do you account for cyber armor plus a dodge? No problem. Go back to that 20 by 20 matrix. Find the column where the monster rolls a 16, with bonuses. Every blank entry in that column, put in an A. Repeat this for all columns to the left. Again, count up the number of A's, divide by 400 (This is quicker if you use a SUM function, but you can do it by hand). THIS is the chance that the cyber armor will absorb damage on a failed dodge when all other armors are gone. I get 78 instances where the armor makes a difference for a failed dodge.
If you change those A's to 1's and count up all the 1's again, you'll have the combined effecitveness of dodging and cyber armor against a given attack. I get 323 instances.
FINAL NUMBERS FOR THESE TWO COMBATANTS:
MONSTER ATTACK VS CYBER-KNIGHT DODGE, NO CYBER-ARMOR: CK dodges 61.25% of the time
MONSTER ATTACK VS CYBER-KNIGHT ARMOR, NO DODGE: Armor takes a hit 75% of the time.
MONSTER ATTACK VS CK DODGE AND CYBER-ARMOR: Either armor takes the hit or Cyber-Knight dodges 80.75% of the time.
CONCLUSIONS:
Dodging helps, but if that extra attack will end the fight sooner, it makes sense in this case to go for simultaneous attacks, since a dodge will improve his odds of surviving an attack by a little under 7%.
The Cyber Armor is a good backup armor against unskilled combatants. Of course, the higher the strike bonuses, the less the armor matters. At +0 to strike, it is effective 80% of the time. At +8 to strike, it is effective 40% of the time. At +16 to strike, the armor is nothing but a decoration.
Higher dodge and parry bonuses also reduce the relative worth of cyber-armor. For this fight, it will protect the defender against a failed dodge 19.5% of the time. The higher the defensive bonus, the less difference the armor makes. An additional +1 to dodge cuts the chance to protect against a failed dodge to 16.5%. An additional +2 to dodge cuts it to 13.75%. An additional +5 to dodge (a +8 total, not unusual for mid-levels) cuts the chance that cyber armor will make protect the character from a failed dodge to 7%. At +15 to dodge or parry, the cyber-armor is only useful for a character who is unconscious, blind, surprise-attacked, or out of attacks.
Bottom line, cyber armor is far from worthless, but its utility drops substantially with increased attacker and defender skill.
Side note: This is a straight-up dice vs dice roll. Bell curves and standard deviations are good for complex systems that tend to push most samples towards similar values. Remember, DocS, that the purpose of math in general and statistics in particular is to bring understanding by way of quantitative modelling. If you throw a lot of advanced terms around in a public forum, you're just sowing confusion, which defeats the purpose of math in the first place
Additional side note: Gloating about your superior math skills doesn't make you look smarter. It makes you look arrogant and condescending. In this case, DocS, you chose the wrong method for analysis. It's time to eat some crow.
How many possible outcomes are there for an attack vs dodge? Exactly 400! 20 equally-possible attack results vs 20 equally-possible dodge results. We're going to account for all 400 possible outcomes. Trust me, it's a piece of cake.
Take an excel spreadsheet. You'll be making a 21 by 21 table. For convenience, set your column width to 2. (Hit Control-A to highlight everything, right click, select Column Width, and enter 2). This will allow you to see the whole table without scrolling side-to-side.
In the far left column, put all 20 values the attacker could get on his roll to strike (he has a +1, so it goes 2, 3, 4... 20, 21). Leave the bottom cell (The 21st one) empty. Then go to the bottom row. Leaving the left-most cell empty (Same cell you just left empty), fill the rest of the row up with the 20 values the defender could get on his rolls to dodge (he has +3 to dodge, so 4, 5, 6... 22, 23).
Now you have an empty field to fill that covers every possible combination of rolls (400 cells). Pick any cell in there. This is one possible instance out of 400. You can look to the left of it and see the attacker's roll. You look to the bottom and you see the defender's roll. If the attack fails, put a 1 in that cell. Repeat that for every cell in the table. Here's a shortcut: Put a 1 in for all tying values. These cells will form a diagonal line going across the table. Everything to one side of the line will be 1's, and everything on the other side will be blank. Remember that a natural 20 beats everything but a natural 20, and any modified roll under a 5 is a miss. That can change things up a little at either end of your triangle of 1's. Count up the 1's and divide by 400. That's the probability that a dodge will work.
I get 245 instances out of 400 where the attacker misses either due to a bad roll (60 instances) or due to a successful dodge (All the rest). In two instances, an attacking natural 20 beats out a modified 21 and 22.
Now the odds that the cyber armor will work with no dodge involved is pretty simple. You take the natural value at which the attacker's roll, with bonuses, equals 16 (15, in this case). Divide that natural roll by 20 and you have your probability of an un-dodged attack either missing or hitting the armor. I get 75%.
The question remains, how do you account for cyber armor plus a dodge? No problem. Go back to that 20 by 20 matrix. Find the column where the monster rolls a 16, with bonuses. Every blank entry in that column, put in an A. Repeat this for all columns to the left. Again, count up the number of A's, divide by 400 (This is quicker if you use a SUM function, but you can do it by hand). THIS is the chance that the cyber armor will absorb damage on a failed dodge when all other armors are gone. I get 78 instances where the armor makes a difference for a failed dodge.
If you change those A's to 1's and count up all the 1's again, you'll have the combined effecitveness of dodging and cyber armor against a given attack. I get 323 instances.
FINAL NUMBERS FOR THESE TWO COMBATANTS:
MONSTER ATTACK VS CYBER-KNIGHT DODGE, NO CYBER-ARMOR: CK dodges 61.25% of the time
MONSTER ATTACK VS CYBER-KNIGHT ARMOR, NO DODGE: Armor takes a hit 75% of the time.
MONSTER ATTACK VS CK DODGE AND CYBER-ARMOR: Either armor takes the hit or Cyber-Knight dodges 80.75% of the time.
CONCLUSIONS:
Dodging helps, but if that extra attack will end the fight sooner, it makes sense in this case to go for simultaneous attacks, since a dodge will improve his odds of surviving an attack by a little under 7%.
The Cyber Armor is a good backup armor against unskilled combatants. Of course, the higher the strike bonuses, the less the armor matters. At +0 to strike, it is effective 80% of the time. At +8 to strike, it is effective 40% of the time. At +16 to strike, the armor is nothing but a decoration.
Higher dodge and parry bonuses also reduce the relative worth of cyber-armor. For this fight, it will protect the defender against a failed dodge 19.5% of the time. The higher the defensive bonus, the less difference the armor makes. An additional +1 to dodge cuts the chance to protect against a failed dodge to 16.5%. An additional +2 to dodge cuts it to 13.75%. An additional +5 to dodge (a +8 total, not unusual for mid-levels) cuts the chance that cyber armor will make protect the character from a failed dodge to 7%. At +15 to dodge or parry, the cyber-armor is only useful for a character who is unconscious, blind, surprise-attacked, or out of attacks.
Bottom line, cyber armor is far from worthless, but its utility drops substantially with increased attacker and defender skill.
Side note: This is a straight-up dice vs dice roll. Bell curves and standard deviations are good for complex systems that tend to push most samples towards similar values. Remember, DocS, that the purpose of math in general and statistics in particular is to bring understanding by way of quantitative modelling. If you throw a lot of advanced terms around in a public forum, you're just sowing confusion, which defeats the purpose of math in the first place
Additional side note: Gloating about your superior math skills doesn't make you look smarter. It makes you look arrogant and condescending. In this case, DocS, you chose the wrong method for analysis. It's time to eat some crow.