borg286
2016-08-11, 09:16 PM
KPR or Kills per round is a measure of combat effectiveness for the goal of killing ASAP and consistently.
KPR is calculated as the expected damage per round / at level monster HP.
Among the many goals that characters try to accomplish (damage, defending, battlefield control, debuffing, social prowess, exploration, investigation...) damage is the easiest to rank objectively.
The goal of the KPR/DPR King candidates thread that I will make after enough consensus has been reached will be to provide lurkers and optimizers a place to submit builds focused on damage both for identifying broken combos, understanding the arms race between monsters and players, and providing builds to serve as a basis of other optimization builds and players.
When I made DPR King Candidates for 4e I had to go through a few iterations to figure out the best way to let builds play nicely. I'd like this thread to focus on discussing the following topics:
What metric[s] to use (KPR, DPR, 1 round nova, 2nd round of mini boss, Single target, Multi-target, Percent of KPR contributed by single target damage
Should we have candidates fight a block of tofu with appropriate armor and saves, or lay out encounters with KPR evaluated over both battles
How to handle daily resources like spells/smites
How to handle the balance between lots of short rests vs few
Qualifications for what should be considered cheese, or would be a flag to a browser that this build requires interpretations or gameplay behavior that would be difficult to get.
What levels candidates compete at (Level 5 favors fighter and discourages multiclass, Level 6 favors a 1 level dip...)
How to treat concentration?
I'd like to stay away from the following topics
Polearm Master + X
Combos requiring level 9 spells
Level 20 abilities
Warlock 2 / Sorcerer X is OP
Here are some reference numbers:
Here (https://docs.google.com/spreadsheets/d/1d-9xDdath8kX_v7Rpts9JFIJwIG3X0-dDUtfax14NT0/edit#gid=2025852255) is a spreadsheet of the DPR of various iconic builds for each class focused on damage.
Here (https://docs.google.com/spreadsheets/d/1d-9xDdath8kX_v7Rpts9JFIJwIG3X0-dDUtfax14NT0/edit#gid=1826558867) is the KPR version.
Here (https://docs.google.com/spreadsheets/d/1d-9xDdath8kX_v7Rpts9JFIJwIG3X0-dDUtfax14NT0/edit#gid=1829796671) are some averages of monster statistics.
I'll be using the following formulas to model monsters based on these (https://docs.google.com/spreadsheets/d/1Bp8Alhfrl3mgHbenZUbgcOu4zkDV_pXMEs0DUq-S5lw/edit#gid=1829796671) numbers
Attack Bonus(CR)=.37*level+3.8
HP(CR) = 13*level + 7.4
AC(CR) = .35*level + 13
Con(CR) = Dex(CR) = level/2
Daily limitations (source (https://docs.google.com/spreadsheets/d/1d-9xDdath8kX_v7Rpts9JFIJwIG3X0-dDUtfax14NT0/edit#gid=2091322934))
Rounds per encounter 5.8 discussion (http://www.giantitp.com/forums/showthread.php?446568-Thought-Process-around-Optimizing-a-Greatsword-Wielder-WIP-PEACH&p=19897323#post19897323)
Encounters per day 4.30 discussion (http://www.giantitp.com/forums/showthread.php?446568-Thought-Process-around-Optimizing-a-Greatsword-Wielder-WIP-PEACH&p=19897425#post19897425)
Number of Rounds per Day 24.9
I think the actual number of rounds per day for most groups would be less than this. But this is what the DMG recommends and what I aim for.
Here are my thoughts on the above topics.
Regarding the different metrics to use, I like fewer metrics to lower the cost of submitting builds. Sadly we have different situations that are put on us that we may want to favor. Pure at-will KPR is easier to calculate but doesn't represent how the build will do when put up against typical day. Also the DMG recommends around 25 rounds per day, thus daily resources can easily be rationed and serve as a good estimate on how much can be expected.
With 4.3 rounds expected per day, being able to frontload damage each encounter and not just have 1-2 spell slots that you rely on for your one trick pony, how do we take into account encounter 3 and 4?
Regarding how to measure nova, the total damage of round 1 and 2 bias towards outliers at the sacrifice of repeatability. Should we toss out builds that fall back to spells 2 levels lower than their max, where their KPR is effectively halved for the rest of the day?
We could take the geometric mean of the 2 KPR values which would favor more balanced builds. We could take the nova damage of round 2 of encounter 2 (assuming rounds 1 and 2 of the current and prior encounters did more damage). A simple way to express this would be "Mini boss round 2, Fight!!!"
Should this nova damage be calculated separately than the sustained KPR? Should the sustained KPR be pulled from either a pure at-will calculation or from the damage in some round in the day like, "third most damaging encounter, third most damaging round". When calculating KPR for my sorcerer build I found I had to do some pretty heavy calculations to get semi-accurate KPR calculations for levels 1-20. Even at a particular level I had to play out each round what I'd do, and the expected damage, how many foes in this area attack...
Qualifications for what is cheese.
Just to be clear, if a particular mechanic is considered cheesy it is often not always outlawed. It is just something that many optimizers end up using to squeek out more damage. For example back in 4e charging was considered a bit cheesy because to get that damage you had to be able to charge every round, which would incur an AoE if your target didn't die. This high risk high reward wasn't incorporated into the KPR calculation, thus we had a cheese tag we attached to all chargers unless you had some way of mitigating the risk. I consider a warlock using Devil's eye and darkness cheesy if he plans on using it in melee because it interferes too much with non-warlock builds. Thus I would be naming this combo cheese and have a tag for it. Any build that uses it would get simply tagged accordingly and ranked among everyone else. If you don't want to use that cheese you can skip those builds. I do not consider Polearm Master + Great weapon master cheesy because it does not interfere with other party members, and, as we can see with the KPR calculations above, it doesn't give a KPR above .8, which I find quite absurd.
If a particular mechanic results in pun-pun or the like it will be outright banned from competition. If it gives you near-unlimited HP it won't be banned, just tagged somehow. It doesn't make encounters end in 1 round, but it enables your character to solo 80% of battles.
Regarding concentration
It is a reasonable expectation that a caster would have someone to protect him, but I feel a build should note that it needs to stay in the rear and someone else needs to take the punches. Thus I feel concentration, while not strictly cheesy, should be a tag.
If a build will concentrate on spells and be in the front we need some way incorporate the odds of him loosing concentration on his spell. The challenge with concentration is that they are comparing apples (save DC and proficiency bonuses) to oranges (damage and HP scaling). They tried to convert it by using the 1/2 factor but it doesn't work at high levels. Once you reach level 20 monsters the damage scalar goes from 6*level to 18*level. The average damage of a level 25 monster, according to the DMG, is 240 damage, which is a 140 DC concentration check. Good luck with that.. But what I found is that the most damage monsters did per attack was only in the 40's, they just get multiple attacks. Even then, the damage dice, if the DM rolled, there is a good chance that the observed damage makes the DC way to high. These high DC numbers are seen as low as CR 8 monsters. We could possibly calculate the probability that at level X damage Y is dealt. Thus each build that needs concentration would need to spit out the damage needed to give the build a 30% chance of losing concentration. This gets even more complicated if we wanted to allow for the shield spell.
Regarding what levels to stat things out at here are my proposals and reasons
3: Gives a good taste of low level, allows some multiclassing, and permits some iconic features to get turned on (Battle Master, Metamagic, 2nd level spells, Rogue specialization),
7: Gives some room for multiclassing while still getting the coveted 2nd attack
13,14,15: I really can't pick one. I need one in the middle but don't know which.
19: Stays away from level 20 crazyness
KPR is calculated as the expected damage per round / at level monster HP.
Among the many goals that characters try to accomplish (damage, defending, battlefield control, debuffing, social prowess, exploration, investigation...) damage is the easiest to rank objectively.
The goal of the KPR/DPR King candidates thread that I will make after enough consensus has been reached will be to provide lurkers and optimizers a place to submit builds focused on damage both for identifying broken combos, understanding the arms race between monsters and players, and providing builds to serve as a basis of other optimization builds and players.
When I made DPR King Candidates for 4e I had to go through a few iterations to figure out the best way to let builds play nicely. I'd like this thread to focus on discussing the following topics:
What metric[s] to use (KPR, DPR, 1 round nova, 2nd round of mini boss, Single target, Multi-target, Percent of KPR contributed by single target damage
Should we have candidates fight a block of tofu with appropriate armor and saves, or lay out encounters with KPR evaluated over both battles
How to handle daily resources like spells/smites
How to handle the balance between lots of short rests vs few
Qualifications for what should be considered cheese, or would be a flag to a browser that this build requires interpretations or gameplay behavior that would be difficult to get.
What levels candidates compete at (Level 5 favors fighter and discourages multiclass, Level 6 favors a 1 level dip...)
How to treat concentration?
I'd like to stay away from the following topics
Polearm Master + X
Combos requiring level 9 spells
Level 20 abilities
Warlock 2 / Sorcerer X is OP
Here are some reference numbers:
Here (https://docs.google.com/spreadsheets/d/1d-9xDdath8kX_v7Rpts9JFIJwIG3X0-dDUtfax14NT0/edit#gid=2025852255) is a spreadsheet of the DPR of various iconic builds for each class focused on damage.
Here (https://docs.google.com/spreadsheets/d/1d-9xDdath8kX_v7Rpts9JFIJwIG3X0-dDUtfax14NT0/edit#gid=1826558867) is the KPR version.
Here (https://docs.google.com/spreadsheets/d/1d-9xDdath8kX_v7Rpts9JFIJwIG3X0-dDUtfax14NT0/edit#gid=1829796671) are some averages of monster statistics.
I'll be using the following formulas to model monsters based on these (https://docs.google.com/spreadsheets/d/1Bp8Alhfrl3mgHbenZUbgcOu4zkDV_pXMEs0DUq-S5lw/edit#gid=1829796671) numbers
Attack Bonus(CR)=.37*level+3.8
HP(CR) = 13*level + 7.4
AC(CR) = .35*level + 13
Con(CR) = Dex(CR) = level/2
Daily limitations (source (https://docs.google.com/spreadsheets/d/1d-9xDdath8kX_v7Rpts9JFIJwIG3X0-dDUtfax14NT0/edit#gid=2091322934))
Rounds per encounter 5.8 discussion (http://www.giantitp.com/forums/showthread.php?446568-Thought-Process-around-Optimizing-a-Greatsword-Wielder-WIP-PEACH&p=19897323#post19897323)
Encounters per day 4.30 discussion (http://www.giantitp.com/forums/showthread.php?446568-Thought-Process-around-Optimizing-a-Greatsword-Wielder-WIP-PEACH&p=19897425#post19897425)
Number of Rounds per Day 24.9
I think the actual number of rounds per day for most groups would be less than this. But this is what the DMG recommends and what I aim for.
Here are my thoughts on the above topics.
Regarding the different metrics to use, I like fewer metrics to lower the cost of submitting builds. Sadly we have different situations that are put on us that we may want to favor. Pure at-will KPR is easier to calculate but doesn't represent how the build will do when put up against typical day. Also the DMG recommends around 25 rounds per day, thus daily resources can easily be rationed and serve as a good estimate on how much can be expected.
With 4.3 rounds expected per day, being able to frontload damage each encounter and not just have 1-2 spell slots that you rely on for your one trick pony, how do we take into account encounter 3 and 4?
Regarding how to measure nova, the total damage of round 1 and 2 bias towards outliers at the sacrifice of repeatability. Should we toss out builds that fall back to spells 2 levels lower than their max, where their KPR is effectively halved for the rest of the day?
We could take the geometric mean of the 2 KPR values which would favor more balanced builds. We could take the nova damage of round 2 of encounter 2 (assuming rounds 1 and 2 of the current and prior encounters did more damage). A simple way to express this would be "Mini boss round 2, Fight!!!"
Should this nova damage be calculated separately than the sustained KPR? Should the sustained KPR be pulled from either a pure at-will calculation or from the damage in some round in the day like, "third most damaging encounter, third most damaging round". When calculating KPR for my sorcerer build I found I had to do some pretty heavy calculations to get semi-accurate KPR calculations for levels 1-20. Even at a particular level I had to play out each round what I'd do, and the expected damage, how many foes in this area attack...
Qualifications for what is cheese.
Just to be clear, if a particular mechanic is considered cheesy it is often not always outlawed. It is just something that many optimizers end up using to squeek out more damage. For example back in 4e charging was considered a bit cheesy because to get that damage you had to be able to charge every round, which would incur an AoE if your target didn't die. This high risk high reward wasn't incorporated into the KPR calculation, thus we had a cheese tag we attached to all chargers unless you had some way of mitigating the risk. I consider a warlock using Devil's eye and darkness cheesy if he plans on using it in melee because it interferes too much with non-warlock builds. Thus I would be naming this combo cheese and have a tag for it. Any build that uses it would get simply tagged accordingly and ranked among everyone else. If you don't want to use that cheese you can skip those builds. I do not consider Polearm Master + Great weapon master cheesy because it does not interfere with other party members, and, as we can see with the KPR calculations above, it doesn't give a KPR above .8, which I find quite absurd.
If a particular mechanic results in pun-pun or the like it will be outright banned from competition. If it gives you near-unlimited HP it won't be banned, just tagged somehow. It doesn't make encounters end in 1 round, but it enables your character to solo 80% of battles.
Regarding concentration
It is a reasonable expectation that a caster would have someone to protect him, but I feel a build should note that it needs to stay in the rear and someone else needs to take the punches. Thus I feel concentration, while not strictly cheesy, should be a tag.
If a build will concentrate on spells and be in the front we need some way incorporate the odds of him loosing concentration on his spell. The challenge with concentration is that they are comparing apples (save DC and proficiency bonuses) to oranges (damage and HP scaling). They tried to convert it by using the 1/2 factor but it doesn't work at high levels. Once you reach level 20 monsters the damage scalar goes from 6*level to 18*level. The average damage of a level 25 monster, according to the DMG, is 240 damage, which is a 140 DC concentration check. Good luck with that.. But what I found is that the most damage monsters did per attack was only in the 40's, they just get multiple attacks. Even then, the damage dice, if the DM rolled, there is a good chance that the observed damage makes the DC way to high. These high DC numbers are seen as low as CR 8 monsters. We could possibly calculate the probability that at level X damage Y is dealt. Thus each build that needs concentration would need to spit out the damage needed to give the build a 30% chance of losing concentration. This gets even more complicated if we wanted to allow for the shield spell.
Regarding what levels to stat things out at here are my proposals and reasons
3: Gives a good taste of low level, allows some multiclassing, and permits some iconic features to get turned on (Battle Master, Metamagic, 2nd level spells, Rogue specialization),
7: Gives some room for multiclassing while still getting the coveted 2nd attack
13,14,15: I really can't pick one. I need one in the middle but don't know which.
19: Stays away from level 20 crazyness