Challenge Ratings

[Lord Rahvin]

> Yep. I know very well what I wrote. The 5% refers to TPK, a situation the party should very well escape from. There`s also a 15% of all encounters that should be considered "deadly;" meaning a good chance of at least one PCC. Now, 5% *will* lose - that`s a bit harsh, isn`t it? In this context, *will* lose is the equivalent of the DM saying, "you die, no save, and you can`t run away, either." Gee, that`d be a fun game to play in - especially if the DM did it more than every 20 encounters, too. How often is the DM supposed to off-handedly kill the PCs? Every 10 encounters? 15.673% of all encounters? 27.894%? (I`m sure all those extra decimals are of supreme importance, since we`re into calculating the statistics of it all, I mean.)


---
I haven`t really been following this, but you guys are (sort of) on the
subject and I thought I`d chime in here with a bit of non-statistical
perspective. Though this is an analysis of the EL system for d20Modern, I
think these ideas are of significance to anyone discussing the way ELs or
CRs work in any d20 game.

I`ve recently picked up D20Modern, and the R&D people seem to have changed
their mind (slightly) on how the CR/EL function works. Mainly, the whole
5%/10%/15%/50% distribution of ELs has all been removed. Their only advice
now on the subject is to keep it varied; have some lower and have some
higher and they`re generally trying to get rid of the idea that parties are
going to frequently encounter ELs equal to their level, and that therefore,
they are always assumed to win combats.

Their only advice now is to keep it varied and keep it fun. They recommend
keeping all ELs roughly within 5 points lower or 5 points higher than the
party level. They do not give experience awards for any encounter higher
than "party level +8" or lower than "party level - 8", not because the DM
shouldn`t throw them in, but because they don`t believe any value they place
on the experience table would accurately reflect this type of encounter.

Their method of calculating the ELs are a bit different, too. Instead of
adding a +2 EL every time you double the number of creatures, you take the
average CR of all opposition (including traps, environmental hazards,
necessary key skill checks, and opponents) that will be encountered
simultaneously, and this average becomes the base EL. Apply +1 EL per
threat beyond one, to a maximum of +6. Raise or lower the EL by 1/3 for
extreme tactical advantage. Finally, if the situation is Low Threat
(non-combat encounter that could potentially erupt into danger) then drop
the EL by half. "No Threat" encounters are dropped by 1/4.

So, 3 Goblins(CR 1) and a Gargoyle(CR 4) in a standard combat encounter
would count as EL 5.5, round down to 5. 11 Goblins(CR 1) would come out to
EL 7. A Gorgoyle(CR 4) and 11 Goblins (CR 1) would come out to EL 7.5,
round down to EL 7. Entering into "hostile negotiations" with this Gargoyle
(CR 4) and his band of eleven Goblin (CR 1) thugs using a (DC 26) Diplomacy
check (CR 3) would drop the encounter level to (9/2=4.5) EL 4.

Although it wasn`t really stated anywhere, there also seems to be a general
emphasis on ELs being calculated and used after playing sessions for
purposes of determining experience, rather than being used as a tool to
construct adventure modules.

Now there are a lot of problems with even this view of CRs and ELs, not the
least of which is still the idea that NPC classes are still CR equal to
level -1. There are many distinct differences between D&D and d20Modern so
I`m not suggesting that what applies to one automatically applies to the
other. D20Modern has less spells and lower base attack bonuses, goblins and
such are more of a threat due to massive damage, and most higher CR monsters
will have conventional weaknesses (unfortunately not taken into account for
purposes of CR and EL) and in a Modern setting not only is there less an
emphasis on toe-to-toe melee combat exchanges, but the use of technology,
communication, firearms, and charisma-based attack options changes even the
nature of those melee exchanges. And, of course, Action Points(D20Modern)
and Magical Items(D&D) change everything.

But I think its worth noting that you could play around with those
5%/10%/15%/50% distributions, even going so far as to disregard them
altogether, and the system does not break down. The CR and EL systems
aren`t that great to begin with and playing around with them doesn`t destroy
them. You could toss in plenty of "Overwhelming" encounters if it fits your
playing style, and it doesn`t destroy the system. PCs don`t have to level
up every 12 encounters. You don`t even really *have to* even know the party
level when designing your adventure. Players will find a way to beat it
because their players and that`s what players do. They might need your help
once in a while. It would help to let them generally know what to expect in
terms of relative difficulty. But you could do it. PCs don`t automatically
die. They still get plenty of experience. DMs don`t necessarily get dice
thrown at them. Encounter Level distribution ratios aren`t the only way to
run the game. Most games don`t even have `em. CRs are meant to provide
adequite rewards (within the limits of the experience system), and provide a
measure of challenge (it`s questionable whether they do this well). It
wasn`t meant to dictate how the game should be played, what kind of
encounters you can and can`t use in your games, or how often PCs should be
put into life-threatening situations.

-Lord Rahvin

************************************************** **************************
The Birthright Homepage: http://www.birthright.net
Birthright-l Archives: http://oracle.wizards.com/archives/birthright-l.html
To unsubscribe, send email to LISTSERV@ORACLE.WIZARDS.COM
with UNSUB BIRTHRIGHT-L in the body of the message.

[Lord Rahvin]

> Yep. I know very well what I wrote. The 5% refers to TPK, a situation the party should very well escape from. There`s also a 15% of all encounters that should be considered "deadly;" meaning a good chance of at least one PCC. Now, 5% *will* lose - that`s a bit harsh, isn`t it? In this context, *will* lose is the equivalent of the DM saying, "you die, no save, and you can`t run away, either." Gee, that`d be a fun game to play in - especially if the DM did it more than every 20 encounters, too. How often is the DM supposed to off-handedly kill the PCs? Every 10 encounters? 15.673% of all encounters? 27.894%? (I`m sure all those extra decimals are of supreme importance, since we`re into calculating the statistics of it all, I mean.)


---
I haven`t really been following this, but you guys are (sort of) on the
subject and I thought I`d chime in here with a bit of non-statistical
perspective. Though this is an analysis of the EL system for d20Modern, I
think these ideas are of significance to anyone discussing the way ELs or
CRs work in any d20 game.

I`ve recently picked up D20Modern, and the R&D people seem to have changed
their mind (slightly) on how the CR/EL function works. Mainly, the whole
5%/10%/15%/50% distribution of ELs has all been removed. Their only advice
now on the subject is to keep it varied; have some lower and have some
higher and they`re generally trying to get rid of the idea that parties are
going to frequently encounter ELs equal to their level, and that therefore,
they are always assumed to win combats.

Their only advice now is to keep it varied and keep it fun. They recommend
keeping all ELs roughly within 5 points lower or 5 points higher than the
party level. They do not give experience awards for any encounter higher
than "party level +8" or lower than "party level - 8", not because the DM
shouldn`t throw them in, but because they don`t believe any value they place
on the experience table would accurately reflect this type of encounter.

Their method of calculating the ELs are a bit different, too. Instead of
adding a +2 EL every time you double the number of creatures, you take the
average CR of all opposition (including traps, environmental hazards,
necessary key skill checks, and opponents) that will be encountered
simultaneously, and this average becomes the base EL. Apply +1 EL per
threat beyond one, to a maximum of +6. Raise or lower the EL by 1/3 for
extreme tactical advantage. Finally, if the situation is Low Threat
(non-combat encounter that could potentially erupt into danger) then drop
the EL by half. "No Threat" encounters are dropped by 1/4.

So, 3 Goblins(CR 1) and a Gargoyle(CR 4) in a standard combat encounter
would count as EL 5.5, round down to 5. 11 Goblins(CR 1) would come out to
EL 7. A Gorgoyle(CR 4) and 11 Goblins (CR 1) would come out to EL 7.5,
round down to EL 7. Entering into "hostile negotiations" with this Gargoyle
(CR 4) and his band of eleven Goblin (CR 1) thugs using a (DC 26) Diplomacy
check (CR 3) would drop the encounter level to (9/2=4.5) EL 4.

Although it wasn`t really stated anywhere, there also seems to be a general
emphasis on ELs being calculated and used after playing sessions for
purposes of determining experience, rather than being used as a tool to
construct adventure modules.

Now there are a lot of problems with even this view of CRs and ELs, not the
least of which is still the idea that NPC classes are still CR equal to
level -1. There are many distinct differences between D&D and d20Modern so
I`m not suggesting that what applies to one automatically applies to the
other. D20Modern has less spells and lower base attack bonuses, goblins and
such are more of a threat due to massive damage, and most higher CR monsters
will have conventional weaknesses (unfortunately not taken into account for
purposes of CR and EL) and in a Modern setting not only is there less an
emphasis on toe-to-toe melee combat exchanges, but the use of technology,
communication, firearms, and charisma-based attack options changes even the
nature of those melee exchanges. And, of course, Action Points(D20Modern)
and Magical Items(D&D) change everything.

But I think its worth noting that you could play around with those
5%/10%/15%/50% distributions, even going so far as to disregard them
altogether, and the system does not break down. The CR and EL systems
aren`t that great to begin with and playing around with them doesn`t destroy
them. You could toss in plenty of "Overwhelming" encounters if it fits your
playing style, and it doesn`t destroy the system. PCs don`t have to level
up every 12 encounters. You don`t even really *have to* even know the party
level when designing your adventure. Players will find a way to beat it
because their players and that`s what players do. They might need your help
once in a while. It would help to let them generally know what to expect in
terms of relative difficulty. But you could do it. PCs don`t automatically
die. They still get plenty of experience. DMs don`t necessarily get dice
thrown at them. Encounter Level distribution ratios aren`t the only way to
run the game. Most games don`t even have `em. CRs are meant to provide
adequite rewards (within the limits of the experience system), and provide a
measure of challenge (it`s questionable whether they do this well). It
wasn`t meant to dictate how the game should be played, what kind of
encounters you can and can`t use in your games, or how often PCs should be
put into life-threatening situations.

-Lord Rahvin

************************************************** **************************
The Birthright Homepage: http://www.birthright.net
Birthright-l Archives: http://oracle.wizards.com/archives/birthright-l.html
To unsubscribe, send email to LISTSERV@ORACLE.WIZARDS.COM
with UNSUB BIRTHRIGHT-L in the body of the message.

[Mark_Aurel]

I definitely agree with most of what you've said here. I don't have d20 Modern yet, so I don't know how that works out; however, all that you say on using CRs and ELs is definitely true. I've seen lots of arguments about CR and EL over a long period of time, in various fora - personally, I find them to be a good tool, though with certain flaws - the main thing people seem to be forgetting is that it isn't a strictly mathemathically accurate construct; you can't universally apply the same CR to all monsters for all parties. If you have a party of four clerics, undead certainly aren't the same challenge as they'd be to a party of four rogues, or bards.

CR is basically two things - it's a tool for rewarding XP, and it's a tool for building dungeons and encounters. It works pretty well for rewarding XP, and less well for building "perfect" encounters; these assume that parties have a certain type of composition; if the actual party composition is different, the level of challenge is also different; over a period of time, the level of challenge will probably even out, though, so always adjusting the CR to take into account party abilities or "walk-over" factor isn't always wise; players will end up not making much of an effort if they actually get less of a reward for making a challenging encounter less challenging through a clever plan or idea.

In my campaign, the fights tend to be pretty hard - there's a good chance of PC casualty in most combat encounters; I'd say that more than 50% of encounters fall above the standard EL for the party level, though few encounters are truly overpowering. Occassionally, though, it's ok to toss the players a goblin or three in order to actually let them know they're in a "realistic" world where they can run into dangers of any level. Works fairly well for my group, especially with a good sampling of non-combat challenges and negotiation situations.

I definitely agree about NPC classes and CR - there's more to it than that, too. In order for a given NPC to be of a certain CR, he must have proper gear for his level - which is generally more than what you get for killing monsters of the same CR. Which means killing an NPC is generally more rewarding than killing a monster. That's one aspect of it; it means you either use it as is, tone down the NPC level of gear, and thus their CR (a naked 20th-level fighter is not a CR 20 encounter - it may be a guy that can pummel you good, and soak up a lot of damage, but a 20th-level party will just walk all over him; probably a 15th-level party too, and maybe a 12th or 10th-level party). Furthermore, while the CR system can be seen as built on the notion of character levels, and one level of a PC class is roughly one "point" of CR, the relative level of challenge posed by NPCs, especially higher-level ones, and especially multi-classed ones, has a certain drop-off effect; the system still appears to work best at middle levels (5-10). A cleric 8/wizard 8/rogue 4 isn't anywhere near as challenging as a single-classed character of the same level. A theorethical 25th-level fighter/16th-level wizard does not make a CR 41 encounter just because of the character level; I'm more fond of calculating CRs according to existing standards. Comparing the aforementioned theorethical individual to Asmodeus or Demogoron (from the BoVD) probably yields a lower level of challenge than those two, who are at 32 and 30. It also probably yields a lower level of challenge than most standard wyrm and great wyrm dragons, who are by themselves practically 40th-level fighters/17th- or 19th-level sorcerers (with access to cleric spells, Heal is especially nasty for a dragon), and some other powers to boot. That's a complete tangent, though, but one I've been giving some thought to lately, for various reasons.

Next, NPC classes. One of the things I amused myself with once, was testing the limits of the city building system in the DMG. As I recall, the highest-level individual possible to make there was a 26th-level commoner. Now, on his own, such a character is probably hardly any more challenging than your average ogre. Give him an NPC level of equipment, though, and you'll have a walking hoard of loot, ripe for the taking. Contemplating the 20th-level commoner farmer with a vorpal scythe that he uses for harvest is funny, though. I'm more fond of adjudicating the CR of any given NPC on the fly than adhering to any strict mathemathical formula for that. Commoners are worth the least, level by level. All the other classes are roughly balanced with each other, and get a slightly higher CR, though nowhere near what a PC class gets. If I say a 20th-level commoner with no significant gear is about equal to a CR 3 or CR 4 encounter, that sounds more than fair to me. A 20th-level adept is worth a lot more - maybe CR 10 or so. If you give either some equipment, their CR goes up a bit, but so does the reward they provide.

All that is strictly based on the idea of finding a CR that represents an accurate challenge, as roughly defined by the book (20% of PC resources spent - meaning basically, some hit points, some spells, possibly some charges or potions) and exemplified in the MM, in how monster CRs work out. I may be wrong to do so, but I feel it works better this way, and was probably intended to be used more in a fashion like this, than as a strict mathemathical construct. We're dealing with an enormous number of variables here, where setting a universal CR for the purpose of anticipating all kinds of encounter design is really impossible, but finding the actual CR for the purpose of giving out XP isn't really all that tricky, once you get used to it, and, I think, is fairly accurate as well.

Another issue that deals with CR is the rate of advancement. For some types of campaigns, I feel that the 3e rate is generally a bit quick. For most campaigns, however, I rather like the progression in the system - it means that we can switch between different campaigns more often, and still feel like we're "getting somewhere" - my group has always been fond of running different styles of games interchangeably; the 3e rate of advancement supports this more nicely than previous editions did. It also lends a more heroic, and less soap opera-like style to the game. Of course, some people like soap operas, and I can be weak for those too (J.R. was a tricky guy), and I would never hold it against anyone that likes to change the rate of advancement like that - whatever way is fun to play is fine.

The system itself has some flaws, but I think a lot of the people that complain about it forget that there are two ends to it - encounter design and experience reward; the former is pretty much impossible to always get right; the latter is pretty easy to get right, and also extremely easy to tweak, unlike previous editions and their rather awkward way of determining XP required to advance.