The Epidemiology of Pathfinder: Live Fast, Die Young

Pathfinder First Edition General Discussion

Hello fellow adventurers! I have updated my blog with a new set of results, testing three types of fighter and including races and feats. I have simulated 100,000 warriors against 100,000 random orcs, but this time my fighters have what I call "purposively selected" ability scores and feats. "Purposive" means they chose one ability score and the other two were 9 or 3d6, whichever was higher (so mimicking ability scores suited to fighters, but retaining some randomness).

I developed three types of fighter: strong, fast and tough. The results: tough fighters never die unless they're halflings or elves, and focusing on dexterity/agile fighting will lead to a high death rate (around 45% for elves and halflings). The detailed results are at my blog.

My next questions: survival rates for different point buy systems, and what are the relative survival rates of OD&D fighters compared to Pathfinder. My prediction of the latter: OD&D fighters live longer. Want to place bets?

Od&d, 5 internet cookies.

Did you include favored class bonus of 1 hp in your simulation? I didn't see it specifically mentioned, however from your examples of calculated hp it appears that you did.

I think if you included the favored class bonus (fighters will always choose hps), your survival rates will increase and also high constitution will not be quite as important, because with a 14 and the favored class bonus you are guaranteed to not be 1 shotted by the Orc.

I meant that you did not include the favored class hp in my earlier post.

g%#$%@n, there's always some tiny new rule isn't there? I suspect including favoured class will mean that constitution bonuses of 2+ grant higher survival.

Turin the Mad, unfortunately I don't know all the rules of OD&D: I am guessing that fighters get 1d6 hps and all weapons do 1d6 damage, THAC0 20 (but maybe they had weird tables), that Orcs have the same THAC0 as fighters and get 1d8+1 hps (but was it 1d8/2? I can't recall). Also what damage do they do and what are armour classes like? It would be simple to run, but I don't have the necessary information - and I'm not sure if I can be bothered spending $5 on a download of an OSR copy-paste.

I appreciate this evolution in the analysis, though I'm inclined to quibble with the mortality curves shown for Dex and AC--I feel like there are too many variables going into those.

I think you have well answered the question of, "If I'm going up against Gruumsh, should I focus on being fast or being tough?" pretty clearly, but you raise more questions.

* I was born fast. Should I focus my workouts on becoming strong, or becoming tough? (For someone really set on a high-Dex fighter, should they put that extra point in Str or Con? Is a high-Dex, high-Con finesse fighter viable?)

* How many eggs should I put in one basket? "Improved initiative, dodge, and weapon finesse" all at first level, on top of a high Dex score? Maybe I should swap out one of those for Toughness and see how I do? Am especially interested in swapping imp-init for toughness, with the idea that surviving Gruumsh's first strike is more important than reducing the chance that he gets to strike first. (Also, is weapon finesse really the one for non-humans to drop? To be honest, I'm mostly curious as to how valid my bias against imp-init is.)

In short, I feel you've failed to disprove the viability of Dex-fighters--but I say this primarily out of an appreciation for science, rather than an appreciation of Dex-fighters. I wouldn't be challenging your methodology if I didn't appreciate it. :)

Od&d fighters got d10 HD, qualified for weapon proficiencies/specialization - longsword doing 1d8 - and were the only ones to get 18/% STR and benefited from a CON greater than 15 or 16 with extra hp/HD > +2.

Base to-hit # at lvl 1 was 20 for everyone IIRC. Fighters get nasty fast though, outstripping cavaliers, rangers and paladins in short order.

There's good reason why I'm betting the way I am. :)

Ok I looked over your assumptions more in depth.

I think for the non human fast fighters dropping weapon finesse makes no sense. Because they can hardly hit if they drop that. I would recommend changing it to dropping improved initiative for the fast non-humans.

I would have recommended dropping shield focus on the tough non-human fighters but than I realized you still need to incorporate the favored class hp. So toughness is less important.

Turin the Mad wrote:

Od&d fighters got d10 HD, qualified for weapon proficiencies/specialization - longsword doing 1d8 - and were the only ones to get 18/% STR and benefited from a CON greater than 15 or 16 with extra hp/HD > +2.

Base to-hit # at lvl 1 was 20 for everyone IIRC. Fighters get nasty fast though, outstripping cavaliers, rangers and paladins in short order.

There's good reason why I'm betting the way I am. :)

That's AD&D, not OD&D. In OD&D, Fighting Men got 1d6+1 hp, there were no weapon proficiencies or percentile strength, all weapons did 1d6 damage (I think), and everyone got (just) +1 hp/HD for Con 15+. The attack table was ... wonky. (1st-level fighters needed a 10 on a d20 to hit an unarmored man and got either +2 or +3 to hit every 3 levels.)

Ooohh ... fair enough.

I'll move my wager to AD&D Fighters for 5 internet cookies then. ^__^

I don't think your "Fast Fighter" build is going to reflect accurately at all, since NONE of your generated characters are using Dex to hit.

faustusnotes's blog wrote:
Fast fighters: dexterity was determined randomly from a uniform distribution between 13 and 18, and the fighters were equipped with studded leather armour, a heavy wooden shield and a rapier. Human fighters had three feats: improved initiative, dodge and weapon focus. Non-humans dropped weapon finesse, and humans put their +2 bonus into dexterity.

If human fighters don't HAVE weapon finesse, and all non-human races that had it dropped it, then of course the build isn't going to survive. They may not get hit as often, but unless they can hit and kill their opponents then the opponent is just going to survive that much longer to continue dealing damage, working towards that inevitable critical hit.

Your feat selection is severely skewed against Dex-built fighters. I would drop improved initiative and replace it with weapon finesse, and make it so that non-humans drop weapon focus or dodge.

I can appreciate the amount of work that went into this, but I don't know what you hoped to prove or disprove. I do not think your variables/parameters are appropriate for the test.

EDIT: I have problems with the Strong and Tough fighter sections as well, but this one is the most wrecked of the three.

Foghammer, sorry, that's a massive typo in my post (what can I say, it was very late and I'd been working till 11:30 pm). Fast fighters had weapon finesse, not weapon focus - I just wrote that wrong (there may be two copy-paste errors from the R code, actually), and I'll change it forthwith.

I'd be interested to hear the problems with strong and tough too -if you think there's a better tough layout then I'd like to hear it (I don't think there is - these guys had very high survival rates!)

Gignere, good idea, I'll give that a go.

Murph, my suspicion is that fighters who have high dex and low strength/toughness are much better off getting the toughness and weapon focus feats and letting their dex do what it can for ac only.

For those who love fast fighters, if you can come up with a build that you think works, that doesn't involve vast amounts of sensitive programming (I'm a s+~! programmer) then I'll give it a go. Just lay down the feats!

Thanks again for the comments!

Distant Scholar, can you tell me the hit rules for fighters and orcs, and the stats for orcs? If so I can do it...

Another issue: The Fast Fighter is using a Heavy Wooden Shield. Note that if you carry a shield, its armor check penalty applies to your attack rolls when using Weapon Finesse. Of course, they can get one point of AC from losing the shield back by using a Chain Shirt instead of Studded Leather. Only Elves and Halflings can get Dex high enough to surpass the Max Dex of a Chain Shirt, and +3/+5 and +4/+4 lead both give +8 AC regardless.

Well, having given it more thought, I suppose the only issues I had were the weapon finesse mix-up, and that non-humans drop toughness as tough tighters, regardless of gnomes or dwarves bonuses. I'd sooner have them drop weapon focus for the purpose of the test.

Originally I had confused a feat in the strong fighter's selection (desperate battler) for something else, but I went back and double checked, so that's not an issue I have anymore.

faustusnotes wrote:
Distant Scholar, can you tell me the hit rules for fighters and orcs, and the stats for orcs? If so I can do it...

Well, there's not much to them.

A 1st-level fighting man will have 1d6+1 hp, +1 if Con is 15 or greater, -1 if Con is 6 or less, and will hit an orc on a 13 or higher (on a d20). He can likely afford any armor class from 2 to 9 (2 being the best).

Strength and Dexterity have no effect on combat(!) unless ranged attacks are used, in which case Dexterity can do a little.

Orcs have 1d6 hp, and hit AC N on a 19-N.

All attacks do 1d6 damage.

Now, once the Greyhawk supplement came out, more variety was added, but I don't have those advanced rules.

1 person marked this as a favorite.

I'll just link to this guy's blog where he took a fighter from (just about) every edition of D&D (0E to 4E) and had them face off against an endless wave of goblins. The results are interesting:

The Evolution of the Fighter in D&D

Awwww maan ... no cookies for me.

So, are we saying now that CON is the best stat?

It looks from the data that as along as you don't dump your STR, CON is a very important stat.

Am I reading this correctly? Strange that the outcome would be so dramatically different.

Are we sure orcs are the best comparison for this There to hit is rather good. If you were harder to hit +1 Ac would reduce average damage of the soemthing like goblins even more while comparitively less of orcs.

Mort, I ignored all armour check penalties - the wooden shield is purely AC bonus. I think that armour check penalty idea is ridiculous actually and if I ever were misfortunate enough to read it, I would immediately houserule it out. Incorporating that would further benefit the CON stat over all others, I think.

Foghammer, that desperate battler feat is a very useful one in this simulation - if one were making a PC for group adventure one would obviously never choose that feat.

In order to explore the particular effect of any given feat the best option would be to go back to my original model, where I randomly generate everything and then randomly add one feat to any PC (I could even drop the minimum str and dex requirements). Then one can see just how useful it is.

I think my next step will be the od&d fighter. That link of hogarth's is very interesting but I don't think goblins are the right comparison, because survival rates are high (I have to guess at the survival rates though because the post doesn't give them directly).

Globetrotter, the change from CON being a dump stat to the best stat came about through a) the maximum hit points and b) the purposive selection of ability scores. The former means a con bonus of +3 or more is very important and the latter means that fighters with high CON will always have reasonable attack bonuses (especially after incorporating feats), so if they can survive one blow they will get enough time to slay their opponent.

Distant scholar, thanks for the info. I think that will be enough for a simulation. Did strength give a +1 to hit/damage for scores above 15?

faustusnotes wrote:
Distant scholar, thanks for the info. I think that will be enough for a simulation. Did strength give a +1 to hit/damage for scores above 15?

No, it did not. Strength did nothing for combat.

Just checking that you are aware that orcs fight till -12 hp (as they dont stop fighting once they hit negatives) which means the average orc takes 18 points of damage to put him out of the fight (being more damage than any most fighters can put out in 1 round at level 1 2d6+9 is only 21 points at max without a crit), meaning that your fighter mortality rates should probably be somewhat higher.

Michael: yes, I was aware of that but haven't incorporated it because it's a pain in the arse. When I do the ODD vs. Pathfinder thing I will do it though.

I originally thought it wouldn't matter because they get penalties as time passes, but now I'm thinking the penalties won't outweigh the extra rounds of damage capability. So next iteration I will do that.

Also, current survival rates are at 80% for slightly over-powered PCs, so incorporating ferocity probably means the orc isn't CR 1/3. I need to test Gruumsh to destruction against increasing levels of fighter to work out what his true CR is.

I find it really hard to beleive that an OD&D fighter is weaker than a pathfinder fighter once you incorporate ferocity!

Scarab Sages

Absolutely the OD&D fighter survives longer. My prediction is somewhere from 3-5x the number of rounds. Just the difference in ability score modifiers results in a basic fight taking about 3 rounds in Pathfinder and 10 rounds in 2nd Edition.

I'm not surprised the Dex build had a lower survival rate, as the AC system is binary. You're much better off guaranteed survival of two hits, and even better off if you can kill the opponent in one.

doctor_wu wrote:
Are we sure orcs are the best comparison for this There to hit is rather good.

I'm also a little skeptical about using orcs: a hit from an orc does 9 damage on an average, non-critical hit, which is almost enough to one-shot a fighter (ignoring Con bonuses). I would argue that that's the exception rather than the rule for most of the creatures a level 1 fighter would face, and it tends to overemphasize the importance of Con (if one or two extra hit points allow you to survive 2 hits instead of 1, you've doubled your effectiveness, sort of).

By the way, are you still in Japan faustusnotes?

Jal Dorak wrote:
Absolutely the OD&D fighter survives longer. My prediction is somewhere from 3-5x the number of rounds.

Talking about survival in terms of number of rounds is fairly meaningless, IMO. A round where the fighter and the orc both miss is not a net positive for either side and shouldn't be counted as such, if you ask me.

I'm using Orcs because they're the quintessential 1st level opponent. Who hasn't had their day ruined by an orc? Also, the damage feeling is entirely mutual - Gruumsh can gank a fighter in one hit, but a fighter can whack Gruumsh too. Whereas goblins are a pack of ankle-biting weirdos who take all day to get the courage up to spit.

A commenter on my blog has suggested testing them to destruction - start with kobolds and work up until survival reaches 0. I have plans for something like that. Currently I'm testing the addition of ferocity to see how much more ferocious it makes an Orc.

Hogarth, yes I'm still in Japan, but not GMing pathfinder in Japanese at hte moment - I'm playing D&D3 homebrew and GMing WFRP3, both in English.

oh and hogarth, as an epidemiologist I'm outcome focused - so I agree with you and I prefer to focus on the disease outcome (death) rather than the process (rounds). If I had incomplete follow-up data I would have to incorporate survival time, but that's not an issue here. So I'm focusing on survival rates.

Hi folks, I think I should let you know that a tiny programming error (3 letters in 375 lines of code) means that my results in this post were wrong :( I have updated the results (briefly, because I'm tired) at my blog, the general conclusion remains the same but constitution is no longer the go-to stat, just not as useless as dexterity.

The programs are cleaned up and sometime in the next few days I will have new material on orc ferocity - an interesting example of how much difference a single ability can make.

Sorry for the errors! good thing this isn't a peer-reviewed journal, or i'd be in deep do-do!

Am I reading your updated blog wrong? It still shows CON to be king.

Is that the conclusion of the analysis?

Globetrotter wrote:

Am I reading your updated blog wrong? It still shows CON to be king.

Is that the conclusion of the analysis?

Here's the updated blog post: -critical-revision/


Perhaps when you figure in "fast fighters" you also should include "ranged fighters" - see if goblinoids with 30' distance weapons can beat / outkill 100' ranged weapons. Or point blank weapons with Dex mod to damage... I don't quite understand what you are testing besides "fighters with more HP and STR live longer in melee". That is sort of a no-brainer, IMHO. Giving an archer a sword and saying "do your best" vesus a sword master doesn't take a lot of thought...

That's what I was missing.

These new numbers hold closer to what the original idea was. I always wonder about the importance of CON vs DEX.

I would love to see a 10th level comparison to see if the data holds constant.

It seems to me like you´ve not included Crits in your calculations.In your Blog you wrote the Orc is using a Falchion ,so if you simulate 100000 fights a Survival rate of 100% for the Con focused fighters should be pretty much impossible.

Liberty's Edge

As others have mentioned, ditching Weapon Finesse from your non-Human Fighters is a huge mistake, and the reason the Elf and Halfling ones have such horrible survival odds. Replace their Improved Initiative with Weapon Finesse and I suspect those numbers will look a lot less bad.

maouse, I'm going to get to this later - perhaps for Rangers - on the general idea that each fight should start in a situation that is most beneficial to the PC: thieves in stealth mode, wizards at a distance, etc. Then definitely the fast fighters will get a better advantage.

Globetrotter, I want to do that too. I think CON will become more important and dex and str less so. CON is the only stat whose benefits scale with level, after all.

Sleet storm, crits are included. I gave the fast fighters a rapier so they could maximize their crit chance, and I wonder whether their survival would be radically different if they had improved crit at first level. I haven't checked to see how important criticals are, but a possible way to do that would be to set two featless, completely random fighters against each other and give one of them crits and one of them no crits. Any survival difference would then be due to critical hit ability.

Deadmanwalking, when I get a bit of time I will investigate that. I think that elf and halflings have terrible odds because elves have too few hps (-2 con) and halflings do significantly less damage due to size. The key to survival against Gruumsh is being able to smash him into the dirt in as few rounds as possible, and the key to that is damage.

addendum to Deadmanwalking: however you're right, it could be that feat selection is much more important for non-humans at first level, because they only get two.

I am also considering a very long simulation where I randomly assign each fighter a single feat, and then assess survival for a very large set of fighters (maybe 10 million - the entire population of Faerun?). Then they can be ranked. If I did that on my work computer (a PowerPC with 32Gb RAM and 12 cores) I could probably also extend the simulation to every possible combination of the feats I've incorporated so far, for every possible combination of ability scores. Maybe I should do that ...

Faustusnotes, might I suggest a significant revision in your experimental procedure? I think your experimental design could be much more focused. Rather than randomly generated fighters, you should try using specific NPC Warriors, so as to focus in on specific abilities

Test A - 1000 fights between orc and a warrior with all 10 ability scores (and a trivial feat - like skill focus (basket weaving).
Test B - 1000 fights between orc and a warrior with a 12 strength and 10 in all other ability scores.
Test C - 1000 fights between orc and a warrior with a 14 strength and 10 in all other ability scores.


and so forth, for a full variety of strength scores, and dex scores, and con scores. This would allow calculation of specific gains in survivability for each increase in ability score. With such a focused test, you might get away with as few as 100 fights per test.

Similarly, you could look at permutations of ability scores (eg 16 str, 12 Dex, 8 con), and a similar routine would work through assessing feats individually, and then in concert

Note that such a repeating test would lend itself very nicely to programming using loops.

Hi folks

I've added a post to my blog in which I compare non-human fighters with and without weapon finesse. As most of you expected, dropping weapon finesse in favour of improved initiative was a bad idea, and increased the odds of mortality against (non-ferocious) orcs by about 30 - 60%.

The details are here

Next I'll be comparing different point buy mechanisms.

Community / Forums / Pathfinder / Pathfinder First Edition / General Discussion / The Epidemiology of Pathfinder: Live Fast, Die Young All Messageboards

Want to post a reply? Sign in.