How does Pathfinder gameplay change when stats are rolled 3d6 in order after selecting race?

Style of play

  • Sandbox setting, estimating risk and reward is up to players.
  • New characters start at first level. Experience from treasure.
  • Character stable: Single player can have several characters, though typically only one and sometimes two are in active play at once.
  • Character death is expected part of play.
  • GM acts as a neutral referee, in the sense that they will not adjust the contents of the game world or fudge dice rolls, or otherwise play to benefit or harm the chances of success of the player characters.

House rules, more detail

To create a character:

  1. Select race (core races, maybe kobolds, goblins and orcs, or maybe only a subset of core races available).
  2. Roll stats, 3d6 in order.
  3. Continue character generation as usual. All Paizo-created Pathfinder material is, as a default assumption, okay. For controversial rules interpretations, consult the GM.
  4. Ability score requirements of feats are reduced by 2. For example, power attack requires strength 11+, not 13+.
  5. 0th level spells have limited uses per day.

How will this affect play?

The obvious consequences are:

  • Humans are a more desirable character species, as they get to place their stat bonus after rolling their stats. This is intended.
  • CR and such estimates don't work as designed. This is of no consequence, considering the sandbox style of play and the role of referee.

What else?

As usual, please answer based on personal experience or published play reports.

  • \$\begingroup\$ I am flagging this as opinion-based. Because none of us will have the required experience to answer this. I have GMed two games with 3d6 in order (2-3 sessions each), the classes and dice rolls during actual gameplay were far more relevant to the outcome than anything else at char creation. For instance, one character had 3 ability scores under 7, but got an int 18, he choose a wizard and had little disvantages over any other Pathfinder wizard. One game ended in a TPK (with weaker characters overall), and the other we simply called quits because two players decided to leave. \$\endgroup\$
    – ShadowKras
    Commented Aug 31, 2017 at 18:15
  • 4
    \$\begingroup\$ @ShadowKras That relevant experience may be hard to find doesn't make a question primarily opinion-based -- it just means it's going to get less answers. (If people start answering with opinion in lieu of any experience, I'll hope for the community's assistance in stopping that.) In point of fact I'm confident many members have played in 3d6-in-order games and understand the implications, and many of the same members will have also played Pathfinder and can reasonably understand how they would affect each other -- so we'll have that at a minimum. \$\endgroup\$ Commented Aug 31, 2017 at 18:42
  • 1
    \$\begingroup\$ @doppelgreener I understand that. But so far, all anwers are correct, in a way or another. \$\endgroup\$
    – ShadowKras
    Commented Aug 31, 2017 at 19:04
  • 1
    \$\begingroup\$ Comments are not for extended discussion; this conversation has been moved to chat. \$\endgroup\$ Commented Sep 1, 2017 at 4:43
  • 2
    \$\begingroup\$ It's not necessary to further (or at all) lecture the asker about their chosen way of playing — doing so only indicates that the lecturer is somewhat unaware of the variety of RPGs that already do this in a way that people enjoy, and therefore verges on unintentional playstyle warring / “badwrongfun” accusations that are not what we expect from our community members. Stick to using comments to ask for non-judgemental clarification. Thanks! \$\endgroup\$ Commented Sep 1, 2017 at 4:46

5 Answers 5


I had experience with this kind of playstyle back in DnD 3.5. In our case the main problem was that in the first few levels, we went through an incredibly high amount of characters (one of the players had 5 in one session :) ) and this created some monsters in the end (one character had 18,17,16,16 and two stats above 12 in the end) and when you end up with something good, you will start to become increasingly wary and paranoid.

Being wary is not necessary a bad thing, though, and I enjoyed playing in a group where all of our battles were carefully planed out involving crazy traps, potions, and other limited use magic items (which we barely use in more recent systems).

Things that you should be very careful of include:

A bit later in the adventure we started to create characters with very boring and stereotypical backgrounds (as they will die soon anyway). So if you want REAL characters you might want to ask them for example to make up one interesting fact about their character and work that thing out later in the game (one character might be the son of a pirate and later in the game you can introduce some NPCs that support this statement, pirate hunters looking for his dad, other pirates looking for his treasure etc.) This is what we used eventually with limited success.

You should also keep in mind that the players of really bad characters may deliberately try to kill their character so they can roll a better one (maybe even going so far as to get themselves killed 5 times in one session). You can circumvent that with a rule that lets you reroll a character if it is COMPLETELY terrible or one that punishes character death somehow (but I would avoid things like that without the full consent of the players).

How this will affect the game and your players otherwise

  • In our game, we almost exclusively played humans because of the random nature of the style and the flexibility of the race. And always picked tier one spell casters if we had the stats to support them (we were relatively lucky because the insanely statted character I mentioned earlier was a dwarf cleric, so healing and tanking was mostly covered just by him; it was also if I remember correctly only his 3rd character).
  • We played until 5-6 level so I can’t say too much about higher level play, but extrapolating, I see a very optimized party that blasts through encounters with ease, but is still afraid to attack two goblins without a plan, which really slows down the game (it was one of the reasons why we stopped playing).
  • This style can also put pressure at your players because if the party is already at a decent level and nicely statted, mistakes that risk a TPK will meet with harsh anger even from mild tempered players (the other reason why we stopped).

With my experience I would not recommend this playstyle to anyone, but if your players are all into it, it can be fun for a time.

  • 2
    \$\begingroup\$ I would add another 2 remarks : Resentment between players with very different/usable PC, and if there is a penalty for reroll, the risk of one player leading the party for a TPK to get rid of his character without falling behind the others players. Also with this setting, a summoner with his eidolon will be a very good choice. \$\endgroup\$
    – Alkano
    Commented Aug 31, 2017 at 13:25

That is not the spirit of Pathfinder.

The most standard way of playing Pathfinder is with point-buy stats and medium rolls for HP per level. It's definitely less random than its predecessors.

A lvl 1 character is also already quite complex: you get Traits, two feats (as most of your characters will be humans), skill points, class features to choose, starting equipment... It can be done quickly, but not in 5 minutes. Some players need hours just to come up with something that pleases them.

PCs are also more sturdy in general: many class get more hp, there is less ways to die instantly... So it is less expected for them to die in general.

To conclude I would recommend you to use an other system as a basis, but if you really want to (which I agree sounds fun), you have to take care of...

The Unfun of playing an unplayable character

Sometimes, a character has just flaws too big to be funny. One obvious example is having a character with a constitution of 1 (it is possible if you roll the minimum and have a race with a constitution malus). With this -5 to your constitution modifier you will be dead in one shot from any monster, and this problem will never go away with level ups (as you will only gain 1hp on most levels and the monster damages scales up more quickly than that). When you roll such a character you know that he will basically only survive until next fight, so you know there will be no character development with him.

On the best case it means you just wait until next fight, and loose something like half a game session waiting. On the worst case your character manages somehow to survive. It means you will have to wait even longer! Expect the PCs to want to commit suicide, or the bad ones to just never be played. As an example of a really terrible flaw you can have someone with an intelligence score into oblivion and no really high score in his other stats (like only 10s). Such a character is not particularly weak, but will have no skill point, no spell beyond cantrips... Basically the best he can do is sitting waiting to die. Not a nice destiny.

General power level

The PC party will be on average way less powerful than standard PC parties, especially on lvl 1. You have the equivalent of a point-buy 0 but without choosing where you spend your points. As a point of comparison a basic NPC gets a set of abilities: 13, 12, 11, 10, 9, and 8 that are supposed to be adjusted to what the NPC is strong at (more or less the same point-buy but without the random factor). Don't expect the PCs to look like heroes, not even to have heroic tendencies. Expect them to be Average Joes that somehow had to go on an adventure, or even crippled persons (as a human is not supposed to have less than a 7 in your abilities if valid).

Many classes will have a very low probability of being played. Expect a very few number of pure casters (but the rare ones could even be melee mages) and a higher number of builds with pets (as pets keep their normal stats they become quite powerful): Summoner, Hunter, Paladin, Cavalier...

You will probably have less two-handed melee and more sword-and-shield or bow warriors (since those rely less on stats and more on feats).

  • 1
    \$\begingroup\$ Is your answer based around actual play experience or reports of such? \$\endgroup\$
    – Tommi
    Commented Sep 1, 2017 at 3:32
  • \$\begingroup\$ No, it is based on experience with the game in general and some basic maths. \$\endgroup\$ Commented Sep 1, 2017 at 12:17

My experience wasn't with Pathfinder, but with an OSR rule system, so I can only extrapolate on what it might mean for Pathfinder.

That being said, from my experience 3d6 in order has two results based on the two parts.

3d6: Means that you are going to get a lot of stat groups close to average, some with a single very high or very low stat, and a few with a bunch of high/low stats.

So a bunch of characters will have stats that are outright bad (all below 10), and a lot of characters won't have the stats needed for high magic (since your highest spell level is limited by stat). You don't usually see a lot of Int 12 wizards.

Though, pure speculation, I guess that might result in characters starting off as wizard (or other spell-casting classes), and then multiclassing into other complementary classes when they run out of stat. But you can expect to see a lot of fighters, rogues, and other classes which can theoretically contribute, even if they are limited to, say, all 8s in their stats.

In Order: Means that players are unlikely to be able to pick their class ahead of time. If you roll 8 Int, you don't get to be a wizard, even if you were really in the mood for that. Your 18 Str might force you into the role of a fighter, even if you really aren't in the mood to play a fighter again.

Unless your players like (not just tolerate, but like) this kind of thing, an outlet is required. Depending on how your stable works, that might be your outlet, but it might also result in a player simply rolling new characters for the stable until he gets his wizard, and then leaving the others unplayed.

In my case high mortality + a leadership-style feature was the outlet. Either your sucky character died (allowing you to roll a new character who was hopefully less sucky), or you lived long enough to gain a comrade (who hopefully would be less sucky).

In a Pathfinder perspective, I'm thinking the change also makes it much more likely that race is picked based on class (an Int 12->14 wizard becomes quite more capable than 16->18, so you wanna be either elf or human). This might be considered a feature.

Both: Magical items that provide stat boosts are going to be more prized, especially if they can be traded between players/within the stable, as the might allow that below-average or average character to.

You should consider really hard what happens when a player rolls a 5, 6, 12, 8, 8, 7 as his first character. That either needs to somehow be a viable character (which depends on playstyle, classes allowed and other specifics) or he needs to know that he might get a new one in the near future, so he can just try his best, and either succeed despite all odds or get that new character. Players should not be stuck hating their character for multiple sessions.

  • 2
    \$\begingroup\$ My speculation is similar, but I'd like to have more than mere speculation. -1 with no malice intended. \$\endgroup\$
    – Tommi
    Commented Aug 31, 2017 at 11:49
  • 4
    \$\begingroup\$ So you want someone who has specifically played Pathfinder with 3d6 in order? Because, as mentioned, a lot of the above is based on my actual experience playing 3d6 in order with OSR. Anything that isn't Pathfinder specific is actual experience: IE spread of stats and their impact on player agency and satisfaction, how it changes the way players choose their characters, the greater search for boosts, etc. \$\endgroup\$
    – Rubberduck
    Commented Aug 31, 2017 at 11:54
  • 1
    \$\begingroup\$ Any ruleset similar do modern D&D should be enough; character builds, detailed rules, etc. Pathfinder, D&D 3, D&D 5, other similar D20 games. Maybe even D&D 4, though that is quite different. Pathfinder would be best. OSR games are quite different, since they almost always lack the interconnected rules components and the default assumption of supercompetent player characters. \$\endgroup\$
    – Tommi
    Commented Aug 31, 2017 at 11:58
  • 1
    \$\begingroup\$ @thanuir that is not speculation. He is extrapolating his real experience playing to PF. It is allowed in this stack. \$\endgroup\$ Commented Aug 31, 2017 at 16:18
  • 2
    \$\begingroup\$ Older editions of D&D are wildly different from Pathfinder with respect to the importance of ability scores and their values. Assuming the OSR game you are referring to is actually similar to 1e or 2e, your experience simply does not apply at all. \$\endgroup\$
    – KRyan
    Commented Aug 31, 2017 at 16:19

This answer assumes you are familiar with Pathfinder 3d6 in order play, and are asking about 3d6 in order after choosing a race rather than before.

1) Your first "obvious" conclusion is completely wrong. This alternate setup makes humans significantly less valuable as a race, though they remain better than most other options. Humans are one of the best races in baseline Pathfinder, one of the best races in 4d6 drop one fixed, one of the best races in 3d6 fixed, and one of the best races when you have to pick your race first. Half-elves are pretty great, too. Half-orcs, not so much.

2) The primary effect is that players play members of their races with so-called 'average' stats, making the characters even more hyperbolically incompetent than normal 3d6 fixed play. This creation system is roughly equivalent to letting players make characters as normal in 3d6 fixed and then giving them a -2 to their highest stat. Normally, in 3d6 fixed, players are forced to pick the race that best boosts their woefully low stats so as to let them play a character with at least some sort of ability to affect the world. In this version, picking a race with two positive stats instead of one flexible one is a great bet because with two positive racial attribute modifiers you have an ~1/3 chance that your highest rolled stat is made higher, while with a flexible modifier the benefit of flexibility is entirely lost since you have to pick your race (and thus the stat the bonus applies to) before rolling your stats and finding out which one is highest. That means you only have around a ~1/6 chance of doubling up stats. Basically this means that instead of PCs always having a rolled 12+ish in their racial bonus attribute, they will now almost never have their highest stats in the attribute their race is known for, before applying racial adjustments.

The other effects are just an intensification of the sort of against-the-odds sort of play one sees in 3d6 fixed. With even less chances that someone's character subverts the creation system via luck and significantly more incompetent characters on average, expect much one-trick-pony-ness, and extreme optimization away from abilities based on attributes.

I've only played in one game like this, which occurred because the GM in question usually ran 3d6 fixed but it never worked the way he wanted it to, because he didn't actually want the PCs to be sideline characters horribly incompetent and everything but the one thing they specialized for, and we stopped shortly after character creation. He also flippantly allowed "all Paizo material" presumably to make up for the lost player agency, and I rolled a Dwarven Wizard with 5 intelligence and 14 Strength, bonded a Chainsaw, and spent my starting cash on batteries. Then we talked about what he wanted from the game and why he was making the rules he was making and found better ways of doing that.

  • \$\begingroup\$ Humans choose which stat to boost after rolling stats. I added this to the post now. \$\endgroup\$
    – Tommi
    Commented Sep 1, 2017 at 3:36
  • \$\begingroup\$ While I understand why it might not be the best play experience, "chainsaw wizard" did make me laugh. Was forcing unusual choices like that part of the intent of in-order rolls? \$\endgroup\$ Commented Sep 1, 2017 at 6:50

Look at the stats

This anydice script shows the probability distribution of the stats. It was modified from this article about 4d6 drop lowest and is worth a read on its own.

If you compare this with the PC elite array {15, 14, 13, 12, 10, 8}, you will roll this well or better less than \$3\over 10,000\$, that is, 0.03% of the time. Compared to the non-elite (peasant) array {11, 11, 11, 10, 10, 10}, you will roll better less than \$2\over100\$ or 2%.

So this means, that unless you roll stats for adversaries in exactly the same way, 98 out of every 100 PCs have have at least 1 worse stat than the average peasant.

Restricted Choice

Players will have to play the hand they are dealt, no playing a wizard if your Intelligence is less than 10 for example. Similarly, barring the kind of luck that would make me raise an eyebrow, any class with MAD is right out. Presumably, this is what you are looking for so I just make the observation.

In addition, you have to take them as they come. Not all stats are created equal - extremely low Constitution and, to a lesser extent, Dexterity are crippling.

More resources will be spent on boosting stats

Pathfinder depends on PCs having appropriate wealth-by-level. More of this wealth will be channeled into stat boosting items than normal meaning less is available for other items. This may need tweaking.

Not comparable with AD&D

Even though this was the default method in AD&D (AD&D 2e had 4d6 arrange to taste as default) this isn't comparable because stats only had an effect at the very extreme ends (15+ or 6-) - middle of the range stats gave neither a bonus nor a penalty.

Pathfinder has a more linear approach to bonuses/penalties - it is extremely likely (83%) you will have a stat at 8 or below and quite likely (41%) you will have 2 that low and not inconceivable (7.5%) you will have 3 stats giving penalties.

  • 2
    \$\begingroup\$ Could you explain what "better" means for you in the first part? Do you have a 0.2% chance that all your stats are 10+ for example, or is the comparison based on something else? \$\endgroup\$
    – Erik
    Commented Sep 1, 2017 at 6:51
  • \$\begingroup\$ The peasant array and the non-elite npc array are both examples with average 10+1/2, which is exactly the same average as rolling 3d6 in order. Since this distribution is symmetric around its mean, I would assume that roughly half the time the rolled average would be higher than or equal to peasant/npc array. \$\endgroup\$
    – Tommi
    Commented Sep 1, 2017 at 10:41
  • \$\begingroup\$ @Thanuir Half the time one of your stats is better than average. All stats to be better than average happens \${1\over{2^6}}=1.5\%\$. \$\endgroup\$
    – Dale M
    Commented Sep 1, 2017 at 21:11
  • \$\begingroup\$ Your meaning is clearer with the edit; thank you. I still think the sentence "Compared to the non-elite (peasant) array {11, 11, 11, 10, 10, 10}, you will roll better less than 2/100 or 2%" is misleading. \$\endgroup\$
    – Tommi
    Commented Sep 2, 2017 at 10:36
  • \$\begingroup\$ Also, half the time a given stat is higher than the average. The probability of there existing at least one stat which is higher than the average is quite a bit larger. \$\endgroup\$
    – Tommi
    Commented Sep 2, 2017 at 10:37

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .