Average Adventurers are Anything But Average (in D&D 5e)

One of the first issues facing any new campaign is how the characters will be created.  I believe there are actually two issues that tend to be addressed a bit sloppily together when it comes to establishing character stats: (1) The technique to come up with values, and, (2) how to make the characters heroic, or amazing, or simply better than average - as the player who wants to play a below average character is fairly rare.  

A similar issue also arises on hit points and in some games critical hits.  Both of these aspects are oh so tempting to buff to make the player characters feel more heroic. 

While the DM can always, with effort, scale challenges to fit any inflated "heroic" party, that just seems unnecessary to me.  The game system also tends to break more easily at its limits.  A character with multiple 20s in stats is going to be taking feats which when stacked can result in synergies that are both fun (for some) and game-breaking for everyone.

In this post I will be spelling out my view on the following topics, generally advocating an average approach, which given that adventurers are far, far from average, is actually heroic.   

1. Starting Stats
2. Hit Points per Level
3. Critical Hits
   

Starting Stats

The PHB defines three methods for initial stats:

1. Roll 4d6, sum three, place in any stat, repeat six times;
2. Arrange stats from the standard array (15, 14, 13, 12, 10, 8) as desired;
3. Point buy with higher scores costing more.

I intensely dislike the first one as it assures players start with different resources and encourages the "tossing" of inadequate results until acceptable stats are obtained.  This typical has a player unhappy that they have too many low values or not enough high numbers to make a viable, they appeal to the DM who often ok's rolling another character.  The cycle can of course repeat.  Eventually, we have a character who is well above average for new characters and primed to push the boundaries of the game.  I could go on, but I'll stop here and point out that I don't allow this technique, period.

The standard array is pretty darn standard.  I like this idea for simplicity and working well in 5E's world of bounded accuracy.  Many races have a +2 and a +1 bonus, so this array might become 16-16-13-12-10-8 which would usually make a prime stat 16, CON or DEX 16, the other 13 and the other three get what is left.  That's not bad.  I've often allowed the use of what I call the improved array: 17-15-13-12-10-8.  This array makes it possible to get those desirable stats even higher while keeping the low end low. 

Point buy allows for more customization while keeping things fairly much in range.  The standard 27 points allows a start with three strong stats: 15-15-15-8-8-8 or a rather flat:13-13-13-12-12-12, or anything in between including the standard array.  The budget can be changed, often inflated to make more "heroic" characters.  Overall, it just feels unnecessarily complicated providing a bit of choice. 

Point buy doesn't allow higher than 15 starting stats. If the table of costs was extended, a stat of 16 might cost 11 and 17 might cost 13 points.  This in turn, implies a budget of 33 to purchase those stats.  I suppose this could be referred to as improved point buy, the general method of obtaining the improved array of stats.

 I end up feeling ok with either the normal or improved array or point buy for starting stats, though for simplicity, I favor the standard array.  

Odd vs Even Stats

One thing to note when looking at stats is the mile stone effect of reaching even values.  The way most of the modifiers work, each even number is exactly as good as the odd value one higher, e.g. a 10 and an 11 are both a +1 to related skills.  The standard array has a pair of odds which are tangibly improved by a racial bump, while my improved array has three odd values.  This allows most races to leverage their stat bonuses to better modifiers.  

The major outlier, the standard old human with +1 to all six stats gets no benefit out of most of those increases with the standard arrays.  Standard humans who use point buy can arrange to have more of their stats with odd values to benefit from their racial bumps.  I really like this effect. 

Extending the Point Buy Table

As I mentioned earlier, I can imagine an extension of the PHB's table to allow 16 and 17 stats with costs of 11 and 13 respectively.  Extending the table allows more extreme stat arrays like: 17-17-9-8-8-8 at 27 points or 17-17-14-8-8 at a 33 point budget.  Stats like that would make a MinMaxer smile and make me sad, so I need some more restrictions, perhaps: No stat over 15 unless the stat buy budget is at least 33 points and then only one is allowed. That to some degree forces spending of points on other stats. 

Sweetening the Fixed Arrays

Another way to encourage the fixed array vs point buy would be to make them a bit over budget.  If this is done on the low end, it doesn't really affect player power and could reward that lowly human by making more stats odd to start with.  I think I want to go with a rules statement much like this:

Initial stats will be set by party size to be either (DM sets budget, player choices method):

1. Standard Array (15-14-13-13-11-9) or 27 point buy,
2. Improved Array (17-15-13-13-11-9) or 33 point buy, only one stat over 15 allowed.
   

In both cases the arrays are three points over budget by bumping the lowest stats up to odd values, not a lot of help, but a bit of a reward for working with spread of abilities.

My preference will be the standard array with a group size of five or more players.  Improved array for lessor numbers.

Hit Points Per Level

By RAW all characters start off with maximum hit points on their first hit die at first level, a Fighter will always have 10 plus his CON modifier hit points at 1st, a Wizard will have 6, etc.  That is a solid change from earlier versions of D&D where rolling a one at first level was very possible and likely a death sentence for the character. 

Hit points gained at levels beyond one are, by RAW, a straight die roll.  That can easily lead to good or bad luck results, though, in the long run it tends toward average results. The small sample size of rolls (no more than 19 and often far fewer) makes really sturdy or fragile characters all too likely. 

RAW does allow an alternative to take the average (rounded up) result from each hit die rolled while leveling up.  This results in a Fighter with a d10, gaining 6 hit points each time they level up.  This results in a calculable hit point total (no wondering if you remembered to roll a die or maybe did it twice leveling up) and it avoids extreme results which is good for game balance.

Its tempting to take both systems as I have done in the past in an attempt to make it better.  Allow the player to roll and if it is better than average use that, otherwise the fixed value is used.  This avoids really low results and allows for the fun die roll.  The problem is it inflates hit points per level for the characters.  That 10th level fighter (with a con of 11... don't ask) would have the following hit points with the three systems:

• Standard Rolling: 60 average (19 to 100) hit points 
• Fixed Gain: 64 hit points
  
• Hybrid Described: 73 average (64 to 100) hit points

The straight rolling system allows a really wide range.  This creates pressure to allow low results to be rerolled which inflates the HP total of the character. The hybrid raises the character's hit points on average 15% and upto 56%.  Both of these approaches devalue in game healing and force the DM to adjust the difficulty of encounters.  

I am not a fan of either effect, so I want to go with Fixed HP gains per level.

Critical Hits

Critical hits are fun.  Roll that Nat 20 and you're going to do something awesome, like double damage! But wait, what if you roll snake eyes right after that awesome critical hit?  That pretty much sucks.  It's also totally possible by RAW:

When you score a critical hit, you get to roll extra dice for the attack's damage against the target. Roll all of the attack's damage dice twice and add them together. Then add any relevant modifiers as normal. To speed up play, you can roll all the damage dice at once.

A snake-eye'ed critical hit is the driving reason for a variety of different rules for dealing with the damage.  Here are versions I have seen in use:

1. Double the result of all dice rolled,
2. Roll all the dice twice (RAW),
3. Roll all the dice twice with a minimum result one more than max of normal hit,
4. Roll the normal dice and add the maximum that can be obtained on the dice.

Approach 1 is optimally bad as it increases the odds of a snake-eye'ed roll and of a max.  I'm done thinking about that one.

Approach 2 is exactly RAW.  It increases the average damage, but still allows the unlikely double one roll.

Approach 3 forces crits to be more damaging than normal hits without increasing the probability of a massive damage roll.  About half the time crits will be one higher than a max hit.  

Approach 4 guarantees impactful critical hits.  It significantly increases the probability of double weapon max damage.  This creates opportunity for very swingy damage which is generally bad for survival of our heroes. 

Approach 5, which I didn't list as I refuse to do it, would be critical hit tables typically with devastating results possible up to and including instant kills. While this can be fun it is bad for the players and difficult for the DM as a more extreme approach.

Why Swingy Damage is B-A-D

Our heroes are involved in many many encounters through out a campaign.  They win most of the encounters, rarely being killed and hopefully having few if any TPKs.  This means on average, the heroes win their fights.  The DM typically sets encounter difficulty such that the party winning is likely.  This is all generally a good thing and works out well.  

If the players get lucky with crits they win more quickly than expected, at worst this leads to an anticlimax, which likely does not negatively affect the arc of the game. If the monsters get lucky with crits players can easily be killed, for example an Adult Red Dragon on average bites for 26 damage, but can crit for 72 a big ouchie.

Big monster crits can quickly put the DM in a quandary, choosing between dead player(s) and fudging the dice. 

Buffed Criticals Favor Certain Damage Types

All of the approaches that raise the average damage of a critical hit favor weapons and attacks that can critical hit.  This helps monsters with larger attack die the most, characters with melee weapons and spells that roll to hit die.  This devalues all of the "roll a save" type attacks in the game.  The same type of attacks that already suffer from a dearth of methods to strengthen them as far as damage goes. 

Any time something is buffed everything else gets comparatively weaker, sometimes in unexpected ways.  

My Preference as DM

I prefer approach 2, 3, 4 in that order and refuse to allow 1 or 5.

RAW handling of criticals (approach 2) is most favorable to the players and easiest for the DM.  It does allow the dreaded snake eyes but it has much less chance of a player being sliced into bits by one lucky hit and doesn't devalue caster damage.

If the players really want 3 or 4, I'll play the campaign that way knowing that any character killed by a critical has signed his or her own death warrant.