Very simple simulator for first level BFRPG combat, used to test out the effects of house rules.

Copyright 2019 Chris Gonnerman

This program simulates a battle between a canonical group of first-level adventurers (all human, one fighter, one cleric, one thief, one magic-user) and a group of four goblins.

It is assumed that the fight takes place in a 10' wide corridor, and that the fighter and cleric form the front line, soaking up all attacks by the goblins as long as they survive. If either of them is killed, the thief moves up and takes the place of the dead adventurer. The magic-user is only subject to attack if two other adventurers are killed.

Ranged attacks can be assigned to characters or monsters, with limited ammunition. Also, our magic-user has access to Magic Missile, which will be cast in his or her first round.

Finally, surprise is ignored in this combat. Initiative is rolled for each combatant, and reach is ignored (so that identical numbers always mean simultaneous attacks).

---

This is an ugly hack, folks. If you want to change the assumptions, try out a house rule, whatever, you edit the source code. No options, no command line business, nothing. And it's probably pretty crappy code too.

When the script runs it generates a sqlite3 database containing the results of the run. This database, sim.db, is deleted before each run! The database has a single table; structurally it looks like this:

create table simulations (
    winner      text,       # one of three values: "pcs", "monsters", "tie"
    rounds      integer,    # how many rounds were fought
    pcs         integer,    # how many PCs survived
    pchp        integer,    # how many hit points the PCs started with
    pcdam       integer,    # how many hit points of damage they took
    monsters    integer,    # how many monsters survived
    monsterhp   integer,    # how many hit points the monsters started with
    monsterdam  integer     # how many hit points of damage they took
);

The script runs 10,000 combats by default; there's a bit of commented-out code that you can use to run two battles back-to-back, with a fresh group of goblins and whatever remains of the player character party.

Generally I start my analysis of the results like this:

select count(*) from simulations where winner = 'pcs';

Now I know how many times they won out of 10,000; if the result is 8,000 (for example), then they won 80% of the time.

Another interesting statistic is how many times all four PCs survived:

select count(*) from simulations where winner = 'pcs' and pcs = 4;

And the average damage the party took when they won:

select avg(pcdam) from simulations where winner = 'pcs';

NOTE CAREFULLY that you can't use double quotes " because sqlite3 treats double quoted values as identifiers; notice that 'pcs' is the value for winner, but pcs (no quotes) is the field name for the number of survivors. You'll get puzzling results if you forget this.