D&D 5e was heavily play-tested prior to its release. I have heard (but have not verified) that, at low levels, the hit rate of the attack rolls of PCs was intentionally optimized to be approximately 2 in 3. Presumably, this probability was chosen deliberately by WotC, then – through simulations, models, and tests – features were added/removed/tweaked until it was realized according to some metric. I am interested in information about this process.
What information, if any, is publicly available about the play-testing process for D&D 5e and about the outcomes (such as this 2/3 hit-rate) that were optimized for during play-testing?

I am not asking why any rules are the way they are or why any of the particular math was used. I am asking what math was used and what information has been published/leaked/discussed about it.

This question is similar to my question. Its top answer links to pages documenting the design goals of D&D, including some details of the play-testing process (for example, here). However that question is about the general design goals while this question is about the math and the methods of simulating/modeling/quantifying the game. It does to link to relevant information, however.


1 Answer 1



Play testing for D&D Next which became 5e was subject to a non-disclosure agreement. WotC know but they aren’t saying.

  • \$\begingroup\$ I mostly assumed this was the answer but wanted some confirmation of it, thanks! \$\endgroup\$
    – nben
    May 7, 2022 at 4:22
  • 4
    \$\begingroup\$ Is this really relevant? All the actual playtesting has been discussed all over the internet. You can easily get different playtesting documents. And you'd be hard-pressed to convince anyone the NDA is even still enforceable. But those aren't related to the internal decisions WotC made when they changed various rules between playtest versions. Things like this article and this one aren't covered by the NDA. \$\endgroup\$
    – MichaelS
    May 7, 2022 at 11:51
  • 1
    \$\begingroup\$ @MichaelS if you have better information please supply an answer, the best refutation would be a contribution. \$\endgroup\$ May 8, 2022 at 14:42

You must log in to answer this question.

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