Reply to thread

Message

<blockquote data-quote="mearls" data-source="post: 9861696" data-attributes="member: 697">Public playtesting has transformed into a brute force tool. Here's the game. Tell us what you think!It's a lot more subtle than that. It works best when you can present people multiple versions of something over time, or a stack of options from which they pick the ones they like best.I don't think this scenario works:&quot;Here are three subclasses for an upcoming book. Give us feedback on them.&quot;This works much better IME:&quot;Here are 10 subclasses. Rank them in your order of preference.&quot; The top three are then published.The core 5e playtest, and the one for Xanathar's, followed the second model. Stuff like the ship rules from Saltmarsh followed model A and delivered meh results.The trick is building a content pipeline that is focused on volume up front (make lots of stuff!) and refinement at the bottom (few things survive, but the ones that do are highly polished).</blockquote>

[QUOTE="mearls, post: 9861696, member: 697"] Public playtesting has transformed into a brute force tool. Here's the game. Tell us what you think! It's a lot more subtle than that. It works best when you can present people multiple versions of something over time, or a stack of options from which they pick the ones they like best. I don't think this scenario works: "Here are three subclasses for an upcoming book. Give us feedback on them." This works much better IME: "Here are 10 subclasses. Rank them in your order of preference." The top three are then published. The core 5e playtest, and the one for Xanathar's, followed the second model. Stuff like the ship rules from Saltmarsh followed model A and delivered meh results. The trick is building a content pipeline that is focused on volume up front (make lots of stuff!) and refinement at the bottom (few things survive, but the ones that do are highly polished). [/QUOTE]

Verification