Reply to thread

Message

<blockquote data-quote="slobo777" data-source="post: 6031310" data-attributes="member: 6694877">I agree, but have to start somewhere, and that is with the simplest scenarios. I see the current simulations as a small step forward from simplistic &quot;DPR&quot; and &quot;number of rounds on average&quot; etc calculations already used to assess PC/player choices in a lot of technical discussions.I'm currently looking at how to model movement, so the test can have positional play, including ranged attacks, blocking access to &quot;squishies&quot; etc. This will however, take a while, as it gets complex very quickly. My current thoughts are I should just embrace that, and dig out some neural nets, evolutionary programming for self-learning, and create something that can play moderately well on a basic grid. It will be fun, and my programming skills are up to it, but it's going to sidetrack me from adding anything else from the playtest docs for a few months probably.</blockquote>

[QUOTE="slobo777, post: 6031310, member: 6694877"] I agree, but have to start somewhere, and that is with the simplest scenarios. I see the current simulations as a small step forward from simplistic "DPR" and "number of rounds on average" etc calculations already used to assess PC/player choices in a lot of technical discussions. I'm currently looking at how to model movement, so the test can have positional play, including ranged attacks, blocking access to "squishies" etc. This will however, take a while, as it gets complex very quickly. My current thoughts are I should just embrace that, and dig out some neural nets, evolutionary programming for self-learning, and create something that can play moderately well on a basic grid. It will be fun, and my programming skills are up to it, but it's going to sidetrack me from adding anything else from the playtest docs for a few months probably. [/QUOTE]

Verification