Reply to thread

Message

<blockquote data-quote="FormerlyHemlock" data-source="post: 7134905" data-attributes="member: 6787650">(1) Please quote someone claiming that rolling will &quot;always&quot; give you &quot;better&quot; numbers. That's so obviously untrue that it's hard to imagine anyone even writing that, let alone it being the &quot;biggest [most common?] lie on this forum.&quot;(2) The math says the opposite of what you say: on average you will get better numbers than with point buy. 56% of the time, rolled stats will get at least one number better than ANY PHB point buy PC has. This is a well-known and provable mathematical result (<a href="http://anydice.com/articles/4d6-drop-lowest/" target="_blank">http://anydice.com/articles/4d6-drop-lowest/</a>). Furthermore, when I took your very own point-buy valuations and did a Monte Carlo sim rolling up a million PCs (<a href="https://repl.it/I9Sd/1" target="_blank">https://repl.it/I9Sd/1</a>), I consistently get rolled PCs who, according to your own valuations, are worth 31 points on average. You lose 4 points by using point-buy.Are you claiming that 27 point-buy and 31 point-buy are about the same &quot;on average&quot;? (Of course, the rolled stats are allocated less efficiently, which is part of the point of using rolled stats--less min-maxing occurs.)That you can claim with a straight face that &quot;the math says&quot; rolled stats should be &quot;about the same&quot; as point buy characters, and that rolled stats which aren't comparable to point buy are signs of cheating, is remarkable. Both closed-form math and Monte Carlo approaches agree: the math says the exact opposite.</blockquote>

[QUOTE="FormerlyHemlock, post: 7134905, member: 6787650"] (1) Please quote someone claiming that rolling will "always" give you "better" numbers. That's so obviously untrue that it's hard to imagine anyone even writing that, let alone it being the "biggest [most common?] lie on this forum." (2) The math says the opposite of what you say: on average you will get better numbers than with point buy. 56% of the time, rolled stats will get at least one number better than ANY PHB point buy PC has. This is a well-known and provable mathematical result ([url]http://anydice.com/articles/4d6-drop-lowest/[/url]). Furthermore, when I took your very own point-buy valuations and did a Monte Carlo sim rolling up a million PCs ([url]https://repl.it/I9Sd/1[/url]), I consistently get rolled PCs who, according to your own valuations, are worth 31 points on average. You lose 4 points by using point-buy. Are you claiming that 27 point-buy and 31 point-buy are about the same "on average"? (Of course, the rolled stats are allocated less efficiently, which is part of the point of using rolled stats--less min-maxing occurs.) That you can claim with a straight face that "the math says" rolled stats should be "about the same" as point buy characters, and that rolled stats which aren't comparable to point buy are signs of cheating, is remarkable. Both closed-form math and Monte Carlo approaches agree: the math says the exact opposite. [/QUOTE]

Verification