Reply to thread

Message

<blockquote data-quote="Silveras" data-source="post: 2399184" data-attributes="member: 6271">Sure you could have some randomization .. but it would be randomization that reflects the developer's personal preferences, at least initially. Sometimes it would be obvious: &quot;Why in the heck does my Orc Fighter 10 with +4 Leather and Combat Expertise/Improved Disarm ... use a +2 Greatsword ?&quot; other times, it might be more subtle, something you only see after using it for a while: &quot;Hmmm... I did not know Orcish Fighters liked Dwarven War Axe so much; 10 out of 100 of the ones I made have used it as their primary weapon.&quot;Any randomization method would need some way to determine &quot;these are the choices&quot; and &quot;this is how likely each choice is&quot;. At the simplest, you give each weapon the same chances, at which point a Halfling Wizard is as likely to wind up with a Greatsword as a Half-Orc Barbarian is to wind up with a Gnomish Hooked Hammer. Adding complexity, you can say &quot;this class is x% likely to want that weapon&quot;, to create affinities, so that Barbarians tend to have bigger weapons, Fighters a broader range, and Wizards lighter ones. But now, the developer has weighted what *s/he* thinks the &quot;best&quot; weapons are by class. Then there is the question of how to add new ones. With hard numbers, you would need to recalculate the whole list for every class every time you add a new weapon or weapons. I can go on, but my point is that any randomization scheme that is sufficiently flexible and accommodates the addition of both new classes and new items requires a good deal of design and testing, and would likely require the gear and classe to have some extra descriptive information that helps the program tie them together (Class A is a &quot;High Damage Low Armor Combat&quot; class ... weapon X is &quot;High Desirability&quot; for &quot;High Damage Low Armor Combat&quot; classes while weapon Z if &quot;Not Desirable&quot; for &quot;High Damage Low Armor Combat&quot; classes).</blockquote>

[QUOTE="Silveras, post: 2399184, member: 6271"] Sure you could have some randomization .. but it would be randomization that reflects the developer's personal preferences, at least initially. Sometimes it would be obvious: "Why in the heck does my Orc Fighter 10 with +4 Leather and Combat Expertise/Improved Disarm ... use a +2 Greatsword ?" other times, it might be more subtle, something you only see after using it for a while: "Hmmm... I did not know Orcish Fighters liked Dwarven War Axe so much; 10 out of 100 of the ones I made have used it as their primary weapon." Any randomization method would need some way to determine "these are the choices" and "this is how likely each choice is". At the simplest, you give each weapon the same chances, at which point a Halfling Wizard is as likely to wind up with a Greatsword as a Half-Orc Barbarian is to wind up with a Gnomish Hooked Hammer. Adding complexity, you can say "this class is x% likely to want that weapon", to create affinities, so that Barbarians tend to have bigger weapons, Fighters a broader range, and Wizards lighter ones. But now, the developer has weighted what *s/he* thinks the "best" weapons are by class. Then there is the question of how to add new ones. With hard numbers, you would need to recalculate the whole list for every class every time you add a new weapon or weapons. I can go on, but my point is that any randomization scheme that is sufficiently flexible and accommodates the addition of both new classes and new items requires a good deal of design and testing, and would likely require the gear and classe to have some extra descriptive information that helps the program tie them together (Class A is a "High Damage Low Armor Combat" class ... weapon X is "High Desirability" for "High Damage Low Armor Combat" classes while weapon Z if "Not Desirable" for "High Damage Low Armor Combat" classes). [/QUOTE]

Verification