Yeonne Greene, on 12 July 2015 - 03:16 AM, said:
Because it's not my job -- unless of course PGI wants to pay me my going rate. But, having said that, I do say how things can be done. Go look for them if you're genuinely interested.
Yeonne Greene, on 12 July 2015 - 03:16 AM, said:
You like to poke and prod at people discussing the subject,
That is what I do. It's to trigger more important brain functions from people. It works in RL well too.
Yeonne Greene, on 12 July 2015 - 03:16 AM, said:
calling all attempts at observation and quantification vain or void
Not vain, just not as easy as many people here seem to always imply.
For example, we see people say things like "It's as simple as just editing an XML file!". Now that's what I call funny.
Yeonne Greene, on 12 July 2015 - 03:16 AM, said:
or "not BattleTech," but you never have anything else to contribute to such conversations. If you care so much about it, do something. That's what the rest of us do; we see a way to break something down within our capabilities and capacities and we do it. You? You sit on that pedestal and frequently talk about too many variables to count as if that actually means anything. Pragmatically, it doesn't mean anything at all.
See my first comment.
Yeonne Greene, on 12 July 2015 - 03:16 AM, said:
You think game developers create such equations? They don't. They make their game, they eyeball the numbers with only maybe a roughly quantified framework of trade-offs, and then they run play-tests to make adjustments. And you know what? It works in most cases. Doing what you suggest is a tremendous waste of resources; the expertise required to build those algorithms is expensive and the finer details in the resulting product would be unappreciated - if not completely unnoticed - by the vast majority of the customer base. The "good enough" theorem is in full effect and I'd also say the "earnest garbage" theorem is, too.
Maybe they should. Many people in other fields do it all the time. It's called quantitative modelling. Some even build computer models based those same equations.
Considering the wild balancing swings we have, it is not so bad an idea.
Yeonne Greene, on 12 July 2015 - 03:16 AM, said:
Don't take this as a hostile attack. I am genuinely perplexed at your...enigmatic existence on these boards. Your comments are almost always condescending and, from where I sit, there's no justification for being so.
TL;DR: If you want to participate, freaking participate, dude.
If that is what you think, then you probably have not been paying close attention
enough.
And by the way, my comment was actually meant as a subtle joke to the "working in academia" comment I was responding to. But then again my so-called jokes (or pokes as you call them) are usually double-edged.
Edited by Mystere, 12 July 2015 - 09:52 AM.