| ReplyMay 14 2003 at 7:21 PM | Sirp (no login) from IP address 211.28.96.38 |
Response to Expectations |
| While I think that T-Hawk has quibbled unnecessarily over irrelevant trivia in the past, I don't think he's doing that here, Sirian.
This is an important issue. You have pointed out several problems with the game, most of which T-Hawk and I have agreed with you on (and some on which we have offered ideas for solutions), however on this one, we disagree. I don't see why we shouldn't say so, and I don't see why we shouldn't investigate to find out the truth of the matter.
You talk as if presenting objective evidence for or against your case is difficult or impossible. It's not. If there is some kind of streakiness in the PRNG that civ uses, it should be fairly easy to set up a test and show a results distribution that is different to what one would expect from random numbers. T-Hawk and I have both done this, and have found that the results distribution lines up with what one would expect from random numbers. I have also pointed you to an article by someone who has reverse engineered the way the civ PRNG works, and found that it closely mimics random numbers.
Writing a PRNG that closely enough resembles true random numbers to provide military-grade encryption is hard. But writing a PRNG that is indistinguishable to any human from true random numbers is not hard. Even imperfect PRNGs will be far more likely to produce negative correlation with previous results, rather than positive correlation, which is what you seem to think happens.
Surely finding the truth of the matter is best for the community. If it is as T-Hawk and I say, then that is best for everyone, not least of all you, Sirian. If you are right, then you seem to have contacts in Firaxis. Why not just send them an email showing evidence of a bug in their PRNG giving correlation with previous results, and ask them to fix it? As I said, writing a decent PRNG is not hard, so surely they would treat it as a bug, and I would anticipate they would make fixing it a high priority. The combat system is frequently critisized, and if someone were to show a genuine bug in it, I'm sure they would fix it quickly.
I'm not even sure why you seem to be getting agitated about it, to be honest, Sirian. It's a fairly easy thing to get credible evidence for. So, why not just generate the evidence, either decide it's not a problem, or else see if we can get it fixed, and then move on? There's no reason to get annoyed or upset over it.
-Sirp. | |
| Responses- Couldn't you test it... - Jester on Jun 2, 12:33 AM
- P.S. - Jester on Jun 2, 12:35 AM
|
|
|