Test: Extra resources on reloaded map?

[RulerofRails]

I didn't mean that you should run it that much. Just the same as before (12 times). But that's fine, not that important to me either.

[RulerofRails]

The way I see it with the number of cycles is that the smaller tests are still good because they still allow us to identify the stronger potential trends (what we are after). Of course there will be some false readings. All potential trends then need testing on a larger scale to confirm their accuracy, but doing the smaller test first can save time. My

---

I ran the test on Pacific NW, 30 cycles. Renamed vs. Normal exe. Only recorded the number of Resource buildings this time.

I know it's hard to believe but the renamed exe returned 232 every single time on my machine! I did a graph to illustrate the variation experienced with the normal exe. The list of raw values is on the LHS. This info is on the first page of the spreadsheet:

[Gumboots]

This is utterly bizarre, but fascinating nonetheless.

So when you say it returned 232 every time: is this identical seedings, or is it a range of seedings but with the same number of resources?

[RulerofRails]

From what I saw they are different locations and mix of types, but of course I didn't take the time to check every single load. A quick glance through my saved screenshots from the Stats - Industry page that I am pulling the figure from shows that the mix among types is different. There is the chance that there was a repeat or two in there, I didn't check them carefully enough for that. But generally, yes, I believe these are different seedings.

[sbaros]

I cannot swear that I have verified 2 ABSOLUTELY identical seeds, yet all the initial seeds of both draft maps I work on present an overall "déjà vu" atmosphere. I don't mean merely specific industries appearing at specific regions (which would be expected from the features that a mapmaker allocates to every place), but whole sets of industries frequently appearing at the very same points, which should be more improbable than winning the lotto !
It sounds too clumsy and I suspect there has been an intention from the part of the creators to employ some kind of pseudo-randomisation, so that different seeds of the same map would not be chaotically different from each other.
A clue to what happens with the algorithm must be the fact that the supposedly random player character allocations always follow specific combinations. For example, whenever the first rolled character is Otto Bismarck, the other 2 are always George Stephen and chairman Mao !
Some time I hope there will be a thorough discussion of the program and its techniques (preferably in some sort of wiki form, where accumulated knowledge would be systematically organised). Nevertheless I believe that intellectual property rights -to the degree they are tolerated- are confined to the program code itself. Users have the right to know what rules are followed by the simulation model, how origin-destination tables are determined etc.