Thanks a lot Adrien. Is your idea rather mathematical/geometric, or does it run along the lines of thrawn's simulation approach? I don't want to burden you with my busywork
Before you start on the big screen I'll try to get a better and uncompressed picture tonight. The limiting factor should be civ's graphics, not my lazy screenshot taking and imgur.
I had an idea about measuring the distances between horizontal grid lines and see if a functional dependency can be fitted, but for one the sample is awfully small, and I would not know which function would be a reasonable basis (hyperbolic/ exponential decay/ power law...)
I feel like projective geometry should provide an analytic answer, but I really understand very little of it (I have a solid understanding of parallel projections, and it appears that the methods may not be too different, just a few extra steps to account for the foci, so I may try yet)
I misread the problem this morning in the train, and didn't factor in the fact that tile height is smaller when farther, which means the calculation I was doing was ten times smaller than the real number. Yeah the simplest solution would be to do a huge map with a lake tile every 10 tiles or so and then compare it with your image...
I'll try to tweak my first idea and come up with a reasonable estimate, it might be worth a shot.
Anybody interested in the details? Basically I did a central projection drawing, following this guide, but backwards. The map dimensions should be something like 98x70, although I don't really have an idea how precise my solution is. Am I correct that on a toroid that means that MaxPlotDistance = (98+70/2)/2 = 66,5?
Embarassingly, after my very tryhard geometric solution got me 90, I felt that that must be too low? So I tried my hand again at flying camera magic and got a much better picture this time:
This more or less confirms the 98 tiles in E-W direction. Well at least I learned something new, I'm actually not unhappy with having spent the time learning a bit on how to draw (and undraw!) perspective.
The game meanwhile, having been launched >11 days ago, can't quite enter turn 5 yet, sadly. The last missing player played their turn after some 46 hours, but apparently there is an issue that the refs have to figure out yet.
So, for those who have not been feeding on the incredible * that Civforum has been delivering over the last few days:
It turns out that the player constituting the tier 2 from my list, Jesper Portus, has maintained a smurf account which he used not only to read other people's threads, but also to ask for copper connections or overview pictures when his main would invade, encourage or dissuade wars, for mapchecking... at one time he even signed up with both identities for a pitboss and proceeded to axe rush himself - and then had the nerve to publish a supposed PM from his smurf explaining how he was just a few hammers short of getting his own axe, and then berated said smurf for violating no diplo rules .
For perspective, as far as I know he is the player with the most MP wins on the forum (however not the perfect streak of Plemo's). He has been a forum member since 2003; the smurf dates back to 2005.
The sad part of course are all the people who used to trust and look up to him, and they also had to cancel the two running PBs where he was participating and got caught now, for posting in another player's thread with his main account it is assumed.
Thankfully this happened when this game was still at turn 4 and he could just be replaced, and more thankfully superdeath hadn't signed up and was available. So we're back at it.
* I understand completely that for everybody involved this is incredibly frustrating and disappointing, but you can't deny it's has a lot of story value as well.
Scouting has completed, revealing even more seafoodlessness:
I also set up an updated sandbox with the hopefully correct map dimensions andall tiles we've revealed, including some fog gazing. It is attached here and will also get attached to post #2.
Tbh, I still have not felt a tickle to sim out this start. Maybe that's the reason why instead I went to learn reverse perspective in order to get the map dimensions. That means that we don't have a concrete plan yet and next turn, having finished Hunting, will automatically follow up with Mining-BW. It is very probably the correct call, considering that it brings the number of useful tiles on the island from 2 up to at least 3 and likely 4 with the copper mine.
We should think of something by the time we finish the worker and reach size 2 though (so basically... a month?). I am increasingly skeptical of a second city on the island, at least before the terrace in the capital. Before that I can see the cap sensibly doing 3->5->3 whip cycles on a food surplus of 7, without much room for sharing food with a second city. Only once we have the terrace sharing becomes efficient. That might mean that the first city goes overseas.
Rusten, I know that you argued against this idea, could you elaborate considering what's available in terms of food?
In the spreadsheet I arrived at:
- Sailing after BW eot 39, galley instabuilt t40 on whip OF and chop
- workboat ready
- 2nd city in south founded t45 with instant fish
- settler for 3rd city by northern fish whipped eot 53 -> found t56
- Pottery around t52 (imprecise)
- 50h into Stonehenge
- 2 Quechuas in the southern settlement, 1 axe in the northern one
I like to found by the southern fish first because there appears to be a hill 1S of it. We can then found for the northern fish on the hill after the cap pops borders.
For additional security we could build Quechuas early instead of dumping hammers into SH, but then we'd start to pay unit maintenance from t38 on with the workboat (and then in quick succession settler and galley). Probably worth to build 1 additional Quechua.
I have to run it through the sandbox yet, and the research numbers, particularly Pottery, are precarious because we definitely start paying unit maintenance from t39 on, and I have not factored in the effects of the second city, (IC)TR, or maintenance in general yet.
Another known unknown is the copper location, if we have any in the BFC . If we do that could speed up the second settler, if not then at least the Quechuas have C1? Damn SIP would have solved that at least, but could we know at the time?
Barbs will come by the dozen; in the stats of the games that just got canceled you see players having killed on the order of 50 warriors. Here it will hardly be better, assuming everybody is on islands and they are free to roam on the mainland. I am quite worried about losing a city to early metal barbs appearing. One of the reasons why I'm keen to found on hills. I guess at least with the fish the overseas cities are less prone to getting their improvements pillaged in the early phases.
* that Civforum has been delivering over the last few days: 
. 
)TR, or maintenance in general yet. 
. If we do that could speed up the second settler, if not then at least the Quechuas have C1? Damn SIP would have solved that at least, but could we know at the time?