[spoilers] Pax Commodorica: Pharaoh of Rome

[Ceiliazul]

I think I finally get your sig! D + 2 = f in hex?

Which direction is city the fifth going? If its the desert plant E of legio, name it Cyrenaica?

Yup, D+2=F. My only original contribution to humor in this lifetime, I think.

i remember that epiphany in 2011, good times!

...so the noble axe's name is 'axeman'?

[Commodore]

...so the noble axe's name is 'axeman'?

Vexillatio I 'Victrix'. Name in the second shot.

[Commodore]

All barbarians, repulsed! Now if we could just convince the next customers to stay out...

[Commodore]

Turn 56, part duex. The reason for the delay on the Oracle is seen here; I can't whip to sufficient overflow to 1t the bugger next turn. So instead I'm going to chop into a mine in the west there and work it plus the copper mine for 10hpt; two turns of that plus 80h of chops makes the Oracle a 2t project. I have hopes that will be enough, as the GLH really does look excellent for Dtay.


Gaspar was last to fourth city and first to fifth, this turn. He's doing unsurprisingly well, only unknown seen here on the top cities screen. The rest here is fairly obvious; Pindicator enjoying his Chm happy cap, Ox and Thoth Creative, and TBS enjoying his long-term potential as Joao of Stonehenge(!).

[Commodore]

TENSION TURNS: 1 of 2


Rather than suck my axe out of position for MP, the now-safe Augusta was pacified by a conversion to Hinduism; I guess I'm clearly committed now and everyone can see who got it. I'm kind of bummed I've not gotten any spreads yet, actually. What is the formula again? If this was a neighbor with Hinduism half my cities would be converted by now.


You guys are a quiet crowd. Anyone have any questions for me?

[WilliamLP]

I'm kind of bummed I've not gotten any spreads yet, actually. What is the formula again?

A more autistic answer than you probably want. Take it as far as you trust my ability to read the source code:

Religion can spread to a city if either the owner can connect it to the holy city, or if the holy city's owner can connect to it. (An complete understanding of how connections and plot groups are processes would be a hell of a lot of work. Code nerds can see: CvPlot::updateRiverCrossing for instance and tell me wtf. Some of the code is really mental, and unnecessarily so. Computational geometry can be hard, but not _this_ hard.)

Then the formula:

Code:

iSpread /= std::max(1, (((GC.getDefineINT("RELIGION_SPREAD_DISTANCE_DIVISOR") * plotDistance(getX_INLINE(), getY_INLINE(), pLoopCity->getX_INLINE(), pLoopCity->getY_INLINE())) / GC.getMapINLINE().maxPlotDistance()) - 5));

So:

1. Take 100 for a holy city, 200 if it has a shrine.
2. Take the distance (not the Euclidean distance, the distance in terms of Civ, the same one that determines cultural borders and air mission ranges). Multiply by 100 and divide by the maximum possible distance between any two squares on the map. Then subtract 5. (I have no idea why.) If this final value is less than 1, use 1.
3. Divide the value from 1 by the value from 2. That's your chance per turn, out of 1000, of a spread occurring!

So if the two cities are the maximum possible distance apart, the chances of a spread should be 0.1% or 0.2% for a shrine (not 0).

If the cities are 55% apart relative to the maximum map distance: the chances of a spread should be 0.2% or 0.4%.

If the cities are 10% apart relative to the max distance: the divisor is 5, and the chances should be 2% or 4%.

The max spread occurs at 6% apart (due to rounding, probably anything less than 7%). Then the chances are 10% or 20% per turn.

I've seen the max spread described elsewhere, but have never seen the info for the scaling. The chance to spread is completely flat anywhere between right next to the holy city and 7% of the maximum possible distance on the map. And then it falls off really rapidly, e.g. 7-8% of the distance is only half as likely to spread per turn, 8-9% of the distance is only a third as likely, 15-16% is a tenth as likely.

(The falloff is so sudden and large that it might actually be a possible consideration in dotmaps.)

And, usual disclaimer, I might have misread or misunderstood something that throws the conclusion off.

[Old Harry]

so is the distance is related to actual map dimensions rather than "small", "large" or "huge"? I'm going to have to go figure it in all my games now... in the demogame I got one spread in the whole game. I wasn't happy...

[Commodore]

Yeah I'll echo that, nicely done WLP. I knew it was dimensions but I didn't think it would be that...interesting...a calculation.

And the upshot is I have been skunked just a touch, but each turn is a new day. I know it always is a crapshoot to rely on natural spreads; I've always thought if you found a religion you ought to plan on either monasteries or OR soon after.

[WilliamLP]

so is the distance is related to actual map dimensions rather than "small", "large" or "huge"?

Yeah, and it even varies for the same dimensions depending on Flat / Cylindrical / Torroidal.

Precisely, it's:

Code:

int CvMap::maxPlotDistance()
{
    return std::max(1, plotDistance(0, 0, ((isWrapXINLINE()) ? (getGridWidthINLINE() / 2) : (getGridWidthINLINE() - 1)), ((isWrapYINLINE()) ? (getGridHeightINLINE() / 2) : (getGridHeightINLINE() - 1))));
}

So if a map is cylindrical and 100x50, the max distance is 50 for x, and 49 for y. The Plot distance of that is 74.5, rounded to 74 (one step right, and 49 steps diagonally, each step diagonally is worth 1.5 distance.) So 7% of that is 5.18. So any city within 5 civ-distance should get the full 10% spread.

And then 15% of it is 11, so at distance 11, spread chance is close to 1% only.

[Ceiliazul]

And the upshot is I have been skunked just a touch, but each turn is a new day.

Also, bettwr to be lucky on foreign spreads anyway, especially early. give me 3 free opponent capitals and i'll happily build missionaries for my own cities!