Conceptual Balance Series (Mod)

[Ivan Pedroso]

UPDATE: Look below for a newer Version of the mentioned Excel-sheet.

- - - - - - - - - - - - - - - - - - - - - - - - - - -

Well back again with some dry math. This time I have used the right 0.3% value for the growth effect (and 6% for order, and 4% for growth). I have tried to take into account the reduction in bad events when taking order. So Order+2 vs. Growth+2, fueled by taking Luck-2, can be compared.

I've also used different values for the gold and growth boosts from taking growth and order in the different calculations. And the "reduction in bad events" bonus from taking order has also been fiddled with.

My conclusion is that "Growth+2, Luck-2" vs. "Order+2, Luck-2" accumulates the same amount of gold after about 50-80 turns, depending on how much of an influence you feel that order has on bad events.

All in all I get the feeling that the values of Order(6%) and Growth(4%, 0.3%) are about right. Maybe Growth(4%, 0.25%) would be better if it is a legal value.

I've attached the Excel-document used to make these comparisons (with comments). It should be usable as a testing template, when horsing around with different values of the scale parameters

[atul]

Very nice sheet. I especially liked the part where you have the huge formula and just note "we assume the situation's mostly identical, therefore". Sure sign of a pro.

Might it be beneficial for the sheet to have some graphs if it were to be used as a testing template? I played with some plots but didn't come up with anything I'd consider too useful (like, a scatter plot of accumulated incomes of order/growth with different values or something...). Might not be a worth the trouble, especially if you intend to use the sheet just to bash your point in.

[Ivan Pedroso]

atul said:
Very nice sheet. I especially liked the part where you have the huge formula and just note "we assume the situation's mostly identical, therefore". Sure sign of a pro.

Hehehe - I take that as a compliment

Might it be beneficial for the sheet to have some graphs if it were to be used as a testing template? I played with some plots but didn't come up with anything I'd consider too useful (like, a scatter plot of accumulated incomes of order/growth with different values or something...).

I tried to come up with some graphs myself, but they weren't too informative, but I'll toy around with it, now that I know that someone has looked it

Might not be a worth the trouble, especially if you intend to use the sheet just to bash your point in.

Well it was never my intention to hammer any point through, I just got curious when Turin mentioned Growth being better than Order when considering income - that seemed a bit odd. And then I got thinking on: "How would I include the Order-event-frequency-reduction-effect in a simple semi-correct manner". It seems that growth is better than order even with this inclusion, but only in longish games, and in a way that seems balanced. Either you want more money early on and with a little more security (lesser risk of bad events that could cripple you economy early on) - then it's Order you want. Or you want more money in the long run, and have the time to wait - then it's Growth.

[atul]

Ivan Pedroso said:

I tried to come up with some graphs myself, but they weren't too informative, but I'll toy around with it, now that I know that someone has looked it

I probably won't need to tell you the benefits of visualizing when viewing that much data. But, whether you can find anything with any informational value in it, I don't know. Good luck (and, incidentally, Christmas).

Ivan Pedroso said: I just got curious when Turin mentioned Growth being better than Order when considering income - that seemed a bit odd. And then I got thinking on: "How would I include the Order-event-frequency-reduction-effect in a simple semi-correct manner". It seems that growth is better than order even with this inclusion, but only in longish games, and in a way that seems balanced.

I think your solution was quite elegant. After all, the effect of randoms is, well, random. But we're dealing with something that affects multiple provinces over multiple turns, so it'd average out on the long run. I'd say the values you have generated would represent the expected values of income over multiple games, each single realization differing a bit. Or something, English isn't my 1st language so this may be a bit incomprehensible.

Of course the actual value of Order's effect is up to debate. I'd maybe rate it a bit higher than 1/0.9 just because in the early game, an unlucky event has both greater chance to hit your capital (only few provinces) and greater impact on your game (a loss hurts more early). If you start with multiple provinces the beginning isn't that big a difference. But I'm a bit risk-averse.

[Ivan Pedroso]

Great comments Atul. You are quite right about the early-game risk of an event hitting your capitol, and that would hurt ones income badly, and that should raise the parameter somewhat above the "naïve" 1/0.90 value.

I have made some graphs on the sheet - it's all quite a bit more informative now.

It is obvious from the graphs that the value chosen for the "Order-effect-on-bad-events"-parameter plays a more important role than the other (well-known) parameters. This means that any conclusions drawn from this method are sensitive to ones choice of this parameter. My estimate of 1/0.90=1.11... for Order+2 came from the observation that in an Order+2 domain event frequencies are reduced by 10%. If a value of 1.40 is chosen (corresponding to an "effective reduction" in event frequencies of 14.2% pr. Order pick - an extreme value in my opinion, even in light of Atul's analysis), then the accumulated income from Growth+2 overtakes that from Order+2 at around turn 120. That indicates that if you are especially scared of an early income hit due to a bad event hitting your capitol (i.e. your private estimate of the parameter is close to 1.40) then Order is what you want even in a long game (with Zen's choice of scales).

I've attached the new and improved sheet to this post (and removed it from the other one above).

[Zen]

I've had a chance to read the data and I'm fairly satisfied with the results. Good job Ivan! It seems my calculations were a slightly off (I was less extreme on the event reduction than your latest spreadsheet).

As such, I think it would be a fair accessment to reduce the population effect of the scale to .25 (as far as I know this is a possible value). This would still give Order the single strongest gold% scale and still regulating Death as painful as other negative scales.

I will also be modifying the Temperature scale back down to 5%.

Barring that debate, anyone have any more feedback about the scales portion of the series? If it doesn't have any outstanding issues I haven't had a good hard look at I will move on in my work with the other aspects of the series.

[Zooko]

Is there a README for the latest balance mods yet?

I'm going to launch a MP game in a few days and I'd like to use the latest or bestest Zen's Balance Mods. The MP game will be slightly non-conventional, in that it will be intended to be a shorter-running game with more international war all along instead of only at the end:

1. Two teams of two players each.
2. Not too large map -- maybe Inland, or The Desert Eye, or even a smaller one if I could find a smaller, wraparound, good-looking map.

Thanks!

Zooko

http://www.shrapnelcommunity.com/thr...rt=&PHPSESSID=