How fast to expand?

[damerell]

I am now in the Modern era in my game and after building a few Banks etc. I did switch to the 0 research occasionally when I wanted plenty of gold to promote my military units. I found that the time saved in not building new units could be used to build further improvements in my cities. Even at 0 I was still producing some research from the contributions of Specialists. So I think I have benefited from this technique used appropriately.

Except you haven't used binary research there at all; you've exchanged less research for more gold and then used the gold. Which is not an invalid approach, but has little to do with binary research.

5 city empire 35 raw commerce per turn, 0 gold in bank. At 40% slider breaks even. In this ten turn period which approach steady or binary [4 at 100% 6 at zero]wins and by how much?

Just because you can produce a case without rounding errors does not mean that there are not cases with rounding errors where binary research will make a difference.

 

[Soirana]

Just because you can produce a case without rounding errors does not mean that there are not cases with rounding errors where binary research will make a difference.

That is exactly not the case without rounding.

Let us see:
Steady:
35x0.4 14 raw beakers.
14+1 hidden>> 15 beakers which get modified by 1.2 -- ends being 18 beakers into tech per turn.
18 x10 180 beakers total.

Binary:
6 turns on zero slider caching 1 hidden beaker - 6 beakers
4 turns on 100%
35+1hidden>> 36 beakers 36x1.2 43.2actual beakers per turn which goes down to 43 beakers.

Total 4x43 +6 -- 178 total beakers.

Binary looses by two beakers. Guess with some effort I could found sample there margin is bigger but that is rather theoretic discussion anyway.

Well, I would gladly take back my words about liars and number manipulators as that is rather typical overstatement of mine, but however...
I would prefer mentioned people take back statements like

"- Binary research is always better in the early game"

Well, and the point about number manipulation proving anything you want in complicated case still stands.

 

[Pisskop]

Without having read all of what what others do, I follow this:

I play usually on marathon (After all, I play this game for the enjoyment, and rarely try to be a brutal and efficient as possible. Half the fun I running a Nationalistic monarchy with serfdom well into the 20th century), so turns reflect that.

The first hundred (100) turns are scouting. A second city, or more rarely a wonder is produced. The next fifty (50) are usually for units, and a barracks is built in #2 while resources are produces are hooked up. A worker* is chopped (usually, again), and a rush is performed if I feel I'm up to it.

(Worst thing ever is to fail a rush. And if its that close, I may not even try but sue for peace until cats).

So, I have 3 to 5 cities by 150-200. By then I'm rushing for currency, which the ai is slow to get even with k-mod. I trade it for select things, but mostly horde it and cottege (only after getting food. Specialists can run off food alone, so I set thta up first).

Then I usually run 50%-30% beaker rates. I expand until I just make enough money to stay in the green. I run higher beakers (and red net profits) on 'important techs' and focus making that one (or two, in large games) friend.

So, by currency, 4-6 cities.
by Astro, it is limited mostly by my economy, war success (I am rarely at peace, usually for 20 turns an era or so to bang whip out infrastructure).

---
And having read I see my input is unneeded. . .

 

[Shaandore]

That is exactly not the case without rounding.

Let us see:
Steady:
35x0.4 14 raw beakers.
14+1 hidden>> 15 beakers which get modified by 1.2 -- ends being 18 beakers into tech per turn.
18 x10 180 beakers total.

Binary:
6 turns on zero slider caching 1 hidden beaker - 6 beakers
4 turns on 100%
35+1hidden>> 36 beakers 36x1.2 43.2actual beakers per turn which goes down to 43 beakers.

Total 4x43 +6 -- 178 total beakers.

Binary looses by two beakers. Guess with some effort I could found sample there margin is bigger but that is rather theoretic discussion anyway.

Well, I would gladly take back my words about liars and number manipulators as that is rather typical overstatement of mine, but however...
I would prefer mentioned people take back statements like

"- Binary research is always better in the early game"

Well, and the point about number manipulation proving anything you want in complicated case still stands.

Now I understand the point you are trying to make. Yes, the word 'always' is almost always an overstatement . Rounding mechanics - as in your example - can make binary better even in the early game. So I take back that binary is always better in the early game. Your example clearly shows that.

We could have jumped to this part in the discussion and we would have saved me the part where I feel offended and have a bad mood for a while. Sigh...

Looks like I'm a very poor manipulator since I have included a paragraph about rounding errors and their magnitude in post #49 and in almost every post after that. I even said that "it can be quite serious in the early game". And even if I didn't include it, it doesn't mean that I was trying to manipulate. It could be because I didn't know this particular game mechanics, I forgot it, or I simply didn't think it was important enough to mention, or I thought that it would make the calculations less comprehensive / more tedious to do.

When I see somebody acting offensively, I prefer to think about a better cause that would make them act the same way, and only after that consider the possibility that they are being intentionally evil.

@Bellringer: Sorry for the thread hijack.