Tech Discussion

[sunrise089]

@Dreylin - The most tangible benefit will be two units in Cape Town rather than one. Having 7 catapults rather than 6 isn't anything to sneeze at, and is raising my confidence about the battle (which isn't a sure thing).

During the golden age though, I'm sure T-Hawk could tell us how many extra gold and hammer's we'll be making. I'm sure it's enough for several additional units, and (if we keep resources online) several more upgrades.

@T-Hawk - If I trigger a golden age this turn does that affect the plans to whip Something Fishy, Pink Dot, or Green Acres? I know it sucks to whip off Golden Age pop, but there is also a large premium on having units sooner rather than later.

@Swiss, the rule expert - Do I need to trigger the GA during the first 24 hours of this turn since the Great Scientist is a unit?

[Krill]

Honestly I think there is an argument for saying that using a unit in anyway counts as moving it (ie starting a golden age, upgrading and promoting), and while I don't know quite how to make it succintly, I do think that it might be against the spirit of the rules. Definately not about to make a judgement though, need to listen to the arguments first.

Be assured that I'm getting a headache over this double move rule as well.

[T-hawk]

Can't give exact Golden Age numbers at the moment without my Civ computer (away till Sunday.) But it's on the order of 30 to 40 each hammers and cash per turn, adding up to something like 35 x 8 = 280 each hammers and cash for the entire Golden Age. That's not all in unit-producing cities, so figure about four extra units worth, plus one more out of Heroic Epic doubling at Cape Town.

Do note that we'll need to save much of that cash for the 100% research run on Nationalism after the Philosophy bulb. I'll figure out exact numbers in a turn or two.

Regarding the "move" rule, I understand the spirit argument too, but the rule does not state that and I think we're only obliged to play by the letter. Consuming is not moving. Adjusting build orders after enemy movement in a turn is allowed (right?) so adjusting Golden Age plans should fall likewise.

[ruff_hi]

Why is it that sometimes binary research doesn't work?

We currently have 459g in the bank
0% tech: +83 gpt, Nationalism in 183
100% tech: -89 gpt, Nationalism in 9

So, to run at 100% tech, we need 801g, or 342g more (a little over 4 turns ... 4.12 turns actually). That means we will have Nationalism in 4+9 ... 13 turns.

However, if we run at 70% tech: -38 gpt, Nationalism in 12 with 3g left over.

Maybe the 4x11 beakers at 0% pushes that 9 down to 8.

[sooooo]

Presumably because if we do it the binary way we end up with more than 3g left over. It can be solved by pushing research to 100% a turn earlier, then on the last turn adjusting the slider to the right amount to finish the tech.

[ruff_hi]

actually, no. binary research needs 4.12 turns to get the required gold. I cut that down to 4 so already it is at negative gold on the final turn.

[ruff_hi]

I put together a spreadsheet looking at this particular issue. Naturally, this all assumes that the next 12 to 14 turns will be exactly the same from a tiles yield point of view - this is totally wrong, but its an assumption that I have to make.

Here is the result:



Col 1 - tech rate while researching
Col 2 - raw beakers generated
Col 3 - gold
Col 4 - Number of turns at 0% tech rate
Col 5 - Beakers Outstanding when you turn tech up
Col 6 - Turns of Tech Required
Col 7 - Gold Left after finishing researching
Col 8 - Estimated Beaker Overflow
Col 9 - Total number of turns (turns @ 0% and turns researching)

I estimated the conversion factor from raw beakers to beakers added to tech to be about 1.22 based on raw beakers generated at the 11 data points and the number of turns required.

So, we have some options ...

• Total 12 turns: tech at 12 turns at 70
• Total 12 turns: tech at 3 turns at 0%, then 9 turns at 90%
• Total 12 turns: tech at 4 turns at 0%, then 8 turns at 100% (and trust that we will find the other 11 beakers required somewhere)
• Total 11 turns: tech at 2 turns at 0%, then 9 turns at 90% (and trust that we will find the other 23 gold required somewhere) Note: Required Figures not shown in above table.

I would go with 4 @ 0% and 8 @ 100% because that is what we have been doing and there is less chance of someone making a mistake. Just keep in mind that we don't need (cross fingers) 5 turns at 0% even though the raw numbers say we do.

Side Issue
It is interesting to note that for each gold piece you spend turning up your tech rate, your best conversion to beakers occurs at 90% tech rate (but not by much).

Tech Rate: Additional Raw Beakers: Cost in Gold: Ratio
10%	24	18	1.33333
20%	49	35	1.40000
30%	73	52	1.40385
40%	97	69	1.40580
50%	122	86	1.41860
60%	146	104	1.40385
70%	171	121	1.41322
80%	195	138	1.41304
90%	220	155	1.41935
100%	244	172	1.41860

So, dialing tech up to 10% costs you 18 gold, nets you 24 beakers for a ratio of 1.333.

[sooooo]

That's ... a bit weird. I have to say that I don't understand the reason why the binary research is working out to be less efficient than the normal research. Where are the beakers being lost to?

I would go with 4 @ 0% and 8 @ 100% because that is what we have been doing and there is less chance of someone making a mistake.

This is flawed thinking - if you have calculated that it's better to go with non-binary research, then let's go with it. Have confidence in your maths!

[ruff_hi]

That's ... a bit weird. I have to say that I don't understand the reason why the binary research is working out to be less efficient than the normal research. Where are the beakers being lost to?

I've said this from day 1 (even under Vanilla) ... there are lots of situations where 80% tech is much better than 100%.

For example: City generates 10 commerce with a library ...

at 100%, you get 12 beakers and no gold
at 80%, you get 10 beakers and 2 gold

The issue is that you lose beakers to the library rounding down. Sure, this is one example, but once tech multipliers (and gold multiplers too) come into play, binary might not be the best result.

Also, in this situation, 100% tech rate and the 11 beakers we generate at 0% are just enough to drop the turns down by 1. If we generated 15 beakers at 0%, I am pretty sure that 100% would be the way to go. The current situation is at a real tipping point.

This is flawed thinking - if you have calculated that it's better to go with non-binary research, then let's go with it. Have confidence in your maths!

No it isn't - I trust my maths (btw, did anyone check it ). I am just building in allowances for a) my initial assumption not reflecting actual experience (highly likely) and b) human nature getting in the way (ie someone logging in and saying ... why are we at 90% tech, that isn't binary ... and then bumping it to 100%.

[ruff_hi]

oh - 1 more thing - 1 small reason to keep with the binary research ...

the later we start teching Nationalism (and assuming PAL get there quickly), the more we benefit from the 'someone else has it' bonus. More tech beakers for your raw beaker bang ... so to speak.

[darrelljs]

Actually, with BTS the rounding is much less of an issue. The calculations are still fixed point for performance reasons, but the radix was shifted so that there is a fractional component.

Darrell

[sooooo]

the later we start teching Nationalism (and assuming PAL get there quickly), the more we benefit from the 'someone else has it' bonus. More tech beakers for your raw beaker bang ... so to speak.

Good point! That's a very good argument for keeping to binary so let's do that.

[ruff_hi]

Update on the tech rate ...

It now seems that running 0% / 100% is the best. At 100% tech, it is costing us 179 gpt and yields 253 raw beakers - 1.413 beakers per gold. 90% is 2nd best at 1.407 ... an obvious win by 100% tech rate.

Run 0% for 2 turns, then turn tech rate up to 100% ... numbers:

current bank: 542
bank in 2 turns: 714
cost at 100%: -93
Nationalism in 8 turns at 100%
Bank once nationalism finished ... -30

Assuming we find 30 gold from somewhere, we are good.

[sooooo]

I think we should go to 100% research next turn. We may be able to get another 100 gold from somewhere - work out a deal with rabbits or borrow gold for a turn from another team.

[T-hawk]

I should just put "agree with sooooo" on hotkey.