Adventure 35 - Results

[T-hawk]

Diplomatic victory:
Selrahc 1665 AD

Spaceship victories:

timmy827 - 1655
zeka - 1690
Olodune - 1710
T-hawk - 1720 (sponsor shadow)
Swiss Pauli - 1725
haphazard1 - 1750
shpoko - 1750
LKendter - 1765
Ruff_Hi - 1785
Ronald - 1840
Kodii - 1860

And as promised, here's a graph showing the research rate over each turn, for each player that reported the data. Also the Excel workbook. Beaker counts are the raw XML values, unadjusted for game size and speed.

Ruff_Hi took an early lead with the Civil Service slingshot, but that didn't hold up through the rest of the game, possibly thanks to fairly late Democracy. Timmy827's early game was middle of the pack, but he took a clear and permanent lead with the Democracy slingshot.

My max-density build worked out probably about as one would expect. I lagged behind the pack through the first third of the game (and no Oracle hurt too), but around T120 after Sid's Sushi, my tech rate clearly improves on par with Timmy's and ahead of everyone else. And my steals show up as small but visible ticks.

Ronald's line shows some surprising activity. He was right with Timmy and even ahead for a short while, but around T120, his tech pace faded significantly. Ronald's report gives little detail, but I suspect the culprit was undersized cities. His screenshot shows cities of sizes 17-16-15-15-14 and it seems Sid's Sushi only in Paris.

If anyone wants to report their tech dates, I'll add them in. Big thanks to Ronald for compiling that data!

[Swiss Pauli]

Thanks again for the game

[Olodune]

I sure wish the in-game graphs used log scales so that the early game didn't disappear by the time the later eras roll around. Neat graph

[timmy827]

Geez, I feel kind of bad now - I had mentioned that I would convert the dates into turns if others did so, and had intended to when I had some spare time this weekend. Thanks to T-hawk and Ronald for assembling.

The option of log charts, or making the graphs zoomable, would be nice.

[Ronald]

Ronald's line shows some surprising activity. He was right with Timmy and even ahead for a short while, but around T120, his tech pace faded significantly. Ronald's report gives little detail, but I suspect the culprit was undersized cities. His screenshot shows cities of sizes 17-16-15-15-14 and it seems Sid's Sushi only in Paris.

I completely ignored corporations. So far I have not played many games in BTS, so I did not know the power of them well enough. Definitely something I will remember for future games.
Also my tech path at the final phase was pretty poor.
Thanks again for a fun game to play

[Ruff_Hi]

Nice work putting the results together T-Hawk.

The option of log charts, or making the graphs zoomable, would be nice.

Log scale isn't available but you can get the graphs to zoom a little ...

• change the time frame from 'all time' to '50 turns' (say)
  
• use the little arrows at the end of the time axis to move back and forth in time

[Ruff_Hi]

The option of log charts

The SVN version of BUG now includes a 'log scale' graph option. Thanks for the suggestion.

[T-hawk]

Nice, Ruff, now I want to peek at some in-game log graphs.

Although I'm not sure that log scale is precisely what we want either. The problem of the vanishing early game is solved, but log scale also compresses the upper end. Timmy and I both got from 100k to 200k beakers considerably faster than the other participants. But in the graph, not much detail there is visible, where the research rate flattens out and is no longer exponential. But the magnitude of the linear numbers still represents a serious factor in speed to the goal.

So is there a such thing in math as a scaling factor that's logarithmic at first then regresses to linear? And can Ruff implement that in BUG by dinnertime?

[Ruff_Hi]

So is there a such thing in math as a scaling factor that's logarithmic at first then regresses to linear? And can Ruff implement that in BUG by dinnertime?

Yes and No. I have installed Log10 (log base 10). If I add a slider to change the '10' then you can change the log scale. For example, if a number looks twice as big as another (log base 10), then the number is actually 10 times bigger. Thus if you use a different base (say 3), and a number looks twice as big, it is really 3 times as big.

That should work.