To avoid clogging up @Truthiness 's data collection thread, here's a separate thread...

Based on an inquiry by @Rocmistro , I decided to look at how many points are generally spent on upgrades, and how those groups fare in competition:

So, the brackets were chosen to be roughly evenly sized. Carriers are 80+ points in squadrons & gunboat fleets have less than that.

results:

Not that informative. Very few fleets use less than 40 upgrade points or more than 95, with most between 52 & 77. But no particular tilt towards christmas trees or nudists in the winning brackets.

Carriers tend to use fewer upgrades than gunships, but the people who brought upgrade heavy carriers or upgrade light gunboats did about as expected.

I suspect that you could find some questions that might give clear results, like "do ackbar or vader fleets use few upgrades successfully?" but nothing that I feel that I could crank out in a hurry. another interesting question would be upgrade points spent on the ship with the most upgrades. (one big hitter vs two threats, etc.)

Are the columns headers supposed to contain information about performance? If not, I don't get it.

27 minutes ago, Andylicious said:

Are the columns headers supposed to contain information about performance? If not, I don't get it.

Looks like the data is from Truthiness' Regionals data spreadsheet, which would imply these headers, categorizing the fleet types by the frequency with which they showed up in the respective buckets in Regionals:

I'm a bit confused by the bracket ranges. They are 6, 9, and 11 across respectively, aside from the tails. I see that they were chosen to be evenly sized in the ALLxALL, but I don't think that makes epistemological sense. Could you explain your rationale for doing so?

4 hours ago, Mikael Hasselstein said:

I'm a bit confused by the bracket ranges. They are 6, 9, and 11 across respectively, aside from the tails. I see that they were chosen to be evenly sized in the ALLxALL, but I don't think that makes epistemological sense. Could you explain your rationale for doing so?

By picking evenly sized starting groups, it would be more obvious to the casual reader if high upgrades was a more successful strategy than average or low.

It was a pretty standard technique in my stats classes to divide populations into quartiles or quintiles.

I used the same brackets for guns vs carriers to make it easy to compare the two.

reviewing the formulas, the brackets are all off by one. I used "less than" because "less than or equal to" is harder in excel, but didn't adjust when I wrote the brackets. Oops.

So that's what I've been doing wrong.

I just needed to add more upgrades!

I think a more accurate way to look at this data would be to say it doesn't matter so much how many upgrade points you bring so long as you can use them effectively.