CN Sanctioned Stats Depot, Special Edition: The Twisted World of Unspeakable EvilThis thread is all about the statistics I produce, and what they mean. I usually know what I mean, but sometimes it isn't obvious to others, so here's the thread. I welcome discussion here of what any of you readers see as indicators from the data (I put my own thoughts about some things below). If you've never seen
my statistics thread, I'm not sure why you are here (although I suppose it's as good a place as any to start).
A word on Tiers:There are only so many hours in the day and there are only so many alliances you can plop into a graph before it becomes difficult to read. Given those constraints, I only produce graphs and Day-to-Day charts for the top 30 alliances. These 30 are divided into two tiers of 15 each. Tier 1 contains the currently sanctioned alliances and the next three biggest by nation strength. On Tier 1, I have a strong bias towards alliances with 290+ members (meaning they meet the membership requirements for being sanctioned), but it isn't set in stone. That's why TOP is there now, they're abnormally large for their membership count. Tier 2 are the next 15 biggest by nation strength.
As of the time of this writing, I update six graphs daily, all for the Tier 1 alliances. Tier 2's Day-to-Day charts are updated daily, also. Everything gets updated on Sundays, for both tiers. If I am going to move alliances around between tiers, it is going to happen on a Sunday, as it can take a fair amount of time to do this step.
The other alliances I track show up in the rankings. They're generally at least 1M NS, but as of now it's 'stronger than the weakest millionaire in the stats'.
A word on nomenclature:All of my normal daily stats are generally produced with a short-hand suffix. This is because I'm lazy. You probably don't care, but if you ever see me type something like 'ANS', 'D2D' or 'TNSPN', know that that's one of the short versions for a graph or chart. I'll document what the short version is as I get to each kind.
The Charts:Charts don't have pretty pictures, but they do have pretty numbers. I try to have some meaningful color-coding to give extra context to the data, because I'm all about efficiency like that.
There is a color code guide for the charts (shown below), and what it means will be explained in further detail below.
(
Posted Image)
Rank:Short form: R
(Posted Image)
The Rank chart is a list of all alliances in the database for which data exists for today and for 7 days prior. Alliances are listed by descending nation strength. Totals for select columns are at the bottom of the chart.
The first column shows the alliance name.
The next two columns are the Rank and Strength of each alliance as of the day the chart was produced. In the above example, you can see that above these columns it says '2007-07-28', which is the date the chart was produced.
The next two columns are the Rank and Strength of each alliance as of one week prior. The date is shown at the top of these two columns to help out those who have always had trouble with subtracting dates.
The last three columns show the change in Rank and Strength.
The change in rank is the difference in rank from last week to the current week. In the example, look at MCXA: the prior week they were the 13th largest alliance by nation strength, and that day they were the 11th; meaning they moved up 2 ranks. If the rank didn't change, the background stays light green. If an alliance moved up the rankings then the change in rank cell is colored green, and if they moved down it is colored red.
The change in Strength is shown as both an absolute nation strength number and as a percentage of the alliance's last week strength. Looking at MCXA again, in the example they were 219,853 NS higher than they were the week before. That is 6.83% growth for the week:
(219,853 NS / 3,218,702 NS) x 100 = 6.86%
The percent growth of all alliances in the chart is at the bottom, in this case it is 3.85% (4.9M/127.5M ~ 3.85%; the spreadsheet carries it out to greater accuracy than I typed out there). If an alliance's percent growth exceeds the overall percent growth, it's change in strength and percent columns are colored green; if they are below average then they are red.
The reason I show both absolute growth and percent growth (and why coloring is based on percent NS growth and not average NS growth) is that whether or not X NS growth is healthy or not depends on how big an alliance is to start with, and how quickly other alliances are growing. 219k NS is 6.86% growth for MCXA, but it would only be 2.33% for IRON. If both alliances could sustain those levels of growth, it'd only be 26 weeks until MCXA surpassed IRON. Naturally, the trick is sustaining that kind of growth, which isn't easy.
In general, smaller alliances can grow at high rates of percent growth (5-15%) for weeks at a time. The stronger alliances tend to have slightly lower growth rates (which tend to equate to a much larger absolute NS change).
Rank by Score:Short form: RS
(Posted Image)
The Rank by Score chart is a list of all alliances in the database for which data exists for today and for 7 days prior. Alliances are listed by descending alliance score. Totals for select columns are at the bottom of the chart.
The first column shows the alliance name.
The next three columns show the rank, score and membership count of each alliance for the date the chart was produced (also listed above the three columns at the top of the chart).
The next two columns show the rank and score for each alliance from one week prior.
These first six columns are color coded to indicate their potential for sanction. Green indicates that the alliance has both 300+ members and a score within the top 12 of those alliances with 300+ members. GPA, NPO, IRON are in this category in the example. 99 times out of 100, these alliances would get sanctioned if an audit happened that day.
Alliances with 250+ members and a score outside the top 12 for alliances with 300+ members are in light-green with italic text. In the example, you can see CSN, MHA and others in this class. I use it to indicate 'near sanctionability', meaning they've gotten to within spitting distance of the membership requirement.
The ugly lightish-blue color (exhibited by TOP, Gramlins and FCC in the example) indicates an alliance with insufficient membership to seriously consider getting sanctioned in the near term.
Green with red, underlined and italic text is a special case. That's where an alliance has 290-299 members and a score that
would be in the top 12 if they had 300+ members. Admin has stated that there is a fudge factor on the membership requirement, for some reason I got it into my head that it is 10 members. So if Admin draws from such alliances then they will get sanction and the last alliance highlighted in green will not. In the example, you can see MCXA is in this limbo-zone. If an audit happened on the June 26 (the day of that chart), and Admin decided MCXA should get a sanction because they were only 1 member off, then Legion would not get sanction. This situation is rare, and now that the score calculations have been adjusted, I imagine it'll be rarer still in the future.
As with the normal Rank chart, the last three columns are for change in rank and score. The color coding and calculations are exactly like those in the Rank chart, but with Score instead of NS.
Day-to-Day:Short form: D2D
(Posted Image)
(Posted Image)
The Day-to-Day chart is a look at one of the tier's constituent alliances. I have two examples here, the first is Tier 1 and the second is Tier 2. Each column is one alliance and each row is a different metric. All of the numbers are calculated from yesterday to today. In the example, it would be the change from the 27th to the 28th (the date is in the top row of the chart).
Percent Growth is the change in alliance NS from yesterday to today. Note that the average in the last column is not the average of the averages (that's not meaningful), it is the total NS growth of all 15 alliances divided the total NS of all 15 alliances from yesterday.
Delta NS is the absolute change in NS from yesterday to today.
Alliance Nation Strength Gain per Nation (ANSGPN) is the change in alliance NS (from yesterday to today) divided by the number of nations in the alliance today. This is a measure of how well the average nation grew yesterday.
Delta Average Alliance Nation Strength (Delta AANS) is the change in the alliance's average NS from yesterday to today.
Delta members, nukes, and score are the change in membership, nukes and score from the day before.
The color coding is: green is above average (from the last column on each row), red is below, and light-green is equal. You won't see the light green very often because (although you don't see it on the chart because it rounds) the averages and values are calculated to many decimal places and the odds of hitting an exact number are bad. You'll see it on members and nukes from time to time (when the average change is 0 or 1, for instance).
Delta AANS and ANSGPN are similar but they are *not* the same (although they can be; confused yet?). In the example, NPO and TOP have the same value for each because they didn't gain or lose any members. This also tells you that they almost certainly did not gain a member and lose a member. The reason for ANSGPN is to help distinguish what changes in membership have on the strength of an alliances members.
For example, MCXA's AANS declined 33 NS. However, the average nation grew 123 NS. Together this helps you understand the alliance's health more than one or the other alone. This can seem counterintuitive, so I'll explain further.
Imagine an alliance with 100 members of 20k NS each. The alliance's total NS is 2M NS, and their average NS is (obviously) 20k. They're happy exactly the way they are and don't do anything to increase their NS, day to day. They even sell the tiny bit of land they grew so that every member stays at exactly 20k NS. However, 10 new members join the alliance one day. They all get aid as they join from the older members, and they all end day 1 at 5k NS.
Now the alliance has 110 members at 2,050,000 NS (+50k NS) total. Their average NS is now only 18,636 (-1,364 NS from the day before). However, the average nation grew 50,000 NS / 110 nations = 455 NS.
So when you see a rising ANSGPN and a falling AANS, it usually means that nations are joining the alliance who are weaker than the average NS. If ANSGPN is falling but AANS is rising, that usually means nations below the average NS are leaving the alliance. That's probably what happened to FAN on the Tier 2 example above. It is possible that both are falling at the same time. That usually means a lot of big nations left or there is a major war going on; it's something that indicates bad things are afoot if it happens for several days in a row. If you see it happen to GOONS, it usually means it's time for another joke alliance or we're at war. :jihad:
The Graphs:Graphs are pretty pictures. There are three sets I maintain: Tier 1, Tier 2 and Team. Tier 1 and Tier 2 are the same graph, just with a different set of alliances. Team graphs are similar but are based on the Team statistics.
Only select Tier 1 graphs are updated daily, everything else is updated on Sundays. Right now, those graphs are Alliance Nation Strength, Average Alliance Nation Strength, Percent Change in Alliance Nation Strength, Alliance Membership Count and Score.
Where possible, graphs will be over a window of the 10 weeks; each dark vertical bar is 7 days from the last. Occasionally the period will be different because of insufficient data (this usually happens on the Percent Change in Alliance Nation Strength graph when new alliances appear or Alliance Affiliation changes), although sometimes a one-off graph is made over a different period of time (such as the Alliance Projected Growth graph).
Alliance Nation Strength:Short form: ANS
(Posted Image)
This is a graph of alliance NS against date, nothing special is done here.
Alliance Average Nation Strength:Short form: AANS
(Posted Image)
This is a graph of Alliance Average Nation Strength versus date. The figure is not actually from the CN stats page, but derived directly from Alliance Nation Strength and Total Nations (the values are the same, but that way I don't have to save AANS values in the database).
Percent Change in Alliance Nation Strength:Short form: PANS
(Posted Image)
This is a graph showing the percent change in the Nation Strength of an alliance versus date. Looking at the X-axis, you'll see that at some point it crosses 0% (usually at the Y-axis). As it moves forward, you see the percent change in NS versus that 0-day. In the example, you can see that NADC grew a bit more than 40% in the first 4 weeks of the measurement period, and a bit less than 95% over the full 10-week period.
The PANS chart is useful for seeing how well an alliance is sustaining their growth. A few days of supurb growth may look great, but in the full context of the 10-week period might not make any real difference in their long term growth. You'll tend to notice the smaller alliances can grow faster; although that isn't a hard rule. Sparta and GOLD are within 2k NS of each other on July 29th, you can see that over the preceeding 10 weeks Sparta grew 130% while GOLD couldn't quite eke out 20% growth. If the pace is continued, you'd expect Sparta to be
much larger than GOLD 10-weeks from the 29th.
You can also see the affect of war by looking at the alliances involved in the FAN conflict (GOONS, IRON and NPO from around June 17th). You can really tell who was targeted the hardest (well, seeing as how FAN isn't on this chart; they're at -71.23% for the same period).
GOLD will be my poster child in this example for what an alliance looks like who is experiencing some other malady (since as far as I know they weren't in any war for the period). You see flat-to-negative growth followed by a return to steady improvement.
If you couldn't tell, I really like the PANS chart, it packs a lot of information into a relatively simple area.
Alliance Membership Count:Short form: AMC
(Posted Image)
This is a graph of alliance membership versus date. It comes straight from the 'Total Nations' column on the CN statistics screens.
In the example I'll point out two excellent examples of alliance membership drives; NADC from about May 27 to about June 18 and Sparta from about July 10 to the last day recorded. When you see something like that, you know they're doing something to get new members. Combined with a falling AANS, you know they're recruiting nations weaker than the average; in this case a crosscheck of Sparta's AMC versus its AANS shows the AANS falls through the floor as their membership begins to increase.
You also see two "new" alliances arrive; MCXA and TOP. The reason they're new is that they changed Alliance Affiliation. Since this process takes a few days, you usually see a sharp increase over a few days followed by a trickle up to the point where all of the members who are going to move do so.
You also see a long and steady decline in NPO membership. This is similar to (but less severe than) the loss you see many alliances face in war, which makes sense seeing as how it starts to accelerate around the time of the FAN conflict. I'm told it is a combination of policing NPO impersonators (and the general tendency of such folks to leave the alliance they're sheltering under when it goes to war) and genuine losses to the war. What I find interesting is that although they lost about 450 members over the period, but only about 2.4M NS (comparing to the ANS graph). Assuming no damage was done to the remaining NPO nations during the time (which is silly), it works out to ~5.3k NS each; given all the nukes flying around I'm sure that there actually
was damage done to remaining NPO nations that would tend to push down the average NS of those departing.
While I'm sure nobody would like losing 400+ members, I can't recall any other time such a loss was predominantly of nations so much weaker than the average (on June 10, about when the decline started NPO's AANS was about 8.6k; at the end of the measurement period it was 10.9k).
By comparison, FAN is down 265 nations and almost 7M NS, which if you could pin it all on nations leaving would mean the average FAN nation was 26k NS when they left, however in this case it doesn't match up. For the first 3 weeks or so of war, FAN's AANS dropped from around 13.3k to about 5.8k. For that period FAN dropped only 65 nations and 4.8M NS. That tells me that most of the NS lost by FAN was probably from the war and not from 73k NS nations leaving. After that, FAN began to lose nations much more quickly (probably to bill lock, but I'm not sure). This is also pretty unique, I can't think of any other alliance in a losing war that held out that way.
Score:Short form: S
(Posted Image)
This is a graph of alliance Score versus date. You get 1 guess as to what day Admin implemented the new score algorithm. :P
Alliance Nukes:Short form: N
(Posted Image)
Finally, some fun stuff. This is a graph of alliance Nukes against date. As you can see, we Goons
love us some mushroom clouds. Someone should probably let our Orange friends know that iron isn't fissile.
Alliance Nukes per Nation:Short form: NPN
(Posted Image)
This is a graph of the ratio of nukes to total nations in an alliance versus date. An alliance with 0 nukes would therefore have 0 nukes per nation. A 100 member alliance with 200 nukes would have 2 nukes per nation, and so forth.
Alliance Nation Strength per Nuke:Short form: NSPN
(Posted Image)
This is a graph of alliance nation strength divided by nukes versus date. It's heavily influenced by two factors: how much an alliance feels like it needs nukes their average NS. As of the 29th, the top 5% ended in a nation with 26,045.561 NS. Thus, you could do a crude guesstimate that an alliance where NSPN is 26k has roughly 1 nuke per nuke capable nation
equivalent. With the exception of some of the alliances with crazy average nation strengths (I'm looking at you, Grämlins :angry:), every 26k NS isn't actually a nuke capable nation. However few nations that are nuke capable buy only one, so there is a balance between how many nukes members in an alliance think they need (or want) and how many of an alliance's members are actually capable.
In general, the lower NSPN the more nuke happy the alliance is, and the higher their average NS is. Having a low NSPN doesn't mean you have a low AANS, but having a low NSPN requires a relatively high AANS.
Alliance Growth Trends:Short form: AGT
(Posted Image)
This is a graph of day-to-day percent growth versus date. It covers the previous 2-week period, and is restricted to the range between -5% and +5%.
This one can be a very busy graph to look at, but the idea is that an alliance whose line stays above the 0% mark most of the time grows. The higher they tend to be each day, the faster they tend to grow. The reverse is the case for shrinking alliances.
The value for any given day is ANS for that day divided by the ANS from the day before (the same as the value from the D2D chart for that alliance on that day).
Team Nation Strength:Short form: TNS
(Posted Image)
Just like the ANS graph, but for team data.
Average Team Nation Strength:Short form: ATNS
(Posted Image)
Just like the AANS graph, but for teams.
Guess when FOK invaded the Orange team, I dare you.
Bonus points (not really) for those who can spot when about the FAN war began.
Team Membership Count:Short form: TMC
(Posted Image)
Just like the AMC graph, but for teams.
You can really see the FOK invasion here, too. As with the AMC and AANS graphs, when you see a rapid rise in membership, you see a dropping average strength.
Team Nukes:Short form: TN
(Posted Image)
Just like the alliance nuke graph, but for teams. It has as a lot to do with how many nations are in a trading sphere.
Team Nukes per Nation:Short form: TNPN
(Posted Image)
Just like the NPN graphs, but for teams.
Because there are more nations involved here, the effect of more members on a team shows up better than the effect of more nations in an individual alliance: the NPN drops noticeably.
Team Nation Strength per Nuke:Short form: TNSPN
(Posted Image)
Just like the NSPN graph, but for teams.
Team Score:Short form: TS
(Posted Image)
Just like the Score graph, but for teams.
I find it funny that there are individual alliance(s) more powerful than the Pink trading sphere. It's not really surprising that No Team is beaten by a bunch of alliances.
Come to think of it, why don't we see the massive change in Score on the Team Scores that we did for the Alliances? I hadn't even thought of that until just now. Odd.
----
Oh well. That's it for now, folks. I might add some more eventually, but hopefully that'll answer any questions about what the heck the numbers are, and some of the questions about what they mean.
I welcome any discussion of interpreting trends on charts. If you look at the URL for the various images you can easily figure out how to pull up historical graphs and charts (where they exist) if you want to pull up a past situation for examination or discussion. Let's try to keep it civil though.
This post has been edited by Unspeakable Evil on Aug 2 2007, 04:02 PM