Take GR survey.
Take it NOW.
http://www.surveymonkey.com/s/QDTCPJW

Why I am not the list?

Either you are in the wrong thread or you think you yourself are a GR subscore.
I would go with the former.

Okay. I am Subscore.

Because the latest ghost rating list is as of March 1st and you don't have any games that ended before March 1st. You will be in next month's rankings.

Oh, that's what he meant.

Oh okay,

I had the 'flu & an asthma attack earlier this week, and currently sit in that awkward place between being too ill to do anything and being well enough to do anything useful. So hello!

fwiw, I'd not take survey results terribly seriously myself. This comes down to two things:
1. I'm a statistics nerd, and I'm going to have as many as I have the energy for, provided:
2. They are meaningful. We could run GR on gunboat games only, but who is to say that it will be a better predictor (in general) of a result in a gunboat game than regular GR, where we have more data (albeit less good). Whether or not it is better is measurable.

That said, if everyone is very keen on an optimised list for <x> then we can try to produce that in a systematic way

@Ghost,
If someone was really interested, they can do an ANOVA analysis to see what correlates with what. I am too lazy to do that and it requires me to use SPSS and such. The key thing would be whether or not overall GR highly correlates with each of the individual subscore.

Though it would make for a fun statistics project, calculate GR and the individual subscores for each person and then run regression analysis on them to see what correlates and what, if any, would most accurately predict performance in each category.

Sorry to hear about your illnesses, TGM.

http://www.surveymonkey.com/MySurvey_Responses.aspx?sm=hXZPstnloIoTVUJtQFjVybq9cVZk37%2fi7iss6%2b%2fHltI%3

Here is summary of the data thus far:
WTA full press non-live classic: 7 out of 17 votes
WTA GB (all variants): 6/17
GB (all variants): 8
PPSC: 4
Live: 7
Two honorable "other" voted WTA GB classic

It looks like if we do WTA full press non-live classic, WTA GB classic, GB, PPSC, and Live, we would capture most of what people want. Whether or not we need two separate categories for GB is debatable.

So I have done a partial regression analysis of 16 people using their peak, current GR, WTA full press (all variants), WTA GB (all variants), and GB (all variants) using SAS and Proc Corr. Here is the document link:
https://docs.google.com/document/d/1MWZOrbIBQZBXmVulFRUdHpr6_ZgZMS2se4I2WVifc5k/edit

So, for the Pearson correlation, the higher the top number, the stronger one value correlates with the other. This is why Peak and Peak has an R value of 1. Peak and current GR has an R value of .738 or 73.8%, which makes sense.
From what little data I have, it looks like the best predictor for your overall GR is how well you do in WTA full press. Gunboat games have only a 3% correlation to overall GR, which means that the overall GR does not at all reflect someone's gunboat ability. GB and WTA GB, however, correlate at 99.3%, which essentially means knowing someone's score in either GB or WTA GB allows you to accurately predict the other. So when we calculate the GR subscore, we would not need to do both GB and WTA GB.

I am not sure if anyone cares about this. If you do, please let me know either here or through private PMs.