To all the flames on primal datacenter...

[Februs]

http://na.finalfantasyxiv.com/lodest...hly/?sort=rate

If this is supposed to prove anything... it doesn't.

The fact is that the loadstone standings are there for personal stats. There is no way to filter results to find out how many matches were won in total for your GC. All it does is show you the top ranked players, or the most prolific players, and how they rank against one another in a list of 100 depending on victories, win rate, or number of matches played.... that really doesn't say much about how the 3 GC's compare to each other.

If you wanted, you could try to compile your own list by counting the individual stats of the players from each GC and than average them for comparison, but that would still give you a wildly skewed result, because the pool being sampled is not even close to being equal or comparable.

For example: this week, on my server, with a sample pool of 93 players in the Victory rate tab, Mael had a total of 75 players represented with a average victory rate of 66.5%. Adders had 3 players represented with an average victory rate of 53%, and Flames had 15 players represented with an average victory rate of 46%. Based just on the victory rate alone, Mael wins the most with the highest average. However, notice that they took up 80% of the listings and not all of the players, including some of the ones at the very top of the list, played an equal number of matches. The top Mael only played 40 matches, but the 5th place Mael had almost 100. Between the two of them, which do you think had an easier time maintaining his win ratio? Obviously the guy who played less matches, since he didn't have as many opportunities to fail as the 100 match guy. The same logic can be applied to the GC's as a whole, and is proved by the Adders/Flames results. Adders "beat" Flames in win rate 66% to 53%, but their sample pool only has 3 players. How is this a fair comparison? It's not. It tells us absolutely nothing in regards to which GC preforms better in terms of win's and losses. Keep in mind that if I had limited the sample pool to the top 10 players in win rate, Adders would not have even been represented at all, meaning Mael would have still won, with Flames trailing just barely behind them... and Adders would be at 0% by default because, according to the listings, they weren't even "good" enough to be there. I'm pretty sure those 3 players who scored above a 50% win rate would beg to differ, but, hey, that's what the standings say, so it must mean Adders totally suck right?

Obviously not.

The only thing this data does tell us is that, on my server, Maelstrom have FAR MORE players queing for PvP than either Flames and Adders combined, but that's literally all we can surmise from this data. It is impressive that they maintained a high average, but there's nothing to say that if Flames and/or Adders had an equal number of players that they couldn't do the same. This tells us nothing about GC performance, only personal performance. To know how the Gc's stack up against one another, we'd need to know the total number of matches played in each GC (which is impossible to calculate because we don't know how many of the players in the listings overlap in matches), and the average win rate of each GC over the course of their total number of matches. That data is not made available to us. So trying to compare which GC is "better" or "worse" using the standings is pointless.

[Blubbers]

The fact is that the loadstone standings are there for personal stats....

You can get some useful, although slightly biased, information out of the standings if you analyze it properly.

See my analysis here for Aether a few months ago for my methodology: http://forum.square-enix.com/ffxiv/t...=1#post3215327

You have to sort by number of matches, not wins or win rate which will be strongly biased toward pre-made groups. You also need to calc the total number of wins & losses so that you weight players with more games more heavily. Last time I did this analysis it only required a small adjustment and gave reasonable looking win-rates.

[Februs]

You can get some useful, although slightly biased, information out of the standings if you analyze it properly.

That's true, but I think people underestimate how skewed the results are. To the best of my knowledge, there is no way whatsoever to account for overlapping matches between players in the match tab. So, for all we know, the actual number of matches played are substantially lower than the calculated number. That by itself has a HUGE effect on the outcome of the average.

You're numbers are by far the closest I've seen to being accurate, partially because you at least tried to account for the margin of error, but there's still no way to completely tell with certainty the actual GC standings. The only thing we can be 100% certain about is our personal standings/performance, and people are dramatically underestimating the value of that information.

The fact is that there are some players who consistently score a certain win rate (which is why I singled out that tab in my example). I, myself, usually score around 40% almost every week, regardless of the number of matches I play (usually around 40-50 a week). Yet there are other players who are consistently scoring lower and pointing the finger at their GC or being ganked by other GC's as the reason. I don't think that's the right way of looking at it. If someone has a lower win rate than the top players in the standings, they should be asking themselves what they're doing differently, not instantly blaming their GC with biased data that somehow explains it away as being typical of their performance.

[Kurogaea]

.

I never said anything about what GC is worst exactly by those standings. The standings themselves can give you a clear idea of what is going on with the GCs in the PvP servers. Btw your example is pretty much what I typed a while ago (more matches and victories), so instead of saying, "This doesn't prove anything", then come with a better solution. It was just to give others something to look at to have a better understandng of what I meant, not an answer to whats the best GC. In fact, premades aren't even a factor. Any GC can have a premade and even then players can still lose, strategies be damned. You can be the best commander out there, but your likely to be wrong, or right. 50-50. Same for the other 2 teams against you.