I did again a bit statistics because I was interested how the GJ alliance influences the game. Here is how often Greyjoy fights vs Lannister or Stark respectively in a game:

https://drive.google.com/open?id=19b5XVjZtsFfY_V-MAzWpyBWf2IXw23Nn

You clearly see two peaks indicating that one can indeed distinguish between games with different alliances.

Unfortunately I cannot resolve fights up to a certain round, since the logfiles are a bit inconsistent or I am just too bad at coding , so i define alliances as:

There is a Greyjoy-Lannister Alliance if: Greyjoy fights 4 or more times vs Stark and 2 or less vs Lannister

There is a Greyjoy-Stark Alliance if: Greyjoy fights 4 or more times vs Lannister and 3 or less vs Stark


Here are the winrates

Greyjoy-Stark Alliance:
https://drive.google.com/open?id=1z1gZNwudlDdWKrbetMnsEpHlH3PtatWX

Greyjoy-Lannister Alliance:
https://drive.google.com/open?id=13QtOJruiHtt5q5NVZy6IPLOr6l_8MDEB

For comparison here again the general winrates in my set:
https://drive.google.com/open?id=1jW71ErXMav9UpT2_ijWwOexADohbvMX9

Interesting I found also the games where Greyjoy fought 8 or more times vs Lannister (4 or less vs Stark):
https://drive.google.com/open?id=1Yl7YUHTY-LnhtIFIp_7k0vV2hqlBL8l3

Let me know what you think!

Last edited by PPPPaco on 2018-May-11 11:57, edited 1x in total

The set contained all 6p games starting at ID 90000 up to 150000. Only restriction was no ToB and second edition House Cards. So yes, there is an offset of EMP games, endless games, beginner games, westeros phase, random start etc.. I was too lazy to differ between them but I think the majority will be the "normal" games.

Last edited by PPPPaco on 2018-Apr-10 20:16, edited 2x in total

Thank you for the analysis PPPPaco. I'm surprised nobody has replied yet.

As for the results, I'd say no big surprises, here's what I think are the main points:

1) Whomever GJ decides to attack is going to get boned. Since they are the biggest power at the start, the house they decide to attack will have a very difficult game.

2) GJ has a better chance of winning when having an alliance with lannister rather than when they are at war. This comes as no surprise to me since it has been my observation as well, both from game experience and theory-crafting .

3) Tyrell is happy when the Lion and Squid are fighting, while Martell is happy when they have an alliance. My biggest surprise is the huge martell spike in winrate with a GJ-Lanni alliance. Yes, it does favor Martell if those 2 are allied, but effectively doubling the win-rate? Are you sure the numbers are correct?

It makes total sense to me that a strong Lanni puts pressure on both of Martell's opponents.  While at the same time, if no one puts pressure on Stark, it basically eliminates anyone else's chances to eke out a 5 castle win with a tie-breaker.  Because Stark is essentially the only house that can win the game without winning a war, Martell's only route to win becomes to completely conquer a neighbor to win before turn 10, and they only have two options for doing that where most have three.  If the board is more balanced, just taking a modest bite from a neighbor can be enough.

I think the biggest "hole" in this data is that based on my observation, Greyjoy fighting Stark means they have an alliance with Lanni.  But just because they are going for Lanni doesn't mean they have an actual alliance with Stark.  Stark alliances are the most passive in any case, but most of the time it seems to me that Greyjoy is just counting on Stark dottering around on in the North long enough for him to crush Lanni, rather than any active support.

nice work once again PPPaco.

Impressive that there are actualy games where GJ is fighting over 20 times against Stark OR Lannister... :-D
Goot take a deeper look into those stats.

---

"Build a man a fire, and he'll be warm for a day. Set a man on fire, and he'll be warm for the rest of his life." , Terry Pratchett

nice work once again PPPaco.

Impressive that there are actualy games where GJ is fighting over 20 times against Stark OR Lannister... :-D
Goot take a deeper look into those stats. Yeah, 20 is a lot, I don't think I've ever had 20 battles with a single other player. But then again, if the opponents are evenly matched, and the game goes to turn 10, that's slightly more than 2 fights per round from turn 3 onward (considering you won't have 2 fights per turn in the first 2 rounds), which is entirely possible.

And like I said in my previous post, it makes total sense for Martell to have a better time during a Lanni-GJ alliance ... but doubling the win-rate? :O

And like I said in my previous post, it makes total sense for Martell to have a better time during a Lanni-GJ alliance ... but doubling the win-rate? :O
I am guessing you would find a lot of round 10 wins there, which are more probable when the game is balanced-out by a Western alliance.

---

“The only way to treat disease is to understand disease.“

Could very well be Doc! But I'm usually skeptic or wary of such large changes, and a doubling of win-rate from 10% to 20% does fit the category of extraordinary claims