I just finished a 1v1 tournament on Project:Oblivion with 2 wins and 5 losses (it was a rough tournament). Checking my luck stats for each game, I noticed that I was in the positives on 1 out of 7, and my average luck for a game was -5.6 (median -8.9). So overall I had quite bad dice, especially for a tournament of 2-player games. So it was comforting, in a way, to be able to blame a 2-5 record on the rolls.
That said, I have no idea how the dice went for other players in the tourney (without looking at each individual game and tallying them up...). So, I was wondering if it might be a good idea to calculate luck stats across an entire tournament? How hard would this be to do? Would it be worthwhile?
As interesting as Luck Stats are, and I like to check them during games, I always question their "after the game" relevancy. On many boards, most people would agree that good luck early in a game is much more valuable than late luck. Consider that once a player is running the board and capturing cards, poor luck may have little impact on the big picture. My point: there's a lot of garbage luck stats in most every game.
That said, I do find value in looking at the LS graph over shorter periods to see, for instance, why a specific campaign was so ineffective. The jury's still out about how streaky WG dice are, and I don't have a strong opinion on the subject (other than I think the jury's still out), but I'm always intrigued when I see spikes on those graphs.
I agree with M57. Early luck(first few rounds) matters a lot more than overall luck.
Anyway, I have another example. 1v1, no forts. My average dice 2 rounds in was -4.5. My best in the entire tournament is -3 after 2 rounds. Only match left to salvage the tournament and prevent me from going 0-7.
Yeah, overall luck stats won't tell you that much. I had a game where I run up +8 early into a completely winning position and then managed to hit -25 before finishing my opponent off, which only served to prolong the game a couple of extra turns.
Then there's the whole impact of scale on things.
http://www.wargear.net/wiki/doku.php?id=general:luck_stats#the_most_common_luck_stat_fallacy
If there was a LS model that could be presented as a Percentile score and based on a Z-score, that would add a lot more context. There are many a threads here on the subject. Bottom Line: They require too many resources and/or they are too difficult to implement.
So despite the issues with luck stats in general, would it not still be interesting to see some aggregate or compilation of your luck across all the games in a tournament? What if you luck graphs for each game in the tournament could simply be overlaid on top of one another in the same chart, such that you could see graphically, to take Pratik's (unfortunate) example, that in every game your dice took a dramatic dip early?
Sure, if it's easy to make something like that I'm sure I'd look at it.
Kjeld wrote:So despite the issues with luck stats in general, would it not still be interesting to see some aggregate or compilation of your luck across all the games in a tournament? What if you luck graphs for each game in the tournament could simply be overlaid on top of one another in the same chart, such that you could see graphically, to take Pratik's (unfortunate) example, that in every game your dice took a dramatic dip early?
+1
Yep, agreed, it would be nice to have it.
As a graph based on average at that particular turn number. And a second graph based on median at that particular turn number.
More info the better!
Don't know that I agree for a site like this. Useful info may be better than worthless info, but I don't doubt that a good many people misinterpret luck stats right off the bat, which (for them at least) turns anything derived from LS into Frankenstats.