Thanks!! So good!I've seen a couple of graphics at the discord:
View attachment 834
View attachment 835
That's a thread clock win rate analysis from reddit:
Having recently seen a few posts that have begun to talk about determining the actual win rate for spells cast based on the threat clock, I wanted to finally dive into it and try to find some answers. To do this, I took quite a lot of screen captures and used an on-screen protractor tool to determine the angle of the various hands displayed. For this experiment, I am assuming that the faint hand closer to the red side of the clock is the base, unadjusted win rate for any given trace. The solid hands have bonuses applied for level and potions, and while determining the formulae for those bonuses would be a fun additional activity, it was not necessary for the immediate task.
By combining the base win rates from the game master and the measured angles of both the sectors of the clock and the hand positions, I filled data into a spreadsheet and began to adjust some of my assumptions to find values which would provide a reasonable fit to the observed data. This is not a supremely scientific approach, and I imagine that others will come along with better techniques to refine and hone the numbers calculated. I still don't like some of the numbers, but this was the best fit to the data I could find until we get more reliable data points.
It appears that within each sector of the threat clock, the win rate is linearly distributed. (i.e. If you go twice as far within a segment, the win rate will change twice as much.) However, the sectors of the clock have wildly different ranges. Sector 1, by itself, contains all win rates from 100% to 40%. If your cast doesn't end up in this sector, you have less than a 40% chance to win. Additionally, sectors 2 and 4 have an unusually small range assigned to them, compared to the other nearby sectors. The upshot of this is that if your range overlaps one of these segments, your cast will very likely land within that region, making it difficult to actually improve your odds.
Consider this: You try to return a trace which has an adjusted base win rate of 35%, and a maximum rate of 50%. Unfortunately, that includes the entirety of Sector 2, which means that its 3% range of win rate will take up something like 80% of your cast bar. A low Fair cast will get the 37% rate of the bottom of Sector 2, while a high Great cast will only get the 40% rate of the top of Sector 2. To actually make a difference to your win rate, you would need a high Masterful to try to take advantage of the small section of darkest green in Sector 1, where you could suddenly jump to a 50% win rate.
I don't expect this to be the end of the discussion, nor do I expect these numbers to be perfect. I do hope we can continue to investigate and refine this information, and hopefully use it to temper our expectations of winning a trace which is "only" in the yellow sector. (Because that means you've got at best a 1 in 5 shot at returning it.)
Thanks to the other posters and the Discord denizens who helped provide information. Come visit the research channel!
Probably, there's a similar "confoundable departure rate" depending on the threat...