English

1. My experience is that accu influences the dodge-adv_dodge relation as well. I as well as my alts used accurate attack to make training on stingas less tedious. The effect can be tested by disabling accurate attack from the stanza. The hit rate then drops remarkably.
2. I don't think so. If the random algorithm is implemented properly (and I ran a couple of tests in fight and crafting corroborating this assumption), the process should be ergodic, meaning that average over space equals average over time. Honestly, I did not test this hypothesis on the wheel proper (would be hard to gather a sufficient number of participants).
If randomization is implemented properly, each wheel turn should be independent from the previous or following ones no matter wether they are happening at a time or subsequently. I shall test a 2 player situation and note the win rate playing alone and together once I have murdered enough eggs for the next 500 tokens :).
3. No. I studied attack ranges as a beginner, going so far that I approached and followed KP in the roots. to my experience, the attack ranges never changed over my level progress, and remained the same for lower level alts.