<div dir="ltr"><div>Markus--</div><div> </div><div>Either I missed your posting when it posted (maybe I didn't get it by e-mail), or, if I replied to it, I probably replied in terms of an earlier deterministic RP version. So let me answer your example for my last-posted HDRP version, which is what i now and in the future will mean by HDRP</div>
<div> </div><div>In your exampl described below, A wins.</div><div> </div><div>BA is discarded because it contradicts AC and CB. </div><div> </div><div>DA is discarded because it contradicts AC and CD.</div><div> </div><div>
Neither of those two discards is affected by any other discard, and so there's no question of which one happens. So this example doesn't bring out any contradiction between two discards (But now I'm not sure whether or not such contradiction is possible). </div>
<div> </div><div>So A is the candidate with no not-discarded defeats. A wins.</div><div> </div><div>But, even though your example doesn't have any two discards that contradict eachother, that example gives me doubt that maybe such an example could be written. An example in which each of two defeats qualifies for discard only if tahe other isn't discarded. If such an example is possible, then it would be necessary to define HDRP in terms of an ordered procedure. That would be regrettable, because I really like the brief time-independent definitions of RP versions.</div>
<div> </div><div>Michael Ossipoff</div><div> </div><div>CD would be discarded because it contradicts DB, BA, and AC. But it isn't, because BA is not a not-discarded defeat. </div><div> </div><div>example:<br><br> Suppose the defeats are (sorted from the strongest to the weakest):<br>
<br> D > B<br><br> C > B<br><br> A > C<br><br> B > A<br><br> C > D<br><br> D > A<br><br>Who is elected in this example by your latest MMV definition?<br><br>Markus Schulze<br></div></div>