[EM] Markus, 18 May, '05, 0423

[MIKE OSSIPOFF]

Markus--

You say:

I proposed that heuristic for the Schulze method that uses Schwartz sets
in 1998:

[...]

I reply:

Yes, Markus, you just finished posting those two links, and I just finished 
commenting on them. I guess it will be necessary for me to repeat what I 
said: You posted a definition of CSSD. SSD is not CSSD. Well, that isn't 
everything that I said, but I'm not willing to say it all over again just 
because you repeat what you said yesterday.

You say:

In Feb 2000, we discussed which Condorcet method should be proposed to
the Debian project. See e.g. here:

I reply:

You weren't included in that discussion, because it was felt that you'd 
advocate one particular method without being willing to consider other 
proposals. Maybe there was also some discussion about that on EM.

You continue:

In your 15 Feb 2000 mail, you discussed 3 methods: "Drop Contradicted
Defeats" (DCD), "Sequential Dropping" (SD), and Schulze:

I reply:

As I said yesterday, and as I said last month, in reply to the same repeated 
statement: At that time I thought that "Schulze's method" meant 
BeatpathWinner. But last month you clarified that "Schulze's method" does 
not mean BeatpathWinner, because BeatpathWinner is an instance of a larger 
class of methods that you call "Schulze's method".

You apparently have nothing to do but repeat the same statements that you 
made yesterday.

You continue:

Then in your 18 Feb 2000 mail, the term "Schwartz Sequential Dropping" (SSD)
was used for the very first time. In that mail, you did not only admit that
you knew my method, you also admitted that you were aware that SSD _is_ my
method:

I reply:

At that time I thought that BeatpathWinner was your method, not realizing at 
that time what "Schulze's method" actually means. And, as I said yesterday, 
and as I said last month, in reply to the same repeated statement, I, at 
that earlier time, carelessly said that SSD is equivalent to BeatpathWinner, 
because it's equivalent to BeatpathWinner when there are no pairwise ties. 
When there are pairwise ties, SSD is not equivalent to BeatpathWinner. CSSD 
is equivalent to BeatpathWinner, but SSD is not. I've admitted that error 
last month, and yesterday, but you keep repeating that quote.

You continue:

Therefore, it is clear (1) that you knew my method when you "devised" SSD

I reply:

I knew of BeatpathWinner when I and Eppley devised SSD. I at that time 
thought that BeatpathWinner was you rmethod, because at that time I didn't 
realiize that, as you said last month,  "Schulze's method" means something 
different.

I didn't know of your CSSD definition, because I hadn't read those postings, 
and didn't know what count rule they described. But that's irrelevant, 
because SSD is not CSSD.

You continue:

and (2) that it was clear to you from the very beginning that SSD _is_ my
method.

I reply:

I and Steve had never heard of SSD or CSSD at the time when we devised SSD. 
But, in any case, SSD is not CSSD, and is not your method in any sense. Even 
if we had known about CSSD, SSD is not CSSD, and is not your method.

I'd said:

>Where did SSD come from? In individual e-mail, Steve Eppley suggested a
>method that successively drops the weakest defeat among the smallest set
>of candidates that is unbeaten from without. SSD is an Eppley-Ossipoff
>method.

You say:

Please forward this communication between Steve Eppley and you.

I reply:

I don't save e-mail back to those days. But my mention of SSD on EM is in 
the EM archives. That's really all you need.

You continue:

I would
like to know why (although you considered neither independence of clones
nor reversal symmetry important and although Steve Eppley decided to
promote Tideman's ranked pairs method) you decided to promote SSD.

I reply:

SSD is clone independent in public elections. Only in small commitee voting, 
where there can be pairwise ties, is SSD not clone-independent.

I advocate SSD for public elections. For small committees I recommend CSSD 
or Beatpathwinner.

So SSD is clone independent for the application for which I recommend it.

Reversal symmetry? You're right. I don't consider reversal symmetry 
important. Pretty much all criteria sound plausible and desirable, but that 
doesn't make them important. I'm not going to argue with you about whether 
reversal symmetry is important. You're welcome to believe that it's 
important.

You ask why I decided to promote SSD. Because it's clone-independent in 
public elections, and because it meets SFC, GSFC, WDSC, and SDSC. And 
because it doesn't require mention of cycles or beatpaths.

And you seem surprised that I'd promote SSD when Steve prefers MAM. Why do 
you expect me to copy Steve on that? Actually, Ranked-Pairs has a very brief 
definition, if we ignore the rules for equal defeats. But those equal defeat 
rules can't be completely left out. And RP's definition, directly or 
indirectly, mentions cycles. So I consider SSD to be a better public 
proposal.

In the matter of SSD vs RP, in terms of pure merit, in public elections, RP 
might very well be _very slightly_ more aesthetically appealing, because RP, 
unlike BeatpathWinner, CSSD, or SSD, never lets a nullilfied defeat 
participate in the nullification of other defeats.

But I consider the merit differences between RP and SSD completely 
negligible in public elections. That's where Steve and I disagree. What's 
wrong with disagreeing?

So I propose what, it seems to me, is more likely to be accepted.

I'd said:

>We've recently been all over that. I won't repeat that argument, but, to
>summarize the conclusion that we reached recently,  I admitted that I was
>mistaken when I thought that "Schulze's method" means the method that I
>call "BeatpathWinner". "Schulze's method" does not mean BeatpathWinner.

You say:

Your "reply" doesn't make any sense. In 2000, you _defined_ the
term "beatpath winner" as a synonym for the Schulze method.

I reply:

At that time I mistakenly believed that 'Schulze's method" meant the count 
rule that I call "BeatpathWinner". I've only explained that about five times 
in the last 24 hours, and at least a dozen times last month.

You continue:

The
term "beatpath winner" hadn't been used before.

I reply:

Correct. I coined that name for the method that I still call by that name. 
The method that I mistakenly thought that "Schulze's method" meant.

You continue:

Therefore, it
doesn't make any sense when you now claim that you only mistakenly
believed that "beatpath winner" was a synonym for the Schulze
method.

I reply:

I coined "BeatpathWinner" as the name for the count rule that I still call 
by that name. Earlier, I mistakenly believed that "Schulze's method" meant 
that count rule.

You continue:

The terms "beatpath winner" and "Schulze method" are
synonymous terms because you _defined_ them to be synonymous
terms.

I reply:

I coined "BeatpathWinner" to refer to the count rule that I still call by 
that name. At the time you're referring to, I mistakenly believed that 
"Schulze's method" meant that count rule.

Now, if you have anything more to say, you might want to try to be sure that 
it's something new, something different from what you've already been 
saying. To most, that might be an obvious good idea, but it's something that 
you don't seem to understand.

If you post more repetition of those same statements that I've already 
answered many times, I won't reply. No one can stop you from posting the 
same statements again, and, knowing you, it's a safe bet that you will. But 
I'm not going to answer them again for you. I won't reply.

Mike Ossipoff

_________________________________________________________________
On the road to retirement? Check out MSN Life Events for advice on how to 
get there! http://lifeevents.msn.com/category.aspx?cid=Retirement