Index ← 3657 CFJ 3658 3659 → text
===============================  CFJ 3658  ===============================

      The Treasuror's report of August 27, 2018, or a portion thereof,
      is doubted.


Caller:                        twg

Judge:                         G.
Judgement:                     TRUE



Called by twg:                                    27 Aug 2018 19:21:58
Assigned to G.:                                   09 Sep 2018 19:17:32
Judged TRUE by G.:                                16 Sep 2018 20:44:33
Motion to reconsider filed by G.:                 23 Sep 2018 07:01:22
Judged TRUE by G.:                                29 Sep 2018 19:13:37


Caller's Evidence:

[Arbitor's note: the underlying question with the below CoE is, if there's 
a "Treasuror's Report" that contains multiple currencies that are listed 
in columns, does a CoE that cites an error in one currency on the report 
cast doubt on the whole report, or just that currency?]

On August 27, 2018 5:47 PM, G. wrote:
> CoE: If incense is defined in the new ruleset, it was never repealed
> and we should all have whatever we had when it was last reported,
> unless a report ratified that explicitly stated Incense was 0.
> Only changes would be if, say, some of us transferred it using "all
> liquid currencies" or the like.
> Since each asset-type report self-ratifies independently (I think), the
> Treasuror's Reports that were missing incense data should be interpreted
> as just not having that data and being incomplete reports, with no
> implication that the missing data were self-ratified to 0.

Caller's Arguments:

I initially assumed the lack of a specified incense balance meant it was 
at its default value (0), but rule 2166/26 defines any "portion of [a 
recordkeepor]'s report" that is "a list of all instances of [a class of 
assets] and their owners" - not the entirety of the recordkeepor's report 
- as a self-ratifying document.

An interesting repercussion, if that interpretation is correct, is that 
CoEs can be made against the balances for _specific asset classes_, 
without blocking the rest of the report from self-ratifying. This might 
potentially mean that many, many previous CoEs against the Treasuror 
report (and possibly other reports?) have never been valid doubts, because 
the way we usually phrase it - "CoE: X is wrong" - is ambiguous: it might 
be a CoE against any one of the self-ratifying documents that make up the 
Treasuror report.


Judge G.'s Arguments:

Imagine three possible claims of error referencing a hypothetical 
officer's Report of Currency X:

A.  "CoE on the Pointlesskeepor's Weekly Report:  Player A does not have
     N units of X".
B.  "CoE on Currency X holdings:  Player A does not have N units of X".

C.  "CoE: Player A does not have N units of X."

Colloquially, these are all within the range of reasonable and clear
communication of the issue at hand (Player A's holdings of X).  Since the
goal of self-ratification is to reduce uncertainty, it would be nice if
these variations all led to the same result.  If variation (A) blocked the
self-ratification of the whole report (including other Currencies), but
(B) blocked the self-ratification of the Currency X holdings only, while
(C) blocked the self-ratification of Player A's X holdings while allowing
other Players' X holdings to self-ratify, then teasing out what facts had
self-ratified would be confusing to say the least. Is there a way to
construe statements A-C so they all lead to the same result?

Rule 2201/7 reads in part:
>     A doubt is an explicit public challenge via one of the following
>     methods, identifying a document and explaining the scope and
>     nature of a perceived error in it:
So with this, a CoE or other doubt (CFJ) has three parts:  The document in 
question, the scope of the error, and the nature of the error.

In all of the above examples, the "nature of the error" is clear:  Player
A does not have N units of X.  So far, so good.

What about "document"?  A "public Document" is defined by R1551 as "part 
(possibly all) of a public message."  This narrows things down in one
direction - a single document cannot be spread over multiple messages. But
within a message, this would allow every section, or even every single
character of a message to be a separate "document".  One could even say
"I CoE on the document made up by taking every other letter of the
following report." 
However, looking to the common definition of "document" gives:
>   A piece of written, printed, or electronic matter that provides 
>   information or evidence or that serves as an official record.

The key point here is that a document provides information or is an 
official record.  At the extreme, a single character provides no 
information on its own, so this suggests there's a minimum information
content for a document to be a document.  And for these purposes, the
"official record" in question is clearly defined in R2166/26:
>    That entity's report includes a list of all instances of that class 
>     and their owners. This portion of that entity's report is self-
>     ratifying.

This list is described as a unit, so a full list of Currency X holdings is
the minimum "unit of information" that could be considered a self-
ratifying document.  Further, the fact that multiple "documents" can be
comingled means that each Currency List is a separate "document", even if
the information is in columns or otherwise combined with other "documents"
(e.g. in columns of multiple currencies). 

This is specified more strongly for the other major type of
self-ratifying record, Switches (R2162/13):
>                                                    a public document
>        purporting to be [the switch values] portion of that officer's 
>        report is self-ratifying, and implies that other instances are at
>        their default value.        
Since missing values here are assumed to be at their default value, a
list of this sort self-ratifies or is CoEd as a whole.  This general
principle is sensible to extend to assets as well as switches.  So in the 
examples above, both (B) and (C) would have the effect of stopping self-
ratification of the complete list of Currency X holdings (including for
any entities who are wholly missing from the list).

But what about example (A)? ("CoE on the Pointlesskeepor's Weekly Report: 
Player A does not have N units of X".)  Is this a CoE on every list (i.e.
all currencies, switches, etc.) in the Report?
Here is were "Scope" of the error comes into play.  "Scope" is the range
of records affected by the error.  And the most reasonable and useful way

This is key.  So, if a message says "CoE on the Report" but doesn't cite
the explicit error, the scope is 0 (null).  It's not a CoE.  If the report
says "CoE on the Report:  Record X is wrong" then the scope is Record X,
and by the argument above the MINIMAL record is the full list of X, but
not the full report.

In other words, regardless of what document the caller CLAIMS is in scope,
the true scope of the CoE (the true document) is always the minimal but
complete list that contains the class of records identified as the error.

In the current situation, the alleged CoE text included the text of the
Report named in the CFJ statement, with the following:
> CoE: If incense is defined in the new ruleset, it was never repealed
> and we should all have whatever we had when it was last reported,
> unless a report ratified that explicitly stated Incense was 0.

The report contained a record of incense (at the time for one player).
Therefore, the CoE was explicitly on the full list of Incense holdings
(a portion of the report).  TRUE.