MatPlus.Net

 Website founded by
Milan Velimirović
in 2006

5:19 UTC
ISC 2024
 
  Forum*
 
 
 
 

Username:

Password:

Remember me

 
Forgot your
password?
Click here!
SIGN IN
to create your account if you don't already have one.
CHESS
SOLVING

Tournaments
Rating lists
1-Jan-2024

B P C F





 
 
MatPlus.Net Forum General What does 'C+' mean?
 
You can only view this page!
(1) Posted by Anders Thulin [Monday, Sep 26, 2022 09:43]

What does 'C+' mean?


What does 'C+' mean? Is there a formal definition or general consensus
formulated somewhere?

ECP ed. 3 says that a problem with this indication 'has been
computer-tested and correct'. (Presumably 'and been found to be correct')

But 'correct' is not defined. Computer-tested and found to have as many
solutions as the stipulation implies? Perhaps even with the same key-moves
as in the author's solution? Perhaps even matching the author's solution?
Or possibly with even more tests?

The reason I ask is that I found a couple of entries in the PDB
database marked as 'C+, except for a dual after ...' which suggested
to me that duals were part of C+-checking, and should not be present.
And while that could be part of computer-testing, it was more than
I had thought was included, myself.

(ECP ed. 3 goes on to imply that 'C-' and 'C?' are synonyms, and mean
'not-C+' but that is surely not fully correct, or possibly not
complete. But again, unless there's an agreed-upon definition ...)
 
(Read Only)pid=23820
(2) Posted by Kevin Begley [Monday, Sep 26, 2022 11:44]

What does 'C+' mean?

C+ should indicate:
a) some solving tool (which should be specified) provides only a solution set which agrees with the "intended solution set" provided by the author,
b) there is no outstanding claim against the solving tool's capacity to correctly render a solution set compliant with the problem (diagram, conditions, stipulation, number of solutions, etc).

If the duals are mentioned, they are part of the intended solution set.
Mention of a minor dual (read: a dual of no thematic consequence) may slightly devalue a problem.
Mention of a major duals (read: a dual of thematic consequence) may significantly devalue a problem (and may completely nullify the intended thematic content).
Duals not mentioned in the intended solution set will not achieve agreement with any functional solving tool.

What does 'C?' and 'C-' mean?

C? should indicate either:
a) no solving tool can provide a solution set which agrees with the "intended solution set" provided by the author, or
b) every solving tool which does provide agreement has an outstanding claim against its capacity to correctly render a solution set compliant with the given problem.

C- should indicate:
a) some solving tool provides a solution set which does not agree with the "intended solution set" provided by the author, or
b) some solving tool which does provide agreement has a valid claim against its capacity to correctly render a solution set compliant with the given problem.

Who determines the validity of these claims?
Unfortunately, there exists no official arbiter -- the moment you create an official body, vested with authority to arbitrate such matters, they will be inundated with Joke Problems (testing our definitions, and the language of our definitions, to the extreme), and their every ruling will spawn more Joke Problems. One faulty ruling from an unappealable authority, it is feared, may lead to the complete collapse of our artform.

Therefore, an overwhelming consensus of the problem community (especially those invested in maintaining order) is generally the goal (anything less is sufficient to sow the choking vines of doubt).
Where consensus of the problem community is not achieved, all parties with an interest in maintaining order and optimizing productivity must resort to passionate debates (often in forums like this one).
We each may have some small influence on the future, but ultimate truth rests in the hands of our progeny.
In other words, someday Artificial Intelligence will have these debates for us. ;)
 
 
(Read Only)pid=23821
(3) Posted by Anders Thulin [Wednesday, Sep 28, 2022 19:23]

While impressive in all details, this answer has the feel of 'what C+ ought to mean' rather than how it is actually is used today. That more restricted
aspect is what I was interested in, but I don't seem to have expressed my question well enough.

So let me rephrase: Chess editors, if any present, what do you need or do before putting a 'C+' indicator (or even C? or C-) on a problem?

And if there are any chess problem historians who know: when did C+ start, and where were any details about it printed?
 
   
(Read Only)pid=23838
(4) Posted by Kevin Begley [Thursday, Sep 29, 2022 19:45]

@Anders,

I'm not currently a problem editor (and I further confess I was an editor for only a very short span), but I can speak to how "C-" is often used today.

For the Win Chloe database, "C-" -- which is synonymous with "incorrect" in Win Chloe searches -- signifies the problem is demonstrably incorrect (whether this was demonstrated by a human solver or by some solving tool).
It may be cooked in as many moves, it may be demolished in fewer moves, it may prove to have no solution (e.g., the intended solution was never legal, or the intended solution did not actually achieve the stipulated aim), or it may have consequential duals which were simply not provided by the author (you'll find this often in studies which are marked "C-" only because the duals were not provided in the intended solution).

Thus, my previous definition for "C-" was inadequate (even when intended, as you correctly surmised, to convey how it ought to be defined), because I simply failed to note that a human might demonstrate a problem deserving of "C-" (even when no computer has found the divergence from the stated intent, and no computer has verified in any way that the accepted divergence which was provided by a human solver does indeed spoil the intent). It is left for the reader (or the author) to challenge that distinction.

In the last ever issue of StrateGems, however, you will find "C-" is often used to signify a problem was simply not demonstrated to be sound by any solving tool (thus, they use "C-" to mean "C?").
This dichotomy is sufficient to illustrate that we have no standard definition for the term, "C-", and no standard usage.
At best, the most universally accepted interpretation for "C-" (wherever you see it) would be: the problem has not been verified correct by a computer.
I expect you already know that (or you would not have asked), so I don't understand the intended point of your question (if not to petition what the standard definition ought to be).

As for "C+"...
I doubt you would find problems marked C+ prior to 1983 (I know Jarosch tested his famous Babson with a computer, but I do not believe he obtained full verification of "C+").
Perhaps #2 problems go back further. Somebody will correct me if my assumption is incorrect.

In the PDB database, any problem marked "C+" will also provide a secondary item ("Computer test"), which generally appears to the right, in bold, just under the Genre classification.
This is where some human will report what solving tool they used, and how long it took (on their computer). That can be helpful for those seeking to reproduce the result, but unfortunately, such an indication is only as good as the integrity of the human who testifies to seeing a computer has verified the problem (and that person's integrity can only be established by reproducing what they report).
If I tell you I verified a proofgame in 52 moves on a supercomputer that ran for seven years, I might as well be telling you that I saw a sasquatch in Sequoia National Forest, back in 1951 (unless there's a plastercast of the computer's steps in solving the problem, you have little to go on, beyond your belief in me, and my description of the sound the creature made -- which you would not want to hear twice in your lifetime).
 
   
(Read Only)pid=23857
(5) Posted by Anders Thulin [Saturday, Oct 1, 2022 09:12]

OK, thanks for all those details. That makes it more of an 'informal way of expressing
aspects of problem quality problem state that in general cannot be interpreted unless a
policy provides such interpretation for the special case.'

My original impression was that C+ was an indication of 'technical correctness', i.e.
the computer software finds the number and length of solutions implied by the stipulation.
And then ECP and PDB usage suggested a more elaborate usage.

I may make an attempt to hunt down the origins of C+ to try and find out what
the original ideas were ... but it seems that it has spread in various directions, and
become difficult to interpret as a general quality indication.
 
   
(Read Only)pid=23895
(6) Posted by Kevin Begley [Saturday, Oct 1, 2022 21:15]

@Anders,

The history of "C+" would make for a fascinating read.
In fact, the greater history of chess problem solving tools (including EGTB) -- especially concerning the steps toward composing algorithms -- would be a real treat (now would be the ideal time to start gathering that, because I suspect that history will making fast advancements in the coming years).

Kindly share any of that history you manage to wrangle -- in this thread (or drop us some links).
Should I stumble upon any information, I will send it your way here.
 
 
(Read Only)pid=23905

No more posts


MatPlus.Net Forum General What does 'C+' mean?