Checking CG for missing tournament games

[Jump to <Biel (2004)>]

What the title says. Or in other words, looking for incomplete tournaments.

As <crawfb5> (a specialist, given his interest in early American tournaments) points out, some tournaments will forever be incomplete, because the missing games are truly missing, i.e. gone from the historical record forever.

Well, forever, barring somebody discovering an obscure record in a dusty notebook in the attic, or some equally obscure newspaper source coming online. These things do happen from time to time.

But very rarely. Incomplete tournaments are a sad fact of life.

Still, those <CG> tournaments missing games are not to be given a free pass so easily. The diligent biographer must completely document what games are truly missing from the historical record, and thus the tournaments (the expected), from those games are missing due to error (the unexpected).

In this business one must expect the expected, and verify, verify, verify!

There are error oversights in the small sample of tournaments examined, that justifies the effort to examine all the tournaments. Presumably this was already done when the games were inducted into the tournament index, but then again, perhaps it wasn’t.  Indeed, perhaps the induction procedure should be re-examined with a wary eye towards building more redundancy and reliability into the system.

I’ve broached the subject of tracking tournaments in a centralized location on <CG>, a sort of parallel to tracking players in this forum post:

http://www.chessgames.com/perl/kibitzing?kid=FBIO&reply=11177

Let me show the output of a new routine I wrote over lunch to try to locate missing games. I’ll demonstrate on a not-so-randomly selected tournament:

>>> mtour(t)  # M for missing, as in missing games in a tournament.

39703 - Dutch Womens Championships (2003)

Player list (6 players):

   Bensdorp, Laura                52194    4.4    3,5
   Hamelink, Desiree              50887    5.5
   Lanchava, Tea                  49853    5.4    3
   Peng, Zhaoqin                  17360    5.5
   Schuurman, Petra               49856    4.5    5
   Van Weersel, Arlette           50677    5.5


N_games =  28

Tournament is assumed to be RR2   (2 missing)

Round counts:

    R1  :  3    R2  :  3    R3  :  2    R4  :  3    R5  :  2
    R6  :  3    R7  :  3    R8  :  3    R9  :  3    R10 :  3


Missing pairings:

    Bensdorp, Laura    52194    4.4    (3,5)  //  Lanchava, Tea           49853    5.4    (3)
    Bensdorp, Laura    52194    4.4    (3,5)  //  Schuurman, Petra        49856    4.5    (5)

Skip down to the end, the <Missing pairings> listing.

There are two games listed, where I’ve adopted <crawfb5>’s use of a slash to denote a pairing with unknown color.

The routine just uses a simple heuristic to determine what kind of RR the tournament is, and then looks for missing pairings. It tracks information for each player, noting which rounds are missing and also counting games with each color. So, given a reasonable guess as to the kind of tournament, missing pairs can be listed, with information to indicate what round and what colors the game should have.

Follow all that? Try writing it down to explain it!

OK, let’s break it down. Two games are missing, the player’s names given in brown, their <CG> id’s in blue. Then, for each player the doublet, (N_white.N_black) is given in green. Finally, the missing rounds are in orange.

Two games are missing for Bensdorp. Who is missing two rounds. But her opponents are only missing one round each. This allows us to identify the round. And we can get the colors too. Bensdorp is 4.4, and so is missing both a White and Black game. This fits, since her opponents are either 4.5 or 5.4, and so are each missing just one game, but with opposite colors.

Life is good, well, as good as can be with two missing games!

<Bensdorp--Lanchava (R3)>
<Schuurman--Bensdorp (R5)>

The actual game results can be deduced from a published xtab, or the leader-board.

This kind of analysis could be programmed into the routine, but I think that would clearly over-complicate the code, and offer little extra.

Ultimately, I would like to find the missing games for any given tournament. But first one must have to be able to find the fact that a tournament has missing games. Which also sounds like finding missing games – ah, the ambiguities of language.

I also plan to write routines to allow easy generation of stub games, in order to track games that are truly missing (a stub being a form of self-documentation). The other advantage of stubs is that, with proper biographical oversight, they can accurately embed game information into the PGN.

That potentially allows SCID, and other software, to know the correct number of games in a tournament, and to be able to construct a correct Swiss xtab (if the colors/rounds can be deduce) or at least a correct RR xtab (if not). And having at least one xtab allows the leader-board to be reconstructed, just from the PGN.

Without Swiss-chess holes in the Swiss, e.g.

Dutch Womens Championships
Leeuwarden NED, 2003.06.25 - 2003.07.05
Average Rating: 2233
                         Rtng  Ti Age Nat    Score       1    2    3    4    5    6    7    8    9   10    Perf Chg
--------------------------------------------------------------------------------------------------------------------
 1: Peng, Zhaoqin        2433  gm  35 NED   9.0 / 10    6w+  2b+  3w+  4w+  5b=  6b+  2w=  3b+  5w+  4b+   2563 +10  (+8 -0 =2)
 2: Lanchava, Tea        2329  im  29 NED   5.5 /  9    5b+  1w- ....  3w+  4b=  5w+  1b=  6w=  4w=  3b=   2311  -3  (+3 -1 =5)
 3: Schuurman, Petra     2293  fm  35 NED   5.0 /  9    4w=  5w+  1b-  2b- ....  4b+  5b+  1w-  6b+  2w=   2282  -2  (+4 -3 =2)
 4: Hamelink, Desiree    2187 wim  22 NED   3.5 / 10    3b=  6w+  5b-  1b-  2w=  3w-  6b-  5w+  2b=  1w-   2132  -7  (+2 -5 =3)
 5: Van Weersel, Arlette 2096 wim  19 NED   3.0 / 10    2w-  3b-  4w+  6b=  1w=  2b-  3w-  4b-  1b-  6w+   2112  +2  (+2 -6 =2)
 6: Bensdorp, Laura      2063 wfm  18 NED   2.0 /  8    1b-  4b- ....  5w= ....  1w-  4w+  2b=  3w-  5b-   2058  +0  (+1 -5 =2)
--------------------------------------------------------------------------------------------------------------------
28 games: +12 =8 -8


Let’s do another tournament, <Biel (2004)>, tid=41663

>>> tfind( "Biel" )

    83126     20   190   Biel Interzonal (1976)
    85469     18   153   Biel Interzonal (1985)
    80051      6    30   Biel (2001)
    39785      6    30   Biel International Chess Festival (2003)

    41663      6    29   37th Biel International Chess Festival (2004)

    41776      6    30   Accentus Turnier Biel (2004)
    47369      6    30   38th Biel Chess Festival (2005)
    47905      6    30   38th Biel Chess Festival (Women) (2005)
    53489      6    29   Biel Int'l Festival (2006)
    53516      6    30   Biel Int'l Festival (Ladies) (2006)
    62373     10    45   Biel Chess Festival (2007)
    65010      6    30   41st Biel International Chess Festival (2008)
    68538      6    30   Biel International Chess Festival (2009)
    72015     10    45   Biel Chess Festival (2010)
    74292      6    30   Biel Chess Festival (2011)
    78291      7    30   Biel Chess Festival (2012)
    78306      8    16   Biel Blitz (2012)
    81935      6    30   Biel (2013)
    84500      6    30   Biel (2014)

>>> mtour( 41663 )

41663 - 37th Biel International Chess Festival (2004)    -- Biel SUI // Biel SUI

Player list (6 players):

   Bacrot, Etienne                29565    5.5
   McShane, Luke J                14287    4.5    6
   Morozevich, Alexander          11719    5.4    6
   Pelletier, Yannick             11676    5.5
   Ponomariov, Ruslan             12109    5.5
   Sasikiran, Krishnan            49247    5.5

N_games =  29

Tournament is assumed to be RR2   (1 missing)

Round counts:

    R1  :  3    R2  :  3    R3  :  3    R4  :  3    R5  :  3
    R6  :  2    R7  :  3    R8  :  3    R9  :  3    R10 :  3

Missing pairings:

 McShane, Luke J    14287  4.5  (6)   //   Morozevich, Alexander    11719  5.4  (6)

I just show a simple substring search I used to get the tid of the tournament of interest. Then run an analysis to find any missing pairings. Which, I do, <McShane–Morozevich (R6)>. It turns out that this game was already missed, as noted in a comment from 2006!

Jun-03-06
Premium Chessgames Member
suenteus po 147: There seems to be a game missing from this tournament. Where is the sixth round game between McShane and Morozevich?

That’s exactly 9 years ago to the day!

zz

Advertisements

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s