Script to classify ICC games as Bullet, Blitz or Standard

Main

Date: 06 Aug 2006 20:27:40
From: Dave (from the UK)
Subject: Script to classify ICC games as Bullet, Blitz or Standard

I've written a small UNIX/Linux shell script that reads from a PGN file
containing one or more chess games and adds to the PGN file either:

[GameTypeOnICC "Bullet"]
[GameTypeOnICC "Blitz"]
[GameTypeOnICC "Standard"]

for each game played. So you can search games in a database based on the
type of game.

The script does this based on the 'TimeControl' header added by ICC in
the PGN file. In other words it extracts the initial time and increment,
then determines whether it is Bullet, Blitz or Standard according to the
formula used on ICC.

So instead of the PGN file having this one line:

[TimeControl "180+1"]

it will now have an additional line:

[GameTypeOnICC "Blitz"]
[TimeControl "180+1"]

It should make it easier split up your games and perhaps search them by
type in a database.

It is a bit slow, doing only 20 lines/s on my 450 MHz machine, so is
perhaps not ideally suited to crunch though a million game PGN file !!

It can be dowloaded from

http://www.drkirkby.co.uk/chess/tools/rating

It reads from a file and prints to standard output.

$ rating mygames.pgn > mygames-with-type.pgn

--
Dave (from the UK)

Please note my email address changes periodically to avoid spam.
It is always of the form: [email protected]
Hitting reply will work for a few months only - later set it manually.

http://witm.sourceforge.net/ (Web based Mathematica front end)

Date: 06 Aug 2006 16:10:46
From: Chris F.A. Johnson
Subject: Re: Script to classify ICC games as Bullet, Blitz or Standard

On 2006-08-06, Dave (from the UK) wrote:
> I've written a small UNIX/Linux shell script that reads from a PGN file
> containing one or more chess games and adds to the PGN file either:
>
> [GameTypeOnICC "Bullet"]
> [GameTypeOnICC "Blitz"]
> [GameTypeOnICC "Standard"]
>
> for each game played. So you can search games in a database based on the
> type of game.
>
> The script does this based on the 'TimeControl' header added by ICC in
> the PGN file. In other words it extracts the initial time and increment,
> then determines whether it is Bullet, Blitz or Standard according to the
> formula used on ICC.
>
> So instead of the PGN file having this one line:
>
> [TimeControl "180+1"]
>
> it will now have an additional line:
>
> [GameTypeOnICC "Blitz"]
> [TimeControl "180+1"]
>
> It should make it easier split up your games and perhaps search them by
> type in a database.
>
> It is a bit slow, doing only 20 lines/s on my 450 MHz machine, so is
> perhaps not ideally suited to crunch though a million game PGN file !!

It's slow because you call an external command, grep, for every
line of the file, and then call more external commands when the
line matches.

The greatest speed-up will come if you replace these lines:

str=`echo $f

Date: 07 Aug 2006 06:55:07
From: Dave (from the UK)
Subject: Re: Script to classify ICC games as Bullet, Blitz or Standard

Chris F.A. Johnson wrote:

> It's slow because you call an external command, grep, for every
> line of the file, and then call more external commands when the
> line matches.
>
> The greatest speed-up will come if you replace these lines:
>
> str=`echo $f

Date: 07 Aug 2006 02:27:32
From: Chris F.A. Johnson
Subject: Re: Script to classify ICC games as Bullet, Blitz or Standard

Much faster than anything else will be to use awk for the whole
script:

awk -F '"' '/TimeControl.*[0-9]+[0-9]/ {
split( $2, t, "+" )
etime = t[1] + t[2] * 40
if ( etime > 900 )
type = "Standard"
else if ( etime > 180 )
type = "Blitz"
else
type = "Bullet"
printf "[GameTypeOnICC \"%s\"]\n", type
}
{ print $0 }' "$@"

--
Chris F.A. Johnson <http://cfaj.freeshell.org >
===================================================================
Author:
Shell Scripting Recipes: A Problem-Solution Approach (2005, Apress)

Date: 07 Aug 2006 12:40:01
From: Dave (from the UK)
Subject: Re: Script to classify ICC games as Bullet, Blitz or Standard

Chris F.A. Johnson wrote:
> Much faster than anything else will be to use awk for the whole
> script:
>
> awk -F '"' '/TimeControl.*[0-9]+[0-9]/ {
> split( $2, t, "+" )
> etime = t[1] + t[2] * 40
> if ( etime > 900 )
> type = "Standard"
> else if ( etime > 180 )
> type = "Blitz"
> else
> type = "Bullet"
> printf "[GameTypeOnICC \"%s\"]\n", type
> }
> { print $0 }' "$@"
>

Yes, that is *much* faster - it went from processing about 50
lines/second to processing about 30,000 lines/second, so is more than
5000x faster.

However, it will not work with the default (/usr/bin/awk) version of awk
on Solaris 10. It will work with /usr/xpg4/bin/awk or nawn or gawk, but
is perhaps a bit less portable than other methods. Sun's awk throws a
wobbler:

% rating 1k.pgn
awk: can't open /TimeControl.*[0-9]+[0-9]/ {
split( $2, t, "+" )
etime = t[1] + t[2] * 40
if ( etime > 900 )
type = "Standard"
else if ( etime > 180 )
type = "Blitz"
else
type = "Bullet"
printf "[GameTypeOnICC \"%s\"]\n", type
}
{ print $0 }

--
Dave (from the UK)

Please note my email address changes periodically to avoid spam.
It is always of the form: [email protected]
Hitting reply will work for a few months only - later set it manually.

http://witm.sourceforge.net/ (Web based Mathematica front end)

Date: 07 Aug 2006 12:48:25
From: Chris F.A. Johnson
Subject: Re: Script to classify ICC games as Bullet, Blitz or Standard

On 2006-08-07, Dave (from the UK) wrote:
> Chris F.A. Johnson wrote:
>> Much faster than anything else will be to use awk for the whole
>> script:
>>
>> awk -F '"' '/TimeControl.*[0-9]+[0-9]/ {
>> split( $2, t, "+" )
>> etime = t[1] + t[2] * 40
>> if ( etime > 900 )
>> type = "Standard"
>> else if ( etime > 180 )
>> type = "Blitz"
>> else
>> type = "Bullet"
>> printf "[GameTypeOnICC \"%s\"]\n", type
>> }
>> { print $0 }' "$@"
>
> Yes, that is *much* faster - it went from processing about 50
> lines/second to processing about 30,000 lines/second, so is more than
> 5000x faster.
>
> However, it will not work with the default (/usr/bin/awk) version of awk
> on Solaris 10.

Solaris is, AFAIK, the only system that still ships the original,
pre-1987 version of awk. But, as you note, it does have more modern
versions. The script will work with all other versions of awk
(including nawk, mawk, gawk, Kernighan's "One True AWK", tawk).

It should also work with the default Sun awk if you change the
beginning to:

awk 'BEGIN { FS = "\"" }
/TimeControl.*[0-9]+[0-9]/ {

> It will work with /usr/xpg4/bin/awk or nawn or gawk, but
> is perhaps a bit less portable than other methods. Sun's awk throws a
> wobbler:
>
>
>% rating 1k.pgn
> awk: can't open /TimeControl.*[0-9]+[0-9]/ {
> split( $2, t, "+" )
> etime = t[1] + t[2] * 40
> if ( etime > 900 )
> type = "Standard"
> else if ( etime > 180 )
> type = "Blitz"
> else
> type = "Bullet"
> printf "[GameTypeOnICC \"%s\"]\n", type
> }
> { print $0 }

--
Chris F.A. Johnson <http://cfaj.freeshell.org >
===================================================================
Author:
Shell Scripting Recipes: A Problem-Solution Approach (2005, Apress)

Date: 07 Aug 2006 14:31:34
From: David Richerby
Subject: Re: Script to classify ICC games as Bullet, Blitz or Standard

[ Cross-post trimmed: this is off-topic in rgc.analysis. ]

Dave (from the UK) <[email protected] > wrote:
> Yes, that is *much* faster - it went from processing about 50
> lines/second to processing about 30,000 lines/second, so is more
> than 5000x faster.

Pssst! 30000/50 == 600. :-)

> However, it will not work with the default (/usr/bin/awk) version of
> awk on Solaris 10. It will work with /usr/xpg4/bin/awk or nawn or
> gawk, but is perhaps a bit less portable than other methods.

Sun value backward-compatibility over standards compliance. The
POSIX-compliant versions of the commands are un /usr/xpg4/bin/ . I'm
tempted to say that it's Sun that's breaking portability here rather
than Chris's awk script!

Dave.

--
David Richerby Electronic Microsoft Ghost (TM): it's
www.chiark.greenend.org.uk/~davidr/ like a haunting spirit that's really
hard to use but it uses electricity!

Date: 07 Aug 2006 15:09:27
From: Dave (from the UK)
Subject: Re: Script to classify ICC games as Bullet, Blitz or Standard

David Richerby wrote:
> [ Cross-post trimmed: this is off-topic in rgc.analysis. ]
>
> Dave (from the UK) <[email protected]> wrote:
>
>>Yes, that is *much* faster - it went from processing about 50
>>lines/second to processing about 30,000 lines/second, so is more
>>than 5000x faster.
>
>
> Pssst! 30000/50 == 600. :-)

Em, seems my maths is not very good!! But in any case it is a very
significant speedup.

I proposed the idea of adding this functionality to pgn-extract. A
couple of hours after sending the email to the author of pgn-extract I
got this reply:

** That should be do-able. I will see what I can manage.

A C version could potentially be even faster. It would make sense to add
this sort of functionality to a program often used to manipulate PGN
files.

>>However, it will not work with the default (/usr/bin/awk) version of
>>awk on Solaris 10. It will work with /usr/xpg4/bin/awk or nawn or
>>gawk, but is perhaps a bit less portable than other methods.
>
>
> Sun value backward-compatibility over standards compliance.

Which is admirable I feel. One of the problems with Linux is that the
goal posts keep moving. I realise POSIX != Linux.

> The
> POSIX-compliant versions of the commands are un /usr/xpg4/bin/ . I'm
> tempted to say that it's Sun that's breaking portability here rather
> than Chris's awk script!

My guess is the script could be modified to work with older versions of
awk. I'm not expert on awk, but if it could be changed so it not reliant
on a modern version of awk, but still works with one, that can only be a
benefit I feel.

Perhaps I'll take a look at the script in more detail and see if I can
work out what it does and what is causing it to not work on Solaris
unless a nonstandard (or standard, depending on your point of view)
version of awk is used.

The same arguments come up about how brain-dead /bin/sh is on Solaris.
But the good thing about writing a script to use /bin/sh on Solaris is
that you can be 99.999% sure it will work on any other UNIX system.

I care quite a bit about portability. One of my programs

http://atlc.sourceforge.net/

has been run on everything from a Sony Playstation games console to a
Cray supercomputer. The old Cray had a portability issue I'd never seen
before, as

sizeof(short) == 8.

The workaround for that had the side benefit of the binary files no
longer having any endianess issues, so they can be moved from big endian
to little endian machines.

--
Dave (from the UK)

Please note my email address changes periodically to avoid spam.
It is always of the form: [email protected]
Hitting reply will work for a few months only - later set it manually.

http://witm.sourceforge.net/ (Web based Mathematica front end)

Date: 07 Aug 2006 16:17:33
From: David Richerby
Subject: Re: Script to classify ICC games as Bullet, Blitz or Standard

Dave (from the UK) <[email protected] > wrote:
> David Richerby wrote:

> A C version could potentially be even faster.

Probably not much faster. Ultimately, I/O is the bottleneck. Perl's
pretty fast as it bytecode-compiles the script before running it; not
sure if modern awks pull the same trick.

> It would make sense to add this sort of functionality to a program
> often used to manipulate PGN files.

Would be a good idea if it did the same for FICS games, too. FICS's
rule is to take the estimated game time to be

(initial time) + (increment) * 80/3

(i.e., assume the game will last forty moves and all of the initial
time plus 2/3 of the increment time will be used). Less than three
minutes is lightning, less than fifteen is blitz, fifteen or more is
standard.

> David Richerby <[email protected]> wrote:
>> Sun value backward-compatibility over standards compliance.
>
> Which is admirable I feel. One of the problems with Linux is that
> the goal posts keep moving. I realise POSIX != Linux.

The goalposts don't move much at all for the standard utilities. It
seems to me that making Solaris users edit all incoming scripts to
invoke the POSIX-compliant utilities in /usr/xpg4/bin causes more
trouble than making people edit their scripts that rely on the old
behaviour to look in /usr/decades-old/bin/ .

> The same arguments come up about how brain-dead /bin/sh is on
> Solaris. But the good thing about writing a script to use /bin/sh
> on Solaris is that you can be 99.999% sure it will work on any other
> UNIX system.

Alternatively, you could write your script in a POSIX-compliant shell,
knowing that 99.999% of Unix systems have a POSIX shell available,
even if it isn't in /usr/bin . The whole point of POSIX is to stop
people having to write their scripts in the intersection of the
various older shells' feature sets.

> The old Cray had a portability issue I'd never seen before, as
>
> sizeof(short) == 8.

64-bit shorts? Wheeeee!

Dave.

--
David Richerby Accelerated Indelible Cheese (TM):
www.chiark.greenend.org.uk/~davidr/ it's like a brick of cheese but
it can't be erased and it's twice
as fast!

Date: 07 Aug 2006 02:12:36
From: Chris F.A. Johnson
Subject: Re: Script to classify ICC games as Bullet, Blitz or Standard

On 2006-08-07, Dave (from the UK) wrote:
> Chris F.A. Johnson wrote:
>
>> It's slow because you call an external command, grep, for every
>> line of the file, and then call more external commands when the
>> line matches.
>>
>> The greatest speed-up will come if you replace these lines:
>>
>> str=`echo $f

Date: 06 Aug 2006 22:00:26
From: Dave (from the UK)
Subject: Re: Script to classify ICC games as Bullet, Blitz or Standard

Chris F.A. Johnson wrote:
> On 2006-08-06, Dave (from the UK) wrote:

>>It is a bit slow, doing only 20 lines/s on my 450 MHz machine, so is
>>perhaps not ideally suited to crunch though a million game PGN file !!
>
>
> It's slow because you call an external command, grep, for every
> line of the file, and then call more external commands when the
> line matches.

Thanks for that. I did realise why it was slow, but did not do anything
about it. Had I thought about it more from the beginning I should have
done it differently.

I guess like many things, I started it off to do a quick job, then
thought I might as well make it available.

But I must admit it is slower than I was expecting. Still, it is still
usable for the number of games any one person will play on ICC in their
lifetime. It is far less usable if you have a huge PGN file of a million
+ games.

> The greatest speed-up will come if you replace these lines:
>
> str=`echo $f

Date: 07 Aug 2006 10:00:09
From: David Richerby
Subject: Re: Script to classify ICC games as Bullet, Blitz or Standard

[ Cross-post trimmed: this is off-topic in rgc.analysis. ]

Dave (from the UK) <[email protected] > wrote:
> Chris F.A. Johnson wrote:
>> It's slow because you call an external command, grep, for every
>> line of the file, and then call more external commands when the
>> line matches.
>
> Thanks for that. I did realise why it was slow, but did not do
> anything about it. Had I thought about it more from the beginning I
> should have done it differently.

If you find yourself writing this kind of thing often, I recommend
using something like Perl or Python.

Chris has given an awk version of your script but I wouldn't
personally recommend awk. It's designed solely for text processing
and, while it can do other things, it gets a bit clunky. The
traditional reason for using awk is speed but the performance
difference is pretty negligible these days and Perl's quite fast
enough. Also, Perl and Python are general-purpose languages.

> Now here I am lost. Not sure what you mean here. I guess once I
> match a line which has TimeControl in it, then I am starting a shell
> for each time awk or sed is started, which is quite often.

You're not starting a shell each time sed or awk is started: you're
starting a copy of sed or awk each time sed or awk is started.

Dave.

--
David Richerby Slimy Book (TM): it's like a romantic
www.chiark.greenend.org.uk/~davidr/ novel but it's covered in goo!

Date: 06 Aug 2006 17:59:57
From: Chris F.A. Johnson
Subject: Re: Script to classify ICC games as Bullet, Blitz or Standard

On 2006-08-06, Dave (from the UK) wrote:
> Chris F.A. Johnson wrote:
>> On 2006-08-06, Dave (from the UK) wrote:
>
>>>It is a bit slow, doing only 20 lines/s on my 450 MHz machine, so is
>>>perhaps not ideally suited to crunch though a million game PGN file !!
>>
>> It's slow because you call an external command, grep, for every
>> line of the file, and then call more external commands when the
>> line matches.
>
>
> Thanks for that. I did realise why it was slow, but did not do anything
> about it. Had I thought about it more from the beginning I should have
> done it differently.
>
> I guess like many things, I started it off to do a quick job, then
> thought I might as well make it available.
>
> But I must admit it is slower than I was expecting.

Even if you use a single call to grep, instead of three, it would
be considerably faster:

str=`echo $f

Date: 07 Aug 2006 01:09:08
From: Dave (from the UK)
Subject: Re: Script to classify ICC games as Bullet, Blitz or Standard

Chris F.A. Johnson wrote:
> On 2006-08-06, Dave (from the UK) wrote:
>
>>Chris F.A. Johnson wrote:
>>
>>>On 2006-08-06, Dave (from the UK) wrote:
>>
>>>>It is a bit slow, doing only 20 lines/s on my 450 MHz machine, so is
>>>>perhaps not ideally suited to crunch though a million game PGN file !!
>>>
>>> It's slow because you call an external command, grep, for every
>>> line of the file, and then call more external commands when the
>>> line matches.
>>
>>
>>Thanks for that. I did realise why it was slow, but did not do anything
>>about it. Had I thought about it more from the beginning I should have
>>done it differently.
>>
>>I guess like many things, I started it off to do a quick job, then
>>thought I might as well make it available.
>>
>>But I must admit it is slower than I was expecting.
>
>
> Even if you use a single call to grep, instead of three, it would
> be considerably faster:
>
> str=`echo $f

Date: 06 Aug 2006 22:53:34
From: Chris F.A. Johnson
Subject: Re: Script to classify ICC games as Bullet, Blitz or Standard

On 2006-08-07, Dave (from the UK) wrote:
> Chris F.A. Johnson wrote:
>> On 2006-08-06, Dave (from the UK) wrote:
>>
>>>Chris F.A. Johnson wrote:
>>>
>>>>On 2006-08-06, Dave (from the UK) wrote:
>>>
>>>>>It is a bit slow, doing only 20 lines/s on my 450 MHz machine, so is
>>>>>perhaps not ideally suited to crunch though a million game PGN file !!
>>>>
>>>> It's slow because you call an external command, grep, for every
>>>> line of the file, and then call more external commands when the
>>>> line matches.
>>>
>>>
>>>Thanks for that. I did realise why it was slow, but did not do anything
>>>about it. Had I thought about it more from the beginning I should have
>>>done it differently.
>>>
>>>I guess like many things, I started it off to do a quick job, then
>>>thought I might as well make it available.
>>>
>>>But I must admit it is slower than I was expecting.
>>
>> Even if you use a single call to grep, instead of three, it would
>> be considerably faster:
>>
>> str=`echo $f

Date: 06 Aug 2006 19:47:46
From: John J.
Subject: Re: Script to classify ICC games as Bullet, Blitz or Standard

Hey Dave, have you been able to connect to the Chessbase server with Pocket
Fritz?

"Dave (from the UK)" <[email protected] >
wrote in message news:[email protected]...
> I've written a small UNIX/Linux shell script that reads from a PGN file
> containing one or more chess games and adds to the PGN file either:
>
> [GameTypeOnICC "Bullet"]
> [GameTypeOnICC "Blitz"]
> [GameTypeOnICC "Standard"]
>
> for each game played. So you can search games in a database based on the
> type of game.
>
> The script does this based on the 'TimeControl' header added by ICC in the
> PGN file. In other words it extracts the initial time and increment, then
> determines whether it is Bullet, Blitz or Standard according to the
> formula used on ICC.
>
> So instead of the PGN file having this one line:
>
> [TimeControl "180+1"]
>
> it will now have an additional line:
>
> [GameTypeOnICC "Blitz"]
> [TimeControl "180+1"]
>
> It should make it easier split up your games and perhaps search them by
> type in a database.
>
> It is a bit slow, doing only 20 lines/s on my 450 MHz machine, so is
> perhaps not ideally suited to crunch though a million game PGN file !!
>
> It can be dowloaded from
>
> http://www.drkirkby.co.uk/chess/tools/rating
>
> It reads from a file and prints to standard output.
>
> $ rating mygames.pgn > mygames-with-type.pgn
>
> --
> Dave (from the UK)
>
> Please note my email address changes periodically to avoid spam.
> It is always of the form: [email protected]
> Hitting reply will work for a few months only - later set it manually.
>
> http://witm.sourceforge.net/ (Web based Mathematica front end)

Date: 06 Aug 2006 22:32:45
From: Dave (from the UK)
Subject: Re: Script to classify ICC games as Bullet, Blitz or Standard

John J. wrote:
> Hey Dave, have you been able to connect to the Chessbase server with Pocket
> Fritz?

I found the problem. It must be a subscription issue, as I changed the
date on my Pocket PC back to the year 2003 and it works!!!

Now is it too much to ask for a programmer to put an error message like

"Sorry, your subscription to use the database has lapsed. Please ..."

rather than the very confusing

"Position not found in the online database!"

Is it me, or their error message a bit confusing?

I'll see if I can get an update - I don't want to have to have my
machines date wrong by 3 years.

--
Dave (from the UK)

Please note my email address changes periodically to avoid spam.
It is always of the form: [email protected]
Hitting reply will work for a few months only - later set it manually.

http://witm.sourceforge.net/ (Web based Mathematica front end)

Date: 06 Aug 2006 22:20:03
From: Dave (from the UK)
Subject: Re: Script to classify ICC games as Bullet, Blitz or Standard

John J. wrote:
> Hey Dave, have you been able to connect to the Chessbase server with Pocket
> Fritz?

No. I am still stuck. I've got no idea what the problem can be. Someone
(was it you) said you need to pay a subscription to use it after 1 year.
But I've no idea how I could enter any code into PocketFritz to indicate
I've paid and I'm sure it was working more than a year after I got the
program. So I'm a bit puzzled.

In any case, if it is a subscription issue (which chessbase have failed
to bother answering), the error message could be a bit more useful than

"Position not found in the online database!"

It does not exactly tell you anything useful, as clearly the position
after the move 1. e4 will be in loads of the games.

I've looked at the firewall logs and keep seeing an outgoing connection
on port 80 (http) to the IP address 80.237.188.72. I can't verify if
that is Chessbases, but I can't see any packets coming back.

Yours, puzzled and a bit disappointed in chessbase's attitude to the
software, which seems to be to no longer develop it and give no support.

--
Dave (from the UK)

Please note my email address changes periodically to avoid spam.
It is always of the form: [email protected]
Hitting reply will work for a few months only - later set it manually.

http://witm.sourceforge.net/ (Web based Mathematica front end)

Date: 06 Aug 2006 23:45:22
From: John J.
Subject: Re: Script to classify ICC games as Bullet, Blitz or Standard

Do you get the message when you are searching for statistics or for games?

The statistics search hasn't worked for me for many months. I wonder of they
discontinued it.

I simply plugged the IP address you posted into my web browser and Chessbase
came up. So yes, it's a chessbase address.

"Dave (from the UK)" <[email protected] >
wrote in message news:[email protected]...
> John J. wrote:
>> Hey Dave, have you been able to connect to the Chessbase server with
>> Pocket Fritz?
>
> No. I am still stuck. I've got no idea what the problem can be. Someone
> (was it you) said you need to pay a subscription to use it after 1 year.
> But I've no idea how I could enter any code into PocketFritz to indicate
> I've paid and I'm sure it was working more than a year after I got the
> program. So I'm a bit puzzled.
>
> In any case, if it is a subscription issue (which chessbase have failed to
> bother answering), the error message could be a bit more useful than
>
>
> "Position not found in the online database!"
>
> It does not exactly tell you anything useful, as clearly the position
> after the move 1. e4 will be in loads of the games.
>
> I've looked at the firewall logs and keep seeing an outgoing connection on
> port 80 (http) to the IP address 80.237.188.72. I can't verify if that is
> Chessbases, but I can't see any packets coming back.
>
> Yours, puzzled and a bit disappointed in chessbase's attitude to the
> software, which seems to be to no longer develop it and give no support.
>
> --
> Dave (from the UK)
>
> Please note my email address changes periodically to avoid spam.
> It is always of the form: [email protected]
> Hitting reply will work for a few months only - later set it manually.
>
> http://witm.sourceforge.net/ (Web based Mathematica front end)

Date: 07 Aug 2006 01:36:38
From: Dave (from the UK)
Subject: Re: Script to classify ICC games as Bullet, Blitz or Standard

John J. wrote:
> Do you get the message when you are searching for statistics or for games?

It was the statistics. I never used it to looks for games (much anyway),
but that seems to work.

But a couple of hours ago I did get the statistics working, but it is
unstable. I reset the date on my PocketPC to 3 years ago, to see if it
was date related. At first it seemed to have cured the problem, but now
it has not. I think something is not quite stable at their end.

> The statistics search hasn't worked for me for many months. I wonder of they
> discontinued it.

I wonder if they know they have discontinued it? I suspect it is a
programming error on their web server.

One possibilty is that someoen has made some sort of tool that extracts
data from their web site. So they changed the format slightly so their
tool would not work. But it also breaks their own software.

Just a thought!

The service pack 2 fixed a problem with the online database. I'm
wondering if they change the format slightly now and again, to stop
someone writing their own appliation to interface to the chessbase server.

> I simply plugged the IP address you posted into my web browser and Chessbase
> came up. So yes, it's a chessbase address.

Now why did I not do that? I used nslookup and it was unable to get me
any useful information.

--
Dave (from the UK)

Please note my email address changes periodically to avoid spam.
It is always of the form: [email protected]
Hitting reply will work for a few months only - later set it manually.

http://witm.sourceforge.net/ (Web based Mathematica front end)