Author: Albert Silver
Date: 11:29:28 11/28/99
Go up one level in this thread
On November 28, 1999 at 11:00:04, Robert Pawlak wrote:
>On November 28, 1999 at 10:08:53, Albert Silver wrote:
>
>>On November 27, 1999 at 18:32:40, Charles Unruh wrote:
<cut>
>>>Yes, once you use CA, you get used to the speed. The tree is also much smaller -
>the entire tree for their 1 million + game DB fits easily on a CD. This means
>you don't have to keep switching amoung multiple trees when working. These two
>features are the main selling points for me.
>
Yes, once you get used to the speed, CB really can cause one to want to tear
one's hair out. I once was getting ready for a tournament and thought of fitting
a CA tree on one CD, and a CB tree on another. The CA tree, from a 1 million +
game DB took about 30-40 minutes to do, then I came around to CB: the first
thing it told me was that I didn't have enough hard disk space. I was surprised,
and checked only to see I had about 1.5 Gigs. Astonished, I split the DB (about
1.1 million games) in two and had it start. 10 hours later it was still
going.... I kid you not. Both of these were done on a K6/233 + IDE HD at the
time.
>>4 elements
>>that really tilted the balance in CB's favor:
>>
>>1) Better board. If you are going to spend hours and hours in front of a board,
>>then it has to be pleasant looking. Frankly, I didn't think (still don't) much
>>of it, meaning that given a choice between studying the same position in CA or
>>CB, I'll choose CB.
>>
>
>This is certainly a matter of personal preference. I think they were trying to
>maintain the same look and piece design as the old DOS CA.
Of course, but it is annoying enough to be a problem in my case. I suspect it
has to do with two things:
1) The type face I am used to seeing in chess literature tends to look much more
like that of CB hence the matter of familiarity.
2) The lines that determine the shape of the pieces are very thin so that the
pieces don't stand out quite as clearly.
Naturally, this may all be just me, but I believe that where one person
complains about it, there are ten other silent miffed ones as well.
>
>>2) Engines. This also killed CA. The equivalent of trying to run in a race while
>>limping. CA only had Dragon (if you never heard of it, you haven't missed
>>anything) and Zarkov. Zarkov was solid enough but still wasn't up to the Fritz
>>or Hiarcs that CB was offering. True, CA allowed one to call up external
>>programs such as Rebel, Mchess, or others, but the time of loading and exiting
>>the program was terrible.
>>
>
>I'd have to agree here. Dragon was ok for blunder checking, but was not in the
>same league as Hiarcs or Fritz.
>
>I use Rebel for analysis, and the load time is not long at all. But you need to
>set up the proper hash table size, because the default can cause the problem in
>load time that you've mentioned.
>
Perhaps this was a memory problem. At the time I used 32 Mb, and though Rebel
would indeed load fairly quickly, the minute I had it start analyzing, the HD
would go crazy for about a minute before Rebel even gave the results for its
first ply.
>>3) Better database. The core base wasn't much better (in fact Big99 or Mega99
>>users might want to check out the dozens of erroneously players rated over 2700
>>and 2800 - poor Gary), and CA did a very good job, but CB's tens of thousands of
>>GM commented games meant that either CA do likewise or allow one to access .cbh
>>files in it. In CA4 they opted for the latter. A wise decision IMO, but buggy to
>>say the least, and as a consequence pretty much unusable. Even plain unannotated
>>games would come out with problems. And fatal errors would appear if I tried to
>>simply convert a large .cbh database into CA's native format.
>>
>
>Yes, there are some bugs in the cbh translation, but I think it was largely an
>'undocumented feature' - I mean I'm not sure any claims were made about it. The
>cbf translation was pretty good though. This is mostly what I use.
I think it should be possible to sort them out.
>
>>4) Multimedia. Here CB really innovated and by allowing the program to handle
>>.html files were able to imbed photos and even videos into textual presentation,
>>allowing them to produce one of the wonders of the chess world: Chessbase
>>Magazine. I was skeptical at first as to the advantage this offered, but after
>>having seen Kasparov throw up his hands in despair after losing material to
>>Anand, and then Anand's enthusiatic retelling of this, suddenly these players
>>became more than stills in a magazine and names on a rating list. It was
>>awesome. I also bought CB's superb Alekhine CD, somewhat nervous at seeing
>>everything in German as the blurb stated, but was relieved to see that the
>>extensive videos of Shirov, Gelfand, and King commenting on Alekhine were all in
>>English. The many old pictures of Alekhine and his opponents in tournaments of
>>yesteryear were also a great pleasure. Here, CA had nothing to offer but to be
>>honest, if this had been the only issue, I could easily have done without
>>(though not my subscription to CBM).
>>
>
>Yes, there is no contest here. Although I find the multimedia stuff to be a bit
>of a curiosity. CBM is really nice, and eye-popping too. I think the main thing
>that I like about the CB training is that you have the timed questions, which is
>something that CA could add, I think.
I don't see why CA could not handle all aspects (except for the multimedia) of
reading/converting .cbh files perfectly. Aside from the pawn-structure symbol,
which causes a small window representing the pawn-structure to appear in
commented CB games, there isn't anything CA doesn't have as well:
- colored arrows and squares in order to highlight moves or squares.
- timed moves... Actually, CA doesn't actually have it, but it is amply used in
all of their training products, so I imagine this would be a simple thing to
include.
- Notation symbols describing the moves and positions.
- Themes: strategical, tactical, etc...
<cut>
>Yes, I am very excited also. For me and others, the engine was a major point o
>be addressed. I got by with using Rebel (which has an awesome playing style),
>but since I've switched to NT, I have to boot up my old Win98 partition when I
>want to use it.
>
>Bob P.
The only thing I still regret is the lack of follow-up on the other programs
such as Strategy, Studies 2.0, etc... These had PGN notation problems which were
eventually sorted out by CA4. These also didn't allow one to use any other
engine than Dragon, despite having Zarkov 4.2 (used by CA3), or the Craftys also
used in CA4. This detracted a lot from them as part of the whole idea was to be
able to train with specific positions against the computer, only, in this case,
it was literally impossible to play against any opponent other than Dragon, or
an external program (no later than Rebel 9). I had wanted to play against Fritz
and Hiarcs (CB) but the PGN notation problem made this impossible. The board in
Strategy was already very limited as it looked absolutely tiny on my screen in
1k x 768 resolution. Again, a small patch to correct this should have been
possible.
Albert Silver
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.