Author: Dann Corbit
Date: 10:19:10 07/18/01
Go up one level in this thread
On July 18, 2001 at 06:03:30, CLiebert wrote: >Posted by Dann Corbit (Profile) on July 17, 2001 at 04:15:06: > >In Reply to: Re: does chessbase care about wb engines posted by CLiebert on July >17, 2001 at 04:02:44: > > >On July 17, 2001 at 04:02:44, CLiebert wrote: > >>On July 16, 2001 at 17:30:43, Dann Corbit wrote: >> >>>On July 16, 2001 at 06:35:07, CLiebert wrote: >>> >>>>On July 16, 2001 at 02:26:53, ERIQ wrote: >>>> >>>>>maybe the problem w/ chessbase interface and wb engines is that, bigbrother does >>>>>not care about them performing at their best. >>>>> >>>>>Why should they ?! wb engines are free, they don't make money from them but they >>>>>compete strongly w/ fritz,nimzo, etc. their *bread and butter* products. >>>>> >>>>>So if wb-engines somehow get dumbed down alittle great. I guess that justifys >>>>>the price of "pro" engines. >>>>> >>>>> sign, >>>>> Eriq >>>> >>>> >>>>First: Why do you think CB put efforts in developing the adapter? >>> >>>To make lots and lots of money, I would suppose. That's generally the purpose >>>of writing code for commercial endeavors. Sometimes, they might do something >>>just to be nice. Is that why CB wrote their adaptor? >>> >>>>Second: there are a lots of engines showing no difference in playing strenght. >>>>If programmers a willing and able to optmize their engines for fritz they could >>>>do this. Borgstädt did it for Goliath, Kai Skibbe for gromit, other examples are >>>>Anmon, Faile or TCB. The Natives and wb-adapter-versions of these engines are >>>>quite at the same level in pratice, did you try one of them? >>> >>>Let's see... >>>5/(100+) seems to be a rather small ratio. >>>I am curious to know why the existing, well-debugged Winboard protocol was not >>>adopted as-is. And even more interesting would be to know why CB protocol must >>>send resets every so often. Is there some purpose to this? This defect has >>>been known for many years. Finally it was fixed recently and then immediately >>>re-broken. Seems a bit odd on the surface. >>> >>>>But if you like to hold on you big brother theories, feel free ... >>> >>>I don't know if it is sinister or not, but I am _personally_ convinced that the >>>defects in the CB version of the Winboard adapter are purposeful, just as the >>>defects in the RS232 adapter were also purposeful. I am not sure if I can >>>actually blame a company for trying to make their products look good at the >>>expense of competitors, especially if the competition is free. >> >>You´re joking?! >> >>Why do you think CB gave the native-code to three of the best wb-programmers >>(gromit, sos and goliath) before if they fear the competition? > >It's called "slave labor" >1. I give you this specification. > >If you would like... > >2. You code to that specification. > >Of course. To the native concept, to uci and / or winboard... > >3. The output will be useless to anyone except for us. > >Wrong! >If people could use an engine with one of the most popular GUIs it is useful >for chessbase (of course), for the engine-programmer and for the users. >I believe Gandalf would be happy getting the native code and rights. >Why do think people were so enthusiastic getting The King working under fritz? >They could use the GUIs of chessmaster or winboard, couldn´t they? The engines that are produced run only under ChessBase. It is true that people like to be able to run all of their favorite engines in one place. I certainly don't disagree with that. I am glad that ChessMaster has taken an open stance and allowed their engine to run under multiple GUI's. Yet another reason that the ChessMaster software will continue to dominate. They can figure out what people want and then try to do it. >As a result of your work, other people will be able to use your engines with CB. > > >And the other way round! How do I run a CB engine under Winboard? >>For what reason should cb fear the competition of wb-engines? > >It isn't easy to compete with "free" -- you have to have significant value >added. I think that CB does add significant value, but it is hard to >communicate. > >Of course it is an significant value for the freak-scene, but not for the mass >market. Chessmaster was (is?) the most sold chessprogramm worldwide with only >one engine, so why do think it is so important from the commercial point of view >to support wb-engines (beside 20 natives!) ? Because once people figure it out, they will want to do it. Because the mass market is completely gobbled up by ChessMaster and nobody else has any chance to get even a fraction of their display space in the mass market. Therefore, they should do their best to appeal to chess freaks. Because the ChessBase products are quite a bit more expensive than ChessMaster (which can be had for well under $20!) and therefore must add significant value to the purchase. Becase a chess tool vendor must anticipate future trends. Do you imagine that the 100+ Winboard engines are not going to eventually capture some attention? >People see SSDF results and make purchase decisions based on that. > People see WMCCC results and make purchase decisions based on that. It's silly >-- they ought to look at useful features. But it is obvious that they don't >really know what they want or need very well. > > >That might be right or not. A lots of people buying a product watching the cover >in the software-market. If there is anything like "Worldchampion" or >"Rankinglist No1" they might opt for this one. What else should they do, not >everbody is an insider. >Making research to find out that fritz or shredder have the best GUI before >buying a "simple computer game"? Behaving like this is not the majority, I >guess. I think you are right about that. There are (however) an educated crowd that looks at features. A real database system like that offered by ChessBase and Chess Assistant is really light years beyond the competition for serious chess players needs. However, it is hard to get that information to the chess playing public. And it is even harder to show people how they might benefit without them actually seeing the product in use. They walk into a specialty store and see a dozen chess programs. They all have nice, shiny boxes. All of them say that they are the best. Most people will just buy the cheapest one. > >>What do think happens if a wb-engines reaches a tiger-or-fritz-level? > >They already have. > > >Do you really believe that any wb-engine can compete with tiger/fritz on the >long run?! I believe that the difference is so miniscule that in play or coaching 99.9999% of the people who buy chess programs could not tell if they were playing against crafty or fritz or chess tiger unless someone told them. All they would know is that the chess engine clobbers them or that the chess engine can suggest much better moves than they can come up with by themselves. Probably, Chess Tiger and Junior and Fritz are 100 ELO above Yace and Crafty. I am guessing that half of that is opening book, but maybe that is an exaggeration. Anyway, that difference is inconsequential to the one who purchases the program. I also think that crafty could be configured on a 64 CPU machine in such a way that it would dominate any other chess engine on the planet. Of course, that has no commercial significance. >Show me a win of an amateur-engine over a long distance match against >tiger, level doesn´t matter. I am shure you can´t! I can't. But I can show you a lot of wins by crafty against any top engine. And Yace. >>Do you think you will get it for free? >>Yes? ;-) > >Crafty and Yace pretty well match that description. > > >I disagree (did you ever try Tiger?). Yes. I have the commercial engines: Chess Tiger Zarkhov Rebel Hiarcs and I like them all. I also have Bookup and Chess Assistant. I don't have ChessBase because the ChessBase programs are unable to process EPD records. I do recognize the general excellence of the ChessBase programs. I especially admire Junior and Fritz as excellent work, and I like some of the features of the ChessBase database. However, for my needs, Chess Assistant is much better. >Of course they are quite good and I like them very much (as some others too), > but it is still a long way to compete the top! >The latest Yace seems to be much stronger, thats a great progress, yes! >Same with gromit 382 or GL2. Crafty remains for nearly a year at the same level. 18.10 is a big jump forward, I think. >Tiger made a great jump (away from the amateurs!) with its last version, >and fritz7 will go further again, so am shure that the > difference between amateurs and profs is getting bigger not smaller in the >near future! It will stay about 100 ELO for all eternity, I think. >>You get Crafty for free. As native. Another one of the strongest >>winboard-engines you get for free too. As native! >>A few days ago CB agreed to publish the GL 1.5 engine as fritz-native. >> >>You will know how it suits to these bigbrother-theories, I am sure >>(and won´t discuss this stuff anymore). > >This is all very interesting, but you didn't bother to answer any of my >questions. > >It is not very interesting for endusers to go into technical details. >It might be interesting in theory and for programmers, of course! >For endusers it makes no difference if it works "good" or "perfect", >if playing works without remarkable problems. Why don't you fix the bug? Just comment out the line that sends the reset. It adds no value and causes problems for the free engines. Better yet, just dump your whole spec and use the Winboard specification. You're falling behind anyway, because there is a new Winboard specification with a lot of improvements. >Whether an wb-engine plays with 97% or 100% performance, with 30 Elo more or >less under fritz is not important for the majority. Sending resets is a 50% performance loss, since it clears the hash tables. >If somebody would like to play with the best engines he wouldn´t opt for >wb-engines. You are wrong, mathematically. 1. I am somebody. 2. I like to play with the best engines 3. I opt for Winboard engines and enjoy them very much q.e.d. >Different style is the most important thing of the wb-engines for me, not 20 or >30 elo more or less under different GUIs, you can´t feel it anyway. Different style is important. There are (for instance) positions that only Phalanx can solve. But when you harm the performance of the engines to the tune of 50% for no reason, it looks like fear. Maybe it isn't but it sure smells bad. >>>It is well known (and incredibly obvious) that sending a reset command during >>>play will not make for optimium performance. Can anyone provide a logical >>>explanation as to why this command is still sent by these tools? Has there been insufficient time to remove this clear and obvious defect >>>? >> >>Where is the point in practice if you compare GL and LG/Winboard or other >>engines I mentioned below wihout any difference after hundreds of games in >>practice? > >If someone spends a great deal of energy working around the bugs in the >protocol, they can produce an engine which has value only to an owner of CB. > >You ´re talking about things you don´t know. >I know Borgstädt for years and can tell you that he did not spend >that "great deal of energy" in working around the protocol what you think! >He did it "by the way" and it wasn´t very difficult for him. >Ok, may be he is in a way like a professional, for real amateurs it >might not be the best solution. You have to write code to work around the bugs in the protocol. I am sure it took him at least 3-4 days and probably more like a week. At the cost of a professional programmer, that is a good deal of money. >So far I would agree and I would be very happy too, if Cb will improve this. I think you will gain a lot of goodwill by fixing it. >But I can also live with the present solution. Add EPD record output, while you are fixing the autoplayer and CB interface issues and then I could live with it too.
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.