Author: Komputer Korner
Date: 22:50:26 11/29/97
The definition of a transposition:. In the Webster dictionary, it means " to change the usual or relative order or position of". NOTE: nowhere in the definition does it say "link". A transposition is more than a link. Mats Winther is treating all transpositions as if they were only links. That is why he thinks that only forward transpositions count because his beloved TascBase shows only forward transpositions. Forward transpositions are shown with definite links and that is what the key tree screen in TascBase does. It is an amazing concept but it doesn't solve the complete problem of transpositions. Do you think that the Fritz 5 and the Chess Assistant programmers are idiots for including backward transpositions in their trees? If we had the time to study all the opening lines of every possible variation in the whole openings tree and if we remembered every one, then the links that Tascbase does would be enough. However we don't have that time and even the top GMs sometimes get caught in a transposition. How would that be possible if they had studied every line? The answer is even they don't have the time. So they study what they know a little deeper. Backward transpositions are extremely important to them for the following reason. All you have to do is follow the path backwards and you will see the possible paths to the position. You can't do that in many situations in the keytree . I used the word MANY and not ALL because the situation is more complicated than that. See my separate response to Mats where I give examples of cases of the English opening leading to the same classical Tarrasch position from 5 separate opening lines. When you are at that position to find some backward transpositions (Note that I said some and NOT all) , you have to move back in the tree clicking CTRL-T each time to show any forward transposition at each point. Eventually if you go back you will discover some of the paths. However, because the key tree database does not actually store the move that transposes to another line, it is impossible to show all the backwards transpositions . A simple example to explain: Take the 2 lines that Mats gave us trying to show that he was right. He didn't actually show anything but here is what you see in the tree. I am going from memory here as to the lines that Mats gave us but if the following 2 lines are in different move orders that does not matter because there is still a transposition at the end. 1st line is 1.e4e5 2.Nf3Nc6 3.Nc3Nf6 4.Bb5 2nd line is 1.e4Nf6 2.Nc3Nc6 3.Bb5e5 and now Tascbase will not let you enter 4.Nf3 into the key tree even though it shows it as a transposition. This is not a bug as that is the way that the key tree is designed. It automatically takes you to the transposed variation position that is already in the tree as the 1st line. If you examine the 1st line in the tree you will see as you go backwards in it and check for transpositions at each point, you will see that there are none shown. That is because you cannot directly transpose to the 2nd line from the 1st line. You can only go the other way around from the 2nd to the 1st. However that is looking at transpositions as if they were only links. Don't you think that it is important to know that the 2nd line will lead to the first even if you hadn't specifically examined the 2nd line and didn't know that it existed as a transposition. If the line is in the key tree and you have 1000's of lines in there you may not specifically look at this line. Then if you decide to play 1...Nf6 on a whim over the board or if you had studied 1...Nf6 a little but not extensively looked at every possible line, you could end up exactly in the above position at the end of line 1. If you had beforehand decided that you never wanted to play that line 1 end position as black, you would have been tricked. Mats will argue that he will beforehand carefully study every line in his repertoire key tree before deciding what to play. That is all well and good but the fact remains that Tascbase gives you the power to import into the tree thousands and thousands of games so that you will build up your repertoire. And in fact unless you play an extremely narrow repertoire, there will be 1000's of games to look at in your repertoire. Nobody has the time to look at them all. Nobody has the time to look at all the variations in one part of your repertoire. Therefore it would be a good idea to know beforehand what are the possible backward transpositions in the positions that you have studied so that you can be prevented from being tricked. Mats will argue that he will make sure that he only puts the lines in his key tree that he has time to study and that is he will look at every one. Therefore he won't be able to import all the relevant games for his repertoire because there will be too many. Fine, then he is using his tree like a limited repertoire keeper and in that respect Tascbase is superior to Bookup because Bookup does not point out any transpositions at all. But using his tree this way is fraught with difficulties because then Mats can't be sure of catching all the transpositions of the lines he knows best because he has a very limited number of lines to look at. However if Mats imports 1000's of games so that more transpositions will show up, he doesn't have the time to look at them all anyway. So Mats is caught in a conundrum. He is damned if he does and damned if he doesn't. The last way is the best of a bad choice, but Chess Assistant and Fritz 5 users don't have to worry about this. All they have to do is import lots of games and then look at positions they know. They will be assured that every transposition (backwards and forwards) will be caught and especially they will be able to discover in the above example that there is a transposition even if they only examine line 1. Of course with Fritz 5 or Chess Assistant they can look at either line to catch the transposition. . With Tascbase you either have to investigate both or either make sure that you investigate the 2nd line. If however you don't know that the 2nd line is in your database, then you are out of luck unless you look at every last line in your database. So readers, I put it to you. Which opening book editor would you rather have?. One that may cause you to miss the transposition or one that will show you any possible backward or forward transposition that the tree contains. If Mats doesn't understand this, then I give up on him. TascBase is a great program but it is NOT the world's best opening book editor and My NAKED TRUTH article stands as written. You all have seen the amount of invective and garbage that Mats has thrown my way. I have tried to restrain myself in my responses to him, but I must admit that I lost my temper too. Please forgive me for any insults I may have hurled his way, but if you have read all of his posts against me you will have realized that he was VERY OBJECTIONABLE. Mats if you understand what I have read here, perhaps we can get on with discussing what TascBase does well, since I have said many times it is a great program and you are probably the most knowledgeable user of it. Don't forget that I had listed it for over a year at the Gambit Soft site as the best all around program for the average user until FRITZ 5 came along.
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.