In 2001 when I purchased my first Fritz/Chessbase program (based on Deep Fritz 6), I ran it vs the Saitek Travel Champion (almost 100% identical specs to the Saitek Cosmos apart from a bigger opening book and subtle things; it should be almost identical in strength too). I gave the Deep Fritz 6 five minutes per game (Fritz 6 is a very strong GM program) and I gave the Saitek Travel Champion the professional time control (2 hours for the first 40 moves + 1 hour for the rest of the game). Well, the Saitek held its own. The game was a draw. Back then I was so impressed with Deep Fritz 6. It calculated more than 14 plies ahead in a classical game. *What's more, it calculated 9 plies ahead in one second! I thought it was a real beast.* Nowadays Fritz 6 looks pathetic vs 3700 Elo engines that calculate more than 30 plies ahead. But then again, even Magnus would look pathetic vs a modern engine.
I still have Fritz 3, Junior 5, Hiarcs 732. Man I used to do knockout bullet tournaments years ago in Chessbase and Hiarcs 732 came out and beat Fritz and I was so excited. I had a collection of 64 engines and I would run knockout after knockout. It was one of my main hobbies.
@@Gadgetify Unfortunately there are compatibility problems nowadays. I ran some emulators on my PC but it is just not how things worked back in the day. I mean I didn't like the emulators of old chess engines at all. Still, I played vs them a few games. Luckily, I still can run my old Fritz 6 software on my current PC (no emulators) but not Fritz 3 or Fritz 4 engines. They are no longer "natively" compatible with the Fritz/Chessbase. I guess If I want a 2300-2500 native engine (not an engine that simulates mistakes), I will need a King Performance. Also, there are hundreds of modern "weak" engines but I don't like them: their options are scarce; their creators give ridiculous Elo numbers. What's more, CCRL ratings are meaningless too. They have always been, but today they are 100% nonsense. I was stupid back in the day to trust chess magazines that published results of games between chess computers, and games of computers vs humans. Many thousands of games, precise ratings, etc. It was all business/marketing tricks and it still is. It's all for hype. I shouldn't have trusted them one iota with all their nonsense ratings. I wonder how many of those games were rigged or artificially manipulated. People are not to be trusted, especially journalists and *especially chess players.* The first chess computer I bought could not play chess at all (some Saitek model) -- it was too weak despite the claims of 1700+ Elo. I returned that junk and I bought the Saitek 2100 Travel Champion which was a superb product.
It is supposed to be around 2100 in rating. Having said that, I need to give it more time. I suppose if you give it a couple of minutes per move, it plays at that max level.
@@Gadgetify Yeah, you are absolutely correct. It's more like 2100+ FIDE (Wikipedia is wrong here). The problem with modern engines is they are too strong (3000+) and they need some time to think (to simulate mistakes) to play badly. Yet, say, 2500 is not really a bad play as far as I'm concerned! By the way, the quality of play (both for humans and engines) goes down a lot when it's not 3 hours per game. For humans in 5 min blitz, the quality of moves is like 600 Elo points lower vs the quality of moves in a 3 hour game. For an engine, it is similar but it depends on the era the engine is from. To be more specific, I can illustrate it by saying, in a long game I won't miss a 2-3 (4-6 plies) move combination but in a blitz it can easily happen. In a long game I often have to calculate 4-5 moves (8-10 plies) ahead and sometimes even 12-14 plies ahead. I also have to recheck my calculations. At some crucial moment in a complicated position, I may spend 20-30 minutes before I proceed (it may decide the outcome of the game). In blitz and bullet I rarely double check any lines and I have to cut off my calculations as early as possible because time is ticking: I have to pace myself to spend 5 seconds per move on average and only in crucial moments I can afford to burn like 30 seconds on a single move. In a one minute bullet game I have to pace myself to spend only one second on my calculations on average. There's not much I can calculate in one second and I also have to calculate for my opponent ideas or I miss some combination. PS: the computer has to have around 3-4 minutes per move or 40 moves per 2 hours to play at its best. This is the control for classical games. Also the permanent brain should be on, i.e. the computer thinks while the opponent thinks. BTW, back in the day chess players used to leave the computer running overnight. That was for preparation in openings. Giving so much time was necessary so that the machine would find good moves and produce good recommendations in openings.
In 2001 when I purchased my first Fritz/Chessbase program (based on Deep Fritz 6), I ran it vs the Saitek Travel Champion (almost 100% identical specs to the Saitek Cosmos apart from a bigger opening book and subtle things; it should be almost identical in strength too). I gave the Deep Fritz 6 five minutes per game (Fritz 6 is a very strong GM program) and I gave the Saitek Travel Champion the professional time control (2 hours for the first 40 moves + 1 hour for the rest of the game). Well, the Saitek held its own. The game was a draw. Back then I was so impressed with Deep Fritz 6. It calculated more than 14 plies ahead in a classical game. *What's more, it calculated 9 plies ahead in one second! I thought it was a real beast.* Nowadays Fritz 6 looks pathetic vs 3700 Elo engines that calculate more than 30 plies ahead. But then again, even Magnus would look pathetic vs a modern engine.
I still have Fritz 3, Junior 5, Hiarcs 732. Man I used to do knockout bullet tournaments years ago in Chessbase and Hiarcs 732 came out and beat Fritz and I was so excited. I had a collection of 64 engines and I would run knockout after knockout. It was one of my main hobbies.
@@Gadgetify Unfortunately there are compatibility problems nowadays. I ran some emulators on my PC but it is just not how things worked back in the day. I mean I didn't like the emulators of old chess engines at all. Still, I played vs them a few games. Luckily, I still can run my old Fritz 6 software on my current PC (no emulators) but not Fritz 3 or Fritz 4 engines. They are no longer "natively" compatible with the Fritz/Chessbase. I guess If I want a 2300-2500 native engine (not an engine that simulates mistakes), I will need a King Performance.
Also, there are hundreds of modern "weak" engines but I don't like them: their options are scarce; their creators give ridiculous Elo numbers. What's more, CCRL ratings are meaningless too. They have always been, but today they are 100% nonsense. I was stupid back in the day to trust chess magazines that published results of games between chess computers, and games of computers vs humans. Many thousands of games, precise ratings, etc. It was all business/marketing tricks and it still is. It's all for hype. I shouldn't have trusted them one iota with all their nonsense ratings. I wonder how many of those games were rigged or artificially manipulated. People are not to be trusted, especially journalists and *especially chess players.* The first chess computer I bought could not play chess at all (some Saitek model) -- it was too weak despite the claims of 1700+ Elo. I returned that junk and I bought the Saitek 2100 Travel Champion which was a superb product.
The retro one played ok for a while
It is supposed to be around 2100 in rating. Having said that, I need to give it more time. I suppose if you give it a couple of minutes per move, it plays at that max level.
@@Gadgetify Yeah, you are absolutely correct. It's more like 2100+ FIDE (Wikipedia is wrong here). The problem with modern engines is they are too strong (3000+) and they need some time to think (to simulate mistakes) to play badly. Yet, say, 2500 is not really a bad play as far as I'm concerned!
By the way, the quality of play (both for humans and engines) goes down a lot when it's not 3 hours per game. For humans in 5 min blitz, the quality of moves is like 600 Elo points lower vs the quality of moves in a 3 hour game. For an engine, it is similar but it depends on the era the engine is from. To be more specific, I can illustrate it by saying, in a long game I won't miss a 2-3 (4-6 plies) move combination but in a blitz it can easily happen. In a long game I often have to calculate 4-5 moves (8-10 plies) ahead and sometimes even 12-14 plies ahead. I also have to recheck my calculations. At some crucial moment in a complicated position, I may spend 20-30 minutes before I proceed (it may decide the outcome of the game). In blitz and bullet I rarely double check any lines and I have to cut off my calculations as early as possible because time is ticking: I have to pace myself to spend 5 seconds per move on average and only in crucial moments I can afford to burn like 30 seconds on a single move. In a one minute bullet game I have to pace myself to spend only one second on my calculations on average. There's not much I can calculate in one second and I also have to calculate for my opponent ideas or I miss some combination.
PS: the computer has to have around 3-4 minutes per move or 40 moves per 2 hours to play at its best. This is the control for classical games. Also the permanent brain should be on, i.e. the computer thinks while the opponent thinks. BTW, back in the day chess players used to leave the computer running overnight. That was for preparation in openings. Giving so much time was necessary so that the machine would find good moves and produce good recommendations in openings.
Thank you.