Playing with Fritz 18 at GM level

by Frederic Friedel
12/26/2021 – In November we launched a new edition of our flagship program, Fritz. Version 18 has a special mode in which Fritz allows its opponent to actually win games, if he is able to find clever moves. Many amateurs and club players have sent in games that they won with brilliant ideas. But how about grandmasters? Fritz determines the playing strength of its opponent, and makes him find more difficult brilliancies. Iniyan Panneerselvam, a young GM, has written a report on his experience with Fritz 18.

ChessBase 18 - Mega package ChessBase 18 - Mega package

Winning starts with what you know
The new version 18 offers completely new possibilities for chess training and analysis: playing style analysis, search for strategic themes, access to 6 billion Lichess games, player preparation by matching Lichess games, download Chess.com games with built-in API, built-in cloud engine and much more.

More...

“Definitely helpful”

Iniyan Panneerselvam, 19, is a second-year Arts and Science student majoring in Business administration in India. He started playing at the age of five, and now devotes a lot of time to the game. Since 2008 he has participated in more than 250 tournaments, in State, National and International events, and recorded a string of successes. Iniyan has won more than 150 medals across all competitions, including multiple medals in Youth Olympiad, World Youth Championships, Asians, Commonwealth, and many tournament victories amongst Grandmaster Open tournaments. He is currently rated at 2555.

Here are two recent reports on Iniyan:

We asked Iniyan to test our new software, Fritz 18, at a very early beta stage. We wanted to see how a strong chess player would evaluate the program as a training partner. We received a lot of encouragement from this young player – and a number of suggestions on how we could improve the user experience.

A grandmaster and Fritz

This is Iniyan's general report on his experience with our latest program, Fritz 18. It includes three game he has annotated:

I have played a bit with Fritz 18. I set the program to Grandmaster Level. Here are my honest thoughts on Fritz 18.

New feature of playing according to the player's level is good. With the computers becoming way stronger, there's no point in playing against them any more, unless at huge material odds. That isn't similar to OTB games, and hence not so useful. You learn very little from it. Because of this, we have to find a Human playing partner, but many times it's difficult to find a player to play training games with, because everyone has their own schedule to keep.

Fritz estimates its opponent's rating a few moves into the game. The first few games I played showed my rating at around 2200, but then after a few games my rating had been changed to around 2700. When it rated me at 2200 the computer would make big mistakes, and I had no trouble winning against it. I wasn't sure if the rating was that way because a) I just started to play in Fritz or b) because that's the way the rating starts each game and then changes according to how the game goes.

Clearly the computer is trying to play at my level of skill. My impression is that if I play the opening dubiously, Fritz 18 decides my rating is very low, then it's going to make huge mistakes against me, much more than it should. To clarify, I tried to play opening moves such as 1.h3, 2.g3, 3.a3, 4.b3, just to see how that affects my rating estimate by the computer. The good thing was that, once I had played a few proper games, my rating got pegged around 2700 at the start of the games, no matter how bad I played the openings. This shows that Fritz remembers the player's strength over all the games played against in the past. This makes the idea of Fritz matching its opponent's level quite successful.

We come to the brilliancy allowing part. When I had a lower rating (in the computer's opinion) I was allowed to play a lot of nice but simple tactical shots. Once my rating was up, that opportunities went way down. It became harder to find brilliancies. You had to really keep a careful watchout for them.

Overall I would say that Fritz 18 gives users a unique opportunity to train, and is definitely helpful for players of all levels. I will be using the program often, for sure.

Here are three examples of games Iniyan played against Fritz 18. He writes:

  • Game one: White was given an opportunity to gain advantage early on after Black played g5. Also, an opportunity to win the game  outright by finding the brilliancy late in the game. Fritz gives such chances  few times in a game and if you use those chances you can maximise your results  against Fritz 18.
  • Game two shows us that although Fritz  makes mistakes time to time, you have to play at your best against it, because Fritz is on the whole more consistent and accurate than the human against it. Without making any obvious mistakes, White's position deteriorated from the start of the middle game till the end of the game.
  • Game three showcases the entire idea behind Fritz 18. The program plays at a similar level to the player after initially judging the player's rating through his moves. This allows for mistakes to be made by Fritz and thus gives a chance to play against computers, which hasn't been possible for a long time, unless there is some kind of material or tempo odds. Also the move 22.Qd3 by White (read my note to this in the annotation), allows the further tactical sequence that came. It shows that Fritz goes for the opportunity for the human to come up with a spectacular idea. Note: Fritz was set at Grandmaster Level and Black's rating was judged high during this game. Fritz allows for brilliancies corressponding to the level of the player. If it is a club player, the tactics are simpler, and if it's a GM level player, the brilliancies are  correspondigly tougher.
 

Super-talents in India

I [FF] have spent time with more than a dozen super-talents – I get the impression that early-teen grandmaster in India are grown in paddy fields. Sixteen of them took part in two training camps with Vladimir Kramnik and Boris Gelfand. The first was in Chens-Sur-Leman, France, the second in Chennai, India. It was great fun interacting with these kids. During the time I started an experiment to figure out if they were also unusually talented in other areas, such as general logical tasks. I have written about it in this report: Are chess super-talents generally smarter than regular kids?

The general result was no, they probably are not. Just average. But there was one exception: Iniyan. I have a standard problem which I give my test candidates. You can watch three of them solve it in our session in France:

This is live video of me giving the problem to Prithu Gupta, 15 years old, P. Iniyan, 16, and R. Praggnanandhaa, 13. All three were already full GMs. Watch their faces, especially Pragg's, towards the end, while they are solving the logical puzzles. And it is clear who is the most talented in solving logical problems.

Links


Editor-in-Chief emeritus of the ChessBase News page. Studied Philosophy and Linguistics at the University of Hamburg and Oxford, graduating with a thesis on speech act theory and moral language. He started a university career but switched to science journalism, producing documentaries for German TV. In 1986 he co-founded ChessBase.

Discuss

Rules for reader comments

 
 

Not registered yet? Register