Monday, December 25, 2017

Invisible moves part 2

Once I got the questions of an exam in advance from a classmate at the university. His cousin followed the same studies at another university. Besides one of those courses was teached by the same professor. As his exams for that course were a couple of days earlier than ours, we were able to study carefully the questions he got. Later it became clear that the professor didn't take this scenario into account as we got exactly the same questions of the cousin. Of course we all scored extremely high at the exam.

This is also valid for chess. Things which we saw earlier, will be recognized and solved much easier. This effect we clearly see at tactic-servers. Although some solvers have very moderate otb-ratings, they manage to obtain very high online tactic-ratings. Chess.com-member 2012VAChamp is the leader today with a stunning rating of 6482 elo (best Belg at chess.com Superdog-II has only 2900). 2012VAChamp explains at his profile that he has memorized all +3000 elo excercies. He estimates that there are about 500-1000.

For me this is the most important reason to not solve more than 5 each day. As non-paying member you are anyway not allowed to solve more than 5 but I could bypass this limit by using my FM-title and request the diamant-status. Besides I see that Warre De Waele has just requested this status as his tournament-victory in Le Touquet (see e.g. holidays part 3) put the foundation of the new FM-title. Nonetheless despite maximum 5 exercises each day, I notice some I have solved once before. The one below I managed recently to solve in only a couple of seconds as it was already the second time presented to me. I was able to recognize the position instantly and only the mouse-clicks took a couple of seconds.

Some people indicate that they solve the same exercises 20 or more times. This has nothing to do anymore with practicing tactics but rather shows how eager they are to get an extremely high tactic-rating. Vanity is still a very wide-spread human weakness and at the same time a source of schadenfreude. It is why the English program Keeping Up Appearances was extremely popular in the 90's.

Once those people are sitting at the board then not much is left of their tactical wizardry. Then simple exercises are unsolvable. Without the memorization they are helpless. Their online tactic-ratings would be very different if the server would process only fresh positions. Unfortunately this won't happen soon as you need a huge database to keep track of all the records of all members (today this is only done for the 25 most recent solved exercises).

The megadatabase seems to me a better tool to define the difficulty of a specific position not only more accurately but also at a much larger scale. In my previous article invisible moves I already indicated this can be done only for opening-positions. This time I want to add that we should only focus at positions not played at the professional-level as mistakes are immediately detected and corrected.

Especially the first round of open tournaments very often generate some interesting stuff to study. Besides the miniatures where the stronger player swiftly punishes the mistakes of the weaker player, we also detect games in which the win occurs less smoothly. That was definitely the case in my first round of the last Open Leuven against Mats Bakker.

Afterwards I found 8 games in the megadatabase with the same position after white's 7th move. In none of them the right move was played while 1 time it was even missed by a grandmaster of Azerbaijan: Azer Mirzoev see game. Further online I also missed it already twice but I very rarely study blitz-games see the (non-)sense of blitz.

This position breaks my previous personal record as most invisible move in my career. In my article scholar's mate I wrote about how popular books about tricks and traps are. So I think it could be a good idea once to collect the most invisible moves from the megadatabase and bundle this into a book. Likely this will be a very original piece of work for which surely some interest will exist.

Brabo

Tuesday, December 19, 2017

Killer-novelties

Playing, playing and playing is the most important ingredient to improve see my article experience. However at some point of time we progress only very little anymore and eventually there is stagnation. We very quickly conclude that this is just natural. You can't forever squeeze a lemon.

Still many things which we do automatically after all those years prevent us of making any major breakthroughs. Only a few dare to question all established routines. A nice example is of course the recent news about Alphazero which tries something different after decades of Alpha-Beta-programming. The top-engines Stockfish, Komodo, Houdini are still making steady progress but Alphazero proves that machine-learning is definitely also a valid programming-track to explore. We could well be at the eve of a significant increase of playing strength of our best chess-engines (we are still far away from playing perfect chess!).

It is not only useful for engines to think out of the box but also we players can benefit from it. Besides we often know very well what is needed to make still some progression as experienced player. Valery Maes wrote a reaction on my article chess-links in which he stated that an IM-title for me is possible but I realize this is not feasible with my current playing- and working schedule. I see 3 domains which can likely make an impact upon my playing level:
- I should play (much) more competitions especially against stronger players (+2300 elo)
- I need to build a much more flexible repertoire so minimum a couple of openings for both colors so I can switch easily.
- Finally I need to dump the Dutch defense or at least I should not play it as my main-opening.

Easy to say of course and much harder to execute. None of the 3 domains will succeed without serious efforts and honestly I don't have the time/ energy for it. Probably my best chance is to profit from my son's chess-career. In a couple of years it must be possible to play together (several) tournaments each year and maybe I also will learn something of his openings when he starts to play better.

Anyway not for everybody it is that difficult to make new progress at a more mature age. I know many players with sometimes decades of experience whom are barely doing any homework ever. They have of course much more margin to improve. The winner of the first Maneblussers-tournament the 38 year old Belgian FM Matthias De Wachter proved this recently with a fide-ratingpeak of 2355 and as told to me with IM-ambitions. Coincidence or not but this rating-gain went along with teaching his daughter Livia chess!

I don't know what exactly Matthias changed at his approach to chess. However I was impressed by his game-preparation of our mutual game which we played in the finished Maneblusserstornooi. As far as I remember correctly, it was the first time in my career that I met a real killer-novelty. Novelties are played in every game (with a couple of exceptions like copycats) but a home-cooked and on top strong (=killer) idea is something very rare on my level. I only was capable of doing it 8 times (at + 800 games!!) see e.g. the list of strength and the expert. Remarkably only 3 of the 8 (e.g the boomerang) are still today not discovered by anybody-else.
So I escaped with a black eye. I was lucky that I played a couple of inferior moves which Matthias had not checked in advance and obliged him to find a non-trivial refutation. After the game there were a lot of speculations about how careless I was. The Dutch defense is a too dubious opening to play non-stop. I am too predictable as an earlier game of me was copied till move 10 which was not only published in the database but also on my blog see a moral victory. These are justified remarks of course. However I still want to nuance the picture. First I really had planned a surprise but to increase the success-rate I decided to answer 3.Bg5 with 3...Nf6 instead of immediately 2...Nf6 which of course allows 3.Bf4. That was a first wrong gamble. The second was that I trusted my very elaborated study of the opening. On my blog I wrote in the article annotations that I only publish a very short summary of my analysis. Of the position raising after the 10th move I had made a lot of extra analysis. Not less than 5 different moves I had studied and even rehearsed for the game-preparation.
So I gambled again wrongly as I missed Matthias' choice. Besides this 6th possibility is very strong. Matthias told me that he found the move after his computer calculated for a while upon the position. I redid the experiment and indeed after 1 hour of calculations and depth 39 in multi-mode (analyzing several lines at once, so here 3) we see Stockfish showing the same preferred first choice.


Once again it is clear that playing a narrow repertoire is risky. It is practically impossible to neutralize all possible killer-novelties in advance by analysis. Besides even if this would be possible then you still need to remember it for months and years. Finally I want to add that this was the very first killer-novelty after more than 800 standard-games. So for now there is no reason to panic.

Brabo

Tuesday, December 12, 2017

Secret

Since July I possess a smartphone. For years I refused to buy one as I considered it expensive and unnecessary. Eventually my employer forced me to acquire one as he would charge me 10 euros per month no matter if I use one or not. Recently the Belgian law changed so no appeal was possible. In short I am definitely not an early adopter of electronic stuff and I will only introduce something new when I am really convinced it has an added value.

So questions about the newest cutest chess apps or programs are better not asked to me. I refer for such questions to a recent article of schaaksite. On the other hand I do warn the reader not to follow blindly the recommendations of the article. Unless you are applying illegal actions, things don't come cheap. Besides for many of the applications exist cheaper or even free alternatives which can be old-fashioned but otherwise function good.

Well I do realize of course that the youth won't listen to my advise. Young people are addicted to quick entertainment and want immediate results by a minimum of effort. A good example of this I already mentioned in my article the Bird. DVDs are surpassing very quickly the classical chess-books. The 12 year old Belgian FM Daniel Dardha is a big fan of the DVDs see a Dutch interview at hln in which Daniel states that he likes to watch them.

However not only amateurs but also professionals enjoy dvds. Former world-champion Viswanathan Anand once again stated in an interview at chess24 that professionals have today to check an enormous amount of information. DVDs are surely easier to digest than books or other sources of information. Besides it doesn't stop here as they still need to work a lot at home individually and create personal analysis. Obviously this work is well shielded from the public. I just read that Chessbase created for that even a special encryption-key to help professionals to secure their databases when they travel to tournaments.

So every professional has secrets which he keeps for himself. It is not a coincidence that often the higher rated player can use the best secrets in his games. A recent example of such secret occurred in the game between Fabiano Caruana and the strong Brittish grandmaster Gawain Jones played at Isle of Man. Both have seen Svidlers Archangels dvd but only Fabiono was aware of a mistake at move 23. Once Fabiano applied his secret on the board, the game was already over.
Between amateurs such secrets are barely popping up. Very few amateurs are up to date with the theory. Many don't have the time to check all publicly available sources, and surely don't spend time at searching novelties. The games are also played in a more relaxed environment. Financially there is little or nothing at stake. The weight of a novelty is rather small upon the result of a game. Finally we as amateurs also have to play against a much wider variety of opponents compared with the very small world of professionals. I am playing more than 20 years of competitions and only 8 times I played against the same player 5 or more times see matches. That is a big difference compared to the world-top playing continuously against each other.

Therefore last I was disappointed and offended when my opponent of the 2nd round in the Belgian interclub: the Dutch IM Xander Wemmers refused firmly to tell what he prepared at home for our game. In the game we got the Avrukh-treatment of the stonewall on the board see for more information about it part 1 en part 2. However as Xander never played this system before (conform the databases) I smelt a rat. I hadn't checked the lines very recently so I thought it would be wise to deviate with a rather new idea which I saw a couple of months earlier. This brought us very quickly on unknown territory so naturally inducing a number of errors.

After the game I was especially interested in what Xander had kept in store for 8...Nbd7 instead of 8... Ne4. Earlier I demonstrated that I made comfortable draws twice in Open Gent against FMs with black. Obviously Xander would not permit me to reproduce such draw. I insisted but Xander didn't give in so the postmortem ended before it even started.

At chesspub I mentioned my case but initially I got very little support. Why would you share something which can still be used later? However the chance is practically non-existent in this particular case even if Xander would never vary his openings anymore. Despite we both play for decades, this was our first game in which I had black. Besides if you look at the database then I am the only player having played 8...Nbd7 more than once see screenshot below.
Games + 2200 elo in the Avrukh Stonewall with 8...Nbd7
Anyway I don't see what we can win here by keeping secrets. It is just very egoistic and absolutely not how I play chess. No, I don't demand that everybody writes a blog to share his deepest secrets but a minimum of altruism is surely necessary if we want to preserve our chess-sport. It is another sad proof that chess-players are extremely individualistic.

Brabo