Nettlesomeness, and the first half of the Carlsen-Anand match

After six games, Carlsen leads by two points, with four draws added to the tally.  Anand seems hell bent on founding a campaign to abolish the advantages of playing with the white pieces.

I find two aspects of the match notable so far.  First, in the last two endgames Carlsen has been outplaying the computer programs (and Anand), sometimes for dozens of moves in a row.  That isn’t easy, to say the least.  And kudos to Alan Turing for realizing early on, in his 1953 paper, that chess-playing computer programs would face special difficulties in understanding some endgames.  The sequences required to establish the importance (or not) of a measurable material advantage can stretch beyond the time horizon of the program, for instance, and the endgame tablebases take us only so far.

Second, Carlsen is demonstrating one of his most feared qualities, namely his “nettlesomeness,” to use a term coined for this purpose by Ken Regan.  Using computer analysis, you can measure which players do the most to cause their opponents to make mistakes.  Carlsen has the highest nettlesomeness score by this metric, because his creative moves pressure the other player and open up a lot of room for mistakes.  In contrast, a player such as Kramnik plays a high percentage of very accurate moves, and of course he is very strong, but those moves are in some way calmer and they are less likely to induce mistakes in response.

Nettlesomeness is an underrated concept in our world, and kudos to Ken for bringing it to our attention.  It should play a larger role in formal game theory than it does currently.  It’s already playing a decisive role in the world of chess.

Addendum: Here are some of Ken’s metrics for “nettlesomeness.”


i want to make sure i'm prepared for the end of average. i don;t want to be left out

I am following the matches. I am a weak chess player, so I have no way of judging the quality. But the comments from Polgar and company make it interesting and offer a window for the uninitiated.

It is fascinating how Carlsen is taking every game to the end. Even obvious draws are played until there is nothing left. Is it because he is playing the long game of tiring an older competitor? The nettlesomeness theory would also predict long games. Unexpectedly long, thus the length is itself a surprise for the oponent. The longer the game, the more likely the other player would make a mistake. All the commentators today were saying that the match was a draw and that many grand masters would have settle for a draw. But Carlsen kept going. One rook each and pawns (Carlsen had an extra pawn and then one fewer), the computer was predicting a draw. And then it wasn't a draw. Amazing.

The alternative hypothesis is that Anand just made dumb mistakes because he is old in chess years.

The problem with that is that Carlson's opponents always seem to be making dumb mistakes, so there is something to the point that Carlson's play has a "make the opponent solve difficult problems" quality to it.

Very presumptuous to call them dumb mistakes. The match isn't some blunderfest, and just because one person wins and the other loses doesn't make one person dumb.. Carlsen is just playing at a different level in the end game. Good match strategy too agreeing to some short draws in two midgame situations and then grinding him down in two rook and pawn endgame situations.

Obviously, the term "dumb" is a relative one, namely relative to what a "good grandmaster" is supposed to be able to do.

The commenting grandmasters (who do not use computers when commenting) caught the mistakes, so for super-grandmasters like Anand to make them is "dumb." Of course, the commenting grandmasters are not under pressure like Anand and would probably find other ways to get in trouble against Carlsen too.

It's not dumb even relative to a "good grandmaster" whatever that means. When someone comes with new ideas and new insight and is a hair better than the next guy, it doesn't mean anyone is "dumb" or got "dumb". And you can be sure the commenting Gm's aren't just pulling lines off the top of their heads in all the time between moves. They're talking to each other and most definitely have access to engines, even if indirectly. There aren't any high level GM's that Anand hasn't beaten.

Take it with a pinch of salt. I don't think @Alex K. meant it literally. It's just his way of alluding to the relative "quality" of mistakes I think.

"When someone comes with new ideas and new insight and is a hair better than the next guy "

That's the problem, the turning points in both decisive games were not due to any new ideas. There was no progressive increase in Carlsen's advange until Anand cracked. Rather, Anand made mistakes that turned drawn games into loses. If calling these mistakes "dumb mistakes" offends you so much, I don't have to use the term.

But the games were not about gradual converting of minor advantages into endgame wins. The most salient features of the games were Anand's mistakes. The question is only what features of Carlsen's game were responsible for Anand making those mistakes.


As far as I can tell Ra4 is not a dumb error. In fact the long run simulations from the engines are consistently showing only a single move is drawable from Anad's position prior to that move - Rb4, a similar but slightly less aggressive maneouver.

Unlike the bloke you replied to, I think "dumb" is not offensive but just inaccurate. More like Anad's was slightly imprecise, and paid the price.

When, in interviews, you claim that you thought the game was lost at point X, and then analysis shows that you still had a drawing position for the next 5 moves, but you didn't take it, the game is a blunderfest, at least at high level standards.

What makes it tougher now is that computers are so good that anyone can see that the GM is making moves that aren't anywhere near what the engine considers the top move.

I think this article capture what makes Carlsen great, and I know where he learned it. To find the practical chances, wiggle and play, do something and then do nothing decisive, act with you whole body, or wait and trick your opponent to do mistakes is classic Simen Agdestein. The game is a fight, enjoy it! Magnus Carlsen is very different from his teacher in how he thinks, but their approach to the psyhology in the game I believe is very similar.

"Nettlesomeness" is most useful when one player can openly and deliberately frustrate the other, but does this describe most non-game problems apart from regulatory evasion and internet trolling?


Right. Nettlesomeness might also be created via non-strictly-gameplay parameters. By mannerisms, etc.

I've no clue if that applies in this case.

Russia is geopolitically weak and Putin's braggadocio doesn't change that. When your strongest allies are Belarus and Kazakhstan, you're in a bad position. Who made mistakes that have benefitted Russia as a result of his behaviour?

In fact, I think it's hard to apply this frustration concept to geopolitics. Unlike chess, it's hard to identify mistakes objectively in retrospect, and harder to prove that better alternatives were forsaken due to a trembling hand.

Putin made us (US) look dumb, in the real sense, recently regarding Syria.

Remember though that engines can find the winning continuation for Carlsen in the fourth game where he couldn't (contrary to Friedel on Chessbase, it involves playing 36... Rd8 37.Ke3 Rd5! 38.Nbc3 Re5! 39.Kf3 Rgxe4! 40.Rxe4 Rxe6! 41.Rxe6 Nd4+ with a won ending for Black). And it's doubtful that even the humans knew whether the ending in today's game was drawn or not, seeing as the win for Carlsen ended up hinging on Anand missing a single tempo. And it depends what you mean by "outplay"; I agree that the moves Carlsen has been finding have done more to press the positions than the engine suggestions, and have thus been more "nettlesome".

Excellent reply! I saw your comment after I posted mine downstream. Indeed, Carlsen failed to see this line, and Anand failed in the press conference to say that he saw that giving check 36...Rf4+ was a mistake, thereby negating any psychological advantage he may have had. That is, Anand should not have commented at all on his moves (which is apparently obligatory per the tournament sponsor), because he failed to say that he thought 36..Rf4+ was a bad move by Carlsen. So Carlsen (just like myself) when he went over this game later must have concluded the same thing I did: that Anand is weak and "blundered" into this draw, quite by accident.

BTW, good spot of GM Friedel's erroneous commentary on Chessbase on this game, which seems to be trying to pander to the Carlsen hero worship. Indeed, my PC in 10 seconds sees that, as you point out, Friedel is not citing the best line for black.

I don't agree that Carlsen is out-playing the computers in the endgames in this match, or any other game he has played in the last 5 years. He is badly outplaying Anand in the late day endgames, however. I am a handful of years older than Anand (I am 47), and I know I have lost the ability to keenly focus on problems for the periods of time I could when I was in my 20s.

As for nettlesomeness, I think this is simply a case where Carlsen is just simply less likely to make a mistake than his opponents, knows he is less likely to make a mistake, and takes advantage of it by refusing to take or offer draws in positions that are not clearly demonstrable draws by human standards.

"As for nettlesomeness, I think this is simply a case where Carlsen is just simply less likely to make a mistake than his opponents, knows he is less likely to make a mistake, and takes advantage of it by refusing to take or offer draws in positions that are not clearly demonstrable draws by human standards."

I agree. As a tournament player (current playing strength ~1950-2000 FIDE) I'd note that you see the exact same behaviour, with often similar consequences in terms of game outcomes, in regular tournament chess games involving players much weaker than these guys. It's common for the higher rated (stronger) player in an equal endgame to refuse a draw and implicitly demand that the lower rated (weaker) player actually demonstrates that he can hold the (theoretical) draw, and often the weaker player can't and cracks under the pressure.

As for the computer comments, I also agree. Unless one is talking about 2005 pocket versions Carlsen doesn't outplay computers in the endgame. Carlsen would be very lucky to survive against a strong computer today and as for outplaying them, that ship sailed a long time ago (ask Mickey Adams... go have a look at the endgame of his fourth game against Hydra from 2005). Also, speaking about problems computers have with endgames seems a bit funny as they play perfectly in a lot of endgames which even the strongest humans alive will mess up, and have done so for a while. For example Anand had a winning endgame with two knights and king against pawn and king against Wang Yue in the 2009 Amber tournament, which he ended up drawing (he had a theoretical win as he could have blocked the pawn with a knight behind the Second Troitzky line - the computer wouldn't have had any problems finding that win). One might argue that Carlsen would have found the win where Anand faild, but I'm pretty confident Carlsen would argue against that notion if one were to ask him.

Thankfully after average is over 47-year-olds will be paid substantially less than 20-somethings of similar peak genetic ability!

Ok, I'm new to this chess-watching thing.

I want to know why it is that on the internet I can only get radio or text only commentary for Cricket. But I apparently have to watch video for Chess!?!?

Is "nettlesome" the right term?

Jimmy Connors was nettlesome in his 1991 comeback US Open match against Aaron Krickstein, in which Connors did things like take long towel breaks to upset his opponent's quick serving rhythm. Connors also rallied the crowd to distract Krickstein, etc.

Here is how you describe Carlsen's nettlesomeness:

"his creative moves pressure the other player and open up a lot of room for mistakes. In contrast, a player such as Kramnik plays a high percentage of very accurate moves, and of course he is very strong, but those moves are in some way calmer and they are less likely to induce mistakes in response."

Sounds quite different from Connors's nettlesomeness.

of course, nettlesome has to be context specific. I am nettle doing things in one place that would not raise an eyebrow in different one. it's one of those, you know it when you see it (or see its effect on others) qualities.

So for you chess nettlesomeness means behaving like Topalov did in 2006...

In golf, nettlesomeness in match play was exemplified by Walter Hagen in the 1920s and Seve Ballasteros in the 1980s. They both tended to get in their opponents heads with bad drives and spectacular recoveries.

At the other extreme, Ben Hogan, Jack Nicklaus, and Tiger Woods got in other players' heads by relentless superiority, especially of the cognitive and emotional side of the game.

I think every avid chess player has played games against inferior opponents where positions arise that are going to lead to draws by repetition because you have made enough mistakes that your logically best course of action is to repeat moves- to do otherwise risks losing. I know I sometimes will take the risk just to avoid the draw because I am confident that I can outplay such opponents enough in even a losing position to turn the game around. Of course, I have been on the other side of this situation, too, where a stronger opponent deviated from his best course of action just to deny me the satisfaction of a draw. Sometimes this pays off, sometimes it doesn't, however, it has been my observation that I usually do better than even overall against weaker opponents, and less than even against stronger ones.

I think Carlsen is doing more of this than is normally seen at this level of chess. He takes risks that expands both sides of the bell curve of outcomes with the firm belief that doing so makes it more likely that he wins rather than loses. I think Carlsen's results prove he is right in doing so.

Example of nettlesomeness: Emanuel Lasker against Aron Nimzowitch, with the "threat" to smoke a cigar.

Nettlesome, or what is called thinking… topic of next book?

A campaign calculated to destroy white advantage? There's a party for that.

They won back in 1860's, but the endgame takes time to play out.

"First, in the last two endgames Carlsen has been outplaying the computer programs (and Anand), sometimes for dozens of moves in a row." Can you tell us more how you arrived at that conclusion, Tyler?

Had he played the Houdini recommendations presented on, for instance, he never would have ended up in a position with so much pressure and so many chances for Anand to go wrong. The game would have remained a deadlocked draw.

You can't know this because you can't know how Anand would have played against those alternatives. Had Carlsen played these moves against Houdini rather than Anand, he might well have lost both the last two games.

Now, Carlsen may well have seen these alternatives over the board, and chose against them because they seemed drier and gave his opponent less scope for error, but I have not heard him say this.

Nettlesomeness is part of the idea behind playing tight-aggressive in poker.

no, loose aggressive play is the equivalent of ''nettlesomeness'' in poker. loose-aggressive play gives up theoretical EV to a hypothetical ideal poker player in order to present real, fallible human poker players with complicated decisions that require more thought, greater uncertainty, and higher variance. tom dwan is the perfect example of a technically unsound professional (see his comments on game theory on 2p2) who made a great deal of money manipulating opponents who knew better into poor play. he's since lost much of it back the inhumanly consistent aspergoids populating the highest stakes games both online and live.

Are you referring to Isildur and also to Antonius and Cates in Dwan's Million Dollar Poker Challenge?

I assume you are Tom Dwan's accountant?

I am a 1900 Elo player, which means Class A, which means I can beat 90% of rated chess players and most of you, but a patzer (weak player) compared to a master, and I agree with Alex K. Anand made dumb mistakes in one of his two losses: game 5. 45. Rc1+ is a HOWLER . A mere class A player can see that this spite check is a helping move. You must not give your opponent outside passed pawns on opposite wings. Any decent player even at my level knows this. No excuse since Anand was not in time trouble. As for game 6, I agree with TC on this one--it was tough for Anand to solve this endgame--and Carlsen is indeed wise to avoid the easy computer generated lines.

BTW the link to Ken Regan is wrong, it merely talks about ratings and makes this curious statement about GM M. Adams which I wish was elaborated upon (indeed Adams, an old timer and the UK's #1 player, has made a resurgence of late): "Adams' resurgence is no accident"

Also the commentary by GM Judith Polgar is quite good, and quite elementary, but that's what you need for a general audience. The commentary by 'hip' IM Lawrence Trent I find annoying, but he's playing to a different audience maybe.

I found GM Anand was weak when he 'blundered' into a draw in Game 4; and I found Carlsen is mortal, failing to find the best move sometimes. See for yourself: (1) go here: (2) click to move 36 and read the commentary.

Finally, Slate is right about the world championship being an anachronism, see

Correction about move #36 of game 4, see the commentary by Saturos upstream and my response--this is better than the commentary on Chessbase by GM Friedel, and what I had in mind when I posted...but it was hard to explain without getting elaborate, and instead I had to rely on an web link that said "looks like a better try than the game continuation", which is true, albeit the line given by Friedel is inferior, see Saturos' post.

What I think is newsworthy is, according to NPR, carlsen is shirtless in a SI spread.

I am a retired 2500+ player. I was many times in situations like this ... of course not on a level like this.

I can imagine the pressure in which Anand is. He is playing against a monster. I mean this positive. And the worst of all, the monster is a symphatic young man, Anand can even not create a hate situation that would give him maybe extra adrenaline and power.

All the other GM's should be ashamed for the past 30 years with all their fast draw agreements.

Here is a new superstar who shows the chess world that chess isn't dead. And that chess is a fighting game! He shows the other GM's that they should have not agreed to drews in theoreticaly drawn positions, but should have checked that the opponent does not collapse under the pressure.

Chess is the UFC of the brain, the mixed marshall arts for the brain. All preparations are nice to be prepared, but to sit on the board is the same as to go in the fighting ring.

In my opinion is chess much more cruel and unforgiving as fist fightings ... and much more beautiful and with less lost teeth!

With Carlsen arrived a true ambassador for the chess sport! In my opinion is it the first time ever, that I accept that is a sport. Because to survive Carlsen, the opponent need to be also physical fit. Carlsen changed the image of chess. It has arrived in the sports world!

I have to register a disagreement with the "nettlesome" thesis. I think there is a different explanation for Carlson's success in games like the last two against Anand: the guy has an unprecedented ability to envision strategic plans. By this I mean that he can take an initial position and visualize extremely subtle strategies for gaining and expanding an advantage. This shows up especially in endgames, because so much of the planning in the middle game now comes out of opening prep.

The ability to discern a "hidden" plan helps focus his calculating efforts. Anand, right from the beginning, was a phenomenon in pure calculating skill; at his peak I think he was matched only by Kasparov. I don't know how far down from that peak he is today, but I suspect he still has a lot of juice left in that department. The reason he failed to hold the endgame in game 6 is that he didn't have a clear sense of Carlson's plan and what it meant for the selection and evaluation of candidate moves. If the idea is to sac the pawns on the queenside to get an f-pawn runner supported by king and rook, you have just a few ideas to focus on: preventing the f-pawn from getting free, using queenside threats to distract either king or rook and neutralize the f-pawn, or use the queenside runners to compensate for the rook that has to be given up to stop the f-pawn. The b4 move he didn't sufficiently calculate was a logical element of counterplan #3. But I don't fault Anand. It is horribly difficult to see these plan possibilities in advance, much less calculate the tactical aspects with precision. Carlson has to be admired for his ability to spot these plans in apparently lifeless positions.

Ironically, Carlson is much stronger at coming up with such plans himself than anticipating them on the part of his opponents. He is hardly invincible, and when he loses it is usually because he was blindsided by the sort of plan that he would probably perceive if he were the instigator. This is his biggest weakness, but it's hard to take advantage of it on a predictable basis unless you're a Carlson yourself. (Anand's nice defensive plan with h5 in game 6 is an example of Carlson missing a stratagem directed at him.)

The great Mr. Nettlesome, at his peak, was Topalov. He had an extraordinary ability to find moves that maintained pressure, that threatened to threaten to threaten. Each individual threat seemed relatively harmless, in that it could be thwarted by several defenses, but typically many of them would be revealed as flawed as further threats were generated. It was not a few moves of murky crisis, as in Tal's games, but an eternity of slow-burning pressure. Topalov could keep this up for 20 moves or more, and only extremely precise calculation on the part of the defender would permit survival.

Tyler seems to be using the notion of nettlesomeness here to capture any play which makes it more likely the opponent will err without altering whether the game is a win, loss or draw with best play. But this notion is too broad. By this notion, Kasparov's excellent opening preparation, Tal's tactical genius and Carlsen's grinding are all "nettlesome" in a large majority of cases when the position remained objectively drawn until the opponent "blundered" (I use quotes to indicate the subtleness of "blundering" in early / complex positions). In fact, almost anything we would describe as "playing chess well" is nettlesome where the moves are not literally altering the game state, which subsumes every notion of good play except the ability to find forcing variations and "only moves".

What Carlsen excels at is really a subset of nettlesomeness which involves creating complications for his opponents in seemingly dry middle games and end games. But that subset of nettlesomeness is less generalizable to circumstances outside chess.

I am just club level player, but I beg to differ with both points in this article and even some of the comments by strong players on this website.

First , I dont think carlsen will outplay houdini in the endgame. In fact the reverse would be true, houdini would do to carslen what he does to his human opponents. Engines can rattle off 10 straight, almost forced moves which will hold the draw from seemingly lost situations. For example in game 6 , 60 .. b4 was drawing, and carslen said in the press conference that it was too slow. One endgame error from Carlsen or any other player for that matter, and the engine will climb over him.

I also disagree that he is nettelsome. He is a low risk player who makes solid moves and waits for opponents to make mistakes, it can be said he is a defensive player. In fact the two games anand lost (especially game 6), Anand could also have made boring, do nothing moves and taken the draw. I found his h-pawn sac astounding. Although he is 2 points behind, I actually think its Anand who has shown the willingness to take a risk and initiative, and for that I think he is the better player so far.

Where does Regan define "nettlesomeness"? Certainly not at the link you posted. Also, what is your evidence that Carlsen outplays engines in endgames?

Comments for this post are closed