When is a sacrifice really a sacrifice?

I think most of us, regardless of who we wanted to win, found ourselves quietly chuffed with Carlsen's 'queen sac' to win the last rapid play game and retain his title. Was it, however, really a sacrifice in the strict sense of the word? Without taking away from what he did with his queen, both in terms of aesthetic appeal and simply spotting the move, I would argue that it wasn't really a sacrifice. Rather it was a combination.

The difference can be expressed in the following terms. A combination as I understand it may involve the temporary surrender of material (calculated in the conventional numerical terms) with the calculable consequence of forcing the gain of more material than originally surrendered or mate. Some combinations are difficult to calculate, others are difficult to spot in the first place, and still others are particularly pleasing on eye (including Carlsen's). However, these are not strictly speaking sacrifices. 

Sacrifices to my understanding involve the surrender of material (understood in numerical terms) for less tangible advantages, for example the control of vital squares or other positional advantages that may be converted into wins. The difference between the two (the combination and the sacrifice) is that the player executing the sacrifice is unable to calculate all the apparently playable continuations and is generally relying on intuition, positional and strategic understanding, and a degree of luck. Sacrifices, as Tal put it, are where you "take your opponent into a deep dark forest where 2+2=5 and the path leading out is only wide enough for one". He was reaching, I believe, for the same idea when he famously said, "Some sacrifices are sound; the rest are mine". 

In short what Carlsen executed was not a queen sacrifice. It was an elegant combination involving the surrender of his queen to force calculable greater gain, specifically mate.

Sorry Magnus! Hope I didn't whizz on your fireworks.

I'd say it was still a sac, but not a dramatic one where you think "God, he's blundered!!" then realise how clever it is.

Methinks you are channeling Ayn Rand here. Giving up a lower value (the Q in this case) in order to obtain a higher value (checkmate!) is not a sacrifice in Randian ethics. An example Rand gives from everyday life is a mother who decides not to buy the new hat she wants (Rand wrote a long time ago, when hats were important) in order to buy milk for her child. She is not making a sacrifice, because her child's nutrition is more important to her than the hat.

Turning back to chess, which is certainly more interesting than hats, or milk for that matter, I would say there are two kinds of sacrifices. One is a concrete part of a combination: you calculate that by giving up one, two, three, or more pieces, you will checkmate your opponent or regain that material and then some. In chess talk, it's still a sacrifice, though my agreement with Randian metaphysics and ethics does make me agree your point from a philosophical standpoint.

The second type of sacrifice is, of course, more speculative. Petrosian's famous exchange sacs and many of Tal's attacking sacs (such as the N sac at e6 in game 9 of his 1960 match with Botvinnik) are the clearest examples I can think of at the moment. You're giving something up without knowing that you're getting something in return. Now that's what I call a sacrifice.

I think of Magnus' move as both a sacrifice and a combination. There are two types of sacrifice, tactical (sound) and positional (Tal's). I have always understood the term to encompass both.

What do we think Magnus would have done if SK had taken the queen? I doubt that he would have allowed 'any number of possible moves afterwards'. Those pawn sacrifices to which you refer are to my mind real sacrifices.

And its more than just semantics. Its about giving credit for two entirely different kinds of idea.

We may say there are two different kinds of idea, a process based on precise calculation leading to a definite gain and a process more based on judgement leading to speculative gain, but the words 'combination' and 'sacrifice' do not correspond to this distinction.

'Combination' means 'combining the pieces', making them work together. Any process which does that is a combination. Thus we are engaged in combinations almost the whole of the game. Of course, most of them are commonplace and not worth a comment. It is those which are deep or surprising that merit praise.

'Sacrifice' means 'giving up or surrendering' something. Thus any process which involves giving up material or trading a unit of higher value for one of lower value is a sacrifice. A sacrifice could also be one of position, such as abandoning the centre for a kingside attack, or allowing the opponent to invade your own camp so as to carry out a quicker attack on his. 

The distinction between the two kinds of sacrifice is valid and worth bearing in mind, but with all respect it seems to me a mistake to try and label them with two terms that really do not express that difference.

This is a really interesting question, thanks @Karpark for bringing it up.

I thought this was already in the literature. A "real" sacrifice basically  means you don't get the material back (or mate) in the foreseeable future. A "fake" sacrifice is you immediately win back the material (usually plus extra) or immediately mate.

Sure Carlsen played a queen sac, but is was the "fake" variety. Sure, you can call it a combination, whatever. I think we all understand. It was still a nice finish to a match.

Some interesting posts. I think we are by and large agreed that there is a distinction to be made between the two kinds of event. I'd like to discuss briefly the terminology used for these and the question of whether this distinction is of any importance.

I haven't come across this distinction in the literature made as a general point, though I'm sure that I would have come across annotations of the kind that might go "Black surrenders his rook but only in the knowledge that all continuations lead to mate" and similar. That said, I would be amazed if this distinction hadn't be made in categorical terms somewhere or other. It certainly does seem, however, as though quite a few of the posters so far hadn't really encountered this distinction expressed as a  general idea, so I'm glad started this thread.

With regard to the term 'combination' I am surprised that more of you haven't encountered this term as it has been used in the chess literature (certainly up until the mid 1970s when most of my chess reading took place). In the English speaking chess world the term 'combination' is/was conventionally used when one player initiated a calculated tactical sequence which even with best play by his or her opponent produced a desirable result in terms of material gain or improved position. "Fake" or "tactical" sacrifices thus constituted in this usage a subset of the larger set of sequences that could be described conventionally as combinations. Some combinations may not involve the temporary loss of material but may produce desirable outcomes in the form of positional advantages such as a dangerous passed pawn or a knight on a good outpost. The point is that combinations are essentially tactical sequences, and that all fake sacrifices are indeed combinations (though not all combinations by any means are fake sacrifices). Has nobody else really come across the use of the term combination in this sense in the chess literature?

The second issue is whether this is a really important distinction. I guess not in the sense that both kinds of sacrifice, real and fake, are played to produce wins and it is that which in the final analysis counts for those involved in games which feature either one of those two kinds of event. However, in the aesthetic sense it is significant, I would argue. I was watching Carlsen's 'queen sac' (as it was presented to us) in real time and, as described above, I was delighted by it. Had I been reading about his victory afterwards and been informed that the last game was finally settled with a queen sacrifice, I might have been a little disappointed once I came, with high expectations, to look at it expecting something like a "real" sacrifice. Sure it is certainly cute and aesthetically pleasing (the more so for appearing in the context of a WCC decider) but even as a fake sacrifice it isn't Marshall's 'Golden Shower' queen sacrifice* (which was equally fake but which had more calculated depth). If I purchased a book entitled 'The 100 Greatest Sacrifices in Chess' (and I'm sure that there a number of books out there with similar titles or content), I would be similarly disappointed if I found that they were all fake sacrifices, even if quite a few of them were as surprising and deep as Marshall's conception. I would want to see Petrosian's famous exchange sacrifices, as mentioned by OldPatzerMike above (enjoyed the moral philosophy analogy by the way with its Game Theory implications, Mike), even some pawn sacrifices so long as they were real sacrifices. That's not to say that I don't enjoy fake sacrifices, but rather to emphasise that the fake sacrifice and the real sacrifice are really quite different animals.

* If you're thinking of Googling up "Marshall's golden shower" be sure to set your browser settings to 'safe search'!

If I may come back to my view on this, when we call a sequence of moves a combination, that simply means that different piece functions are being used together (that is, in combination) to produce a result.

The word 'combination' on its own does not tell us whether the result, or the sequence itself, was good or bad. We may speak of a good or a bad combination, a failed combination or a pointless combination. All these are valid expressions.

When a difficult or brilliant sequence of moves is played, we may say it is a deep, great or brilliant combination, etc. Then it is the adjectives which express the quality of the play, not the word 'combination'.

Suppose I push a pawn, which drives off an enemy piece, opening a line for my rook, and I then play rook takes knight. If the rook is taken, then my knight gives check on a now undefended square, forking king and queen.

In this case, the second move, surrendering the higher nominal value of rook for knight, is a sacrifice. The sacrifice is just that one move, nothing else. It is simply the act of giving up that value, and that remains true, no matter what the result is.

If the result is not material but positional gain, it is still a sacrifice. If there is a flaw in the combination (suppose the rook move leaves the back rank unguarded) then the move rook takes knight is still a sacrfice and it is still part of a combination, even though it is a blunder allowing the opponent to mate.

On this understanding of the words, there is no such thing as a 'fake sacrifice'. The move either surrenders nominal value or it does not. If it does, that move is a sacrifice, regardless whether the result is good, bad or indifferent and regardless whether it is part of a deep or a shallow combination, or a clear or unclear one.

Note that this definition is based on nominal value: that is, the traditional numerical value of a rook as equal to five pawns, a knight to three, etc. That is a convenient objective standard, expressing the typical or statistically average value of the pieces.

Once we discard the nominal valuation and start talking about the real value of a piece in an actual position on the board, then values will vary from one position to the next and we lose the basis of comparison for calling it a sacrifice. Our benchmark is gone.

On the other hand, one of the key features of a valid sacrificial combination is precisely that it treats pieces according to their real value in the position. It discounts the nominal value of the pieces.

This involves freeing our mind from the expectations created by those nominal values. That mental leap or emancipation is what makes it difficult to find the sacrificial combination, and that is what makes it beautiful and satisfying to achieve and to contemplate.

Without wishing to sound pedantic, wandle, I would disagree with you about the term 'combination'. Certainly you are right about this word in its everyday sense, but chess authors have used it for a long time to mean something different. Combinations in chess are, by definition, sound and successful tactical sequences that produce desired results. 'Brilliant combinations' are those that are particularly hard to see or calculate, but they are no different in kind insofar as less 'brilliant' combinations are equally sound and successful. Certainly we encounter 'failed combinations', but these are not then combinations in the chess sense, though players of these (maybe missing a forcing move or something else) may not realize that they are flawed when they initiate these tactical sequences. I don't have Reinfield's book (link below) but I am sure that good definitions of 'combination' and 'sacrifice' as these terms are conventionally used in the chess world (as opposed to the ways they are used in the wider non-chess world) would be found in this volume.

https://www.amazon.co.uk/1001-Winning-Chess-Sacrifices-Combinations/dp/0879801115

I stand by my view, for the chess-based reasons I have given. It is the result of discussion, reading and reflection over the years on the chess application of these terms, and particularly of my own analysis, for what it is worth, of the meaning of the game.

Okay, we'll agree to disagree and shake hands as gentlemen. I'll be interested, though, to know how others have hitherto understood these terms.

Certainly. Same here.

@alexm2310 - I agree that the term 'sacrifice' has historically been used to describe both 'fake' sacrifices (like Carlsen's Qh6+) and 'real' sacrifices, though I am arguing that the two are quite different and perhaps ought to be terminologically distinguished. Where I disagree more with wandle is how the term 'combination' is and has been used in the chess literature.

term u l00king f0r is sham sacrifice

Real versus sham[edit]

Rudolf Spielmann proposed a division between sham and real sacrifices:

• In a real sacrifice, the sacrificing player will often have to play on with less material than his opponent for quite some time.
• In a sham sacrifice, the player offering the sacrifice will soon regain material of the same or greater value, or else force mate. A sham sacrifice of this latter type is sometimes known as a pseudo sacrifice.^[2]

In compensation for a real sacrifice, the player receives dynamic advantages which he must capitalize on, or risk losing the game due to the material deficit. Because of the risk involved, real sacrifices are also called speculative sacrifices.

Thanks DBG. Do you remember by any chance where Spielmann published his thoughts about this division?