Help! Why don't fairy chess pieces have values?

If you would look at the game it’s obvious those values fail. I expected you wouldn’t watch the game because it shows your academic preaching is worthless. Theory in a vacuum without application in the real world is pointless. Show me one bad trade I made in that game where your piece values show my ‘mistake.’ Until you do, you’re admitting I’m correct.

It is funny to think a single game can teach you anything conclusive about piecevalues, which are in fact averages over a huge game tree. So no, I am not going to waste my time looking at that game; there are plenty of free 10x8 engines, and if you are interested you can feed the game to such an engine in analysis mode to see what trades it considers a bad deal.

Tactics is extremely important in Chess, and even more so in 10x8 Chess. As a consequence the player that can think ahead deeper than his opponent usually wins. Games between players with unequal tactical abilities therefore say little to nothing about any strategic misconceptions they might have (such as piece values), if these are not wildly crazy. So yes, when you set Q=1 and P=9 in an engine, that engine will lose almost any game. But if you set Q=7, instead of the usual 9.5, the engine does not become all that much weaker. The occasional wrong trade it would make is easily offset by an extra ply of search depth. You don't have to take my word for it; it is something you can easily test yourself.

So a match between Letchworthshire and 'my program' is completely pointless. As I explained, 10x8 programs are typically weak, so I am pretty sure it would lose to a strong human. That doesn't prove it used sub-optimal piece values; because it can be weak for a zillion other reasons. To be able to ascribe the result to the piece values you have to be sure the piece values were the only difference between the two players. (Or, more efficiently, make sure there is no difference between the players at all, but just give them different pieces to start with, and may the most-valuable piece win.)

This discussion also seems to go off at a tangent by focussing on 10x8 Chess, and the value of the Archbishop. But this was not the main point I raised. The failure of the safe-checking theory is demonstrated much more clearly for the non-royal King than for the Archbishop. So let's focus on that.

And to demonstrate that I am actually prepared to play a game against any of you, as I already write: I replace my Queen by 8 non-royal Kings. Since I am probably only something like 1600 USCF, you should have no problem beating me with such a large advantage, a Queen worth 9.5 against 8 non-royal Kings each worth 0, right? Yet I am about as confident you cannot even beat a player as weak as me at this. After all, I can also beat a GM in KQK; when ahead enough a very minimal skill is enough to harvest the win.

So yes, let's play; you can start. (We will have to remember which King is the royal one, but I don't think that will be a problem, as I will likely never have to move it....)

So basically “No, playing a match of you vs. my program with bad pieces weights won’t prove anything, but play me in this make believe game that nobody ever plays will?” No wonder nobody respects academics like yourself. You live in your own dream world that is disconnected from reality. Thanks for losing the debate in front of such a large audience. Your absurd counter proposal was very entertaining.

I was not aware there was any 'debate'. So far we have not seen a single argument on the subject matter from you, just rants on what you think about 'academics', and me in particular. And what this allegedly 'large audience' thinks about that is really for them to decide, not you. In fact I would be surprised if there were many others for which the explanation I gave was over their head. But why not just ask them:

Is there anyone else here who agrees with Letchworthshire that the non-royal King has zero value?

Sometimes it only takes a peasant to point out the emperor is naked. And denying the fact, or stressing the humble origins of the peasant won't make the crowd stop laughing!

I was white in this game. Proof Queen is not greater than Archbishop. The strongest Gothic player we know of had black. He avoided Arch takes Queen and Arch takes Chancellor. Because piece weights can’t be used with supermajors. The positional aspects are more important.

Why are you still babbling about the Archbishop? That is almost completely off topic. The issue I was discussing is the value of the non-royal King, which this 'safe checks' theory predicts to be zero, as it cannot deliver any. As I think most readers here (if there actually still are any...) know very well (from some other variants that can be played on chess.com) its value is about similar to that of the Knight (and many FIDE GMs have acually rated the Kings 'fighting value' as large as 4), this shows in a very obvious way the safe-checking theory utterly sucks (100% wrong value!).

And if it sucks for one piece, the logical consequence is that a correctly predicted value for some other pieces will be purely accidental. Even a broken clock indicates the correct time twice a day.

So to stay on topic: do you still maintain non-royal Kings have no value, or are you ready to admit safe checking has no relation to value? It doesn't seem rocket science to me...

Taylor’s paper in 1876 was widely hailed as brilliant. The safe check concept produced a baseline for piece values, not an ‘evaluation function’ for every chess position. An apple cart is not built for oranges.

Well, so we have your word for it that people in 1876 were pretty stupid.

So what? That doesn't alter the fact that it gives a totally wrong 'baseline value' (whatever that may mean) for the non-royal King.

'Piece value' is a well-established concept, to be used as a guideline for which trades are typically favorable and which not. They are not an 'evaluation function for every chess position', so it mistifies me that you even mention that.

You'd better elucidate the reader as to your position in this. Does this save check concept described in the paper provide us with piece values in the above meaning (which for the orthodox pieces boils down to something like P=1, N=B=3, R=5, Q=9), or do they provide some other quantity ('baseline value'?), that has nothing to do with the traditonal value concept, and is completely useless for tactical considerations?

Read the paper. You’re not paying me to educate you. I’ve beaten your 10x8 program enough to know it’s coded incorrectly.

It has always been my vocation in life to educate people by disseminating information that might benefit others. In fact I cannot see any other reason why people would post in forums like this at all, and I will happily do it without being payed.

Of course I have read the paper. How else could I know it sucks, and explain in detail where it fails?

So some decades ago I took a very weak Chess program, and converted it to play on a 10x8 board. And, surprise, surprise, by making the board larger it did not suddenly become a super strong program! That you repeatedly mention this as if it would have any bearing on the value of the non-royal King, which doesn't participate in any of the variants that program plays, is a good way to make yourself look silly. (Are you payed for that?)

Since you are apparently not willing to 'educate us' by saying something sensible and has any bearing on the issue, I think there is little point in continuing this discussion. Although it would be interesting to see how much further you want to drag down your reputation by pretending you fail to see the problem, and trying to distract readers from the issue with irrelevancies. For as long as you don't explain us how the safe checks method predicts a plusible value for the non-royal King, the situation will be clear to everyone, and there will be no reason for me to point out the problem ad nauseam.

Your non-royal king is as worthless as your arguments. You can count on one finger the number of people interested in a non royal king. Nobody plays such a game. My performance in 10x8 chess variants speaks for itself.

It might speak for itself, but the only thing it says something about is your ability as a player of 10x8 games with the Capablance piece set. Not really very relevant for a discussion on how to determine values of arbitray pieces, methinks...

So your so-called piece values are supposed to be a measure for how many people use the piece, and not what good it would do you in an actual game? That makes your 'contributions' here pretty irrelevant, as in this forum section people are interested only in the latter.

Anyway, the important thing was to warn people away from this flawed safe-checking idea, and I think you have achieved that goal very well by now. Don't use it for any piece that Letchworthshire considers worthless, as this will produce totally non-sensical results. It should only be used for pieces he considers important (which can be counted on two fingers ), and then the results should be taken as gospel, never mind it is garbage in every other case...

Ok here’s a game not by me but my nemesis. He had several chances to win material back after he sacked The Exchange. He passed up each “winning” opportunity to apply even more pressure. Piece weights such as yours would have taken Bishop x Chancellor, would have never played Rook x Bishop, and this pretty miniature would never have been the result.

So you want to argue now a Bishop should be valued higher than a Rook? Good luck with it!

If you don’t understand the point of the intuitive trades made in that game, that explains why you have easily defeatable software programs. Read Yasser Seirawan’s book, “Take My Rooks.”

White to move. Who is better, and why?

Should we read this as an admission that this nonsensical theory about piece values based on safe checking is garbage, because piece values are garbage in the first place, and trades can only be judged 'intuitively'?

It’s so obvious you can’t answer a simple question because your values are wrong. Great players know when exceptions apply to the general knowledge. You would never be confused with a member of such a group.