Julia's Fairies

No.811,812 (PH)

No.811, 812 
Peter Harris (South Africa)


Original Problems, Julia’s Fairies – 2015 (I): January – June

   →Previous ; →Next ; →List 2015(I)

Please send your original fairy problems to: julia@juliasfairies.com

No.811, 812 by Peter Harris – “The Ultra Maxi condition offers unique possibilities. I have tried to encourage its use.” (PH)

No.812.1 by Peter Harris – An improved version to No.812. (JV)


Royal piece: Piece that executes a function of the King on the board.

White Maximummer Ultra: White’s only legal move is his longest – calculated per maxi rules.

White Maximummer: White must play the geometrically longest move or may choose from among longest moves of equal length, distances being measured from the center of each square. Diagonal and oblique distances are measured from the orthogonal coordinates by using Pythagora’s theorem (take the square root of the sum of the squares of the orthogonal distances). All other orthodox chess rules apply.

Sentinels: When a piece (Ks included but not pawns) moves, a pawn of the same colour appears on the vacated square unless that square is on the first or eighth ranks or there are 8 pawns of that colour on the board already.

Sentinels Pion Advers: When a piece (not a Pawn) moves, a Pawn of the colour of the opposite side appears on the vacated square if it is not on the first or the last rank, and if there are less than 8 Pawns of that colour on the board.

No.811 Peter Harris
South Africa

original – 26.05.2015

HINT by author: (click to show/hide)

white royal ph7 black qg1

hs=4            2 solutions             (1+1)
Royal Ph7
White Maximummer Ultra
(no bK)

Solutions: (click to show/hide)

No.812 Peter Harris
South Africa

original – 26.05.2015

Solutions: (click to show/hide)

white qg1 black ka2

h#6              b) a2→a8            (1+1)
Sentinels Pion Advers
White Maximummer Ultra
(no wK)

No.812.1 Peter Harris
South Africa

Version of No.812 – 27.05.2015

Solutions: (click to show/hide)

white qg1 black kc2 pb4

h#6              b) c2→a5            (1+1)
Sentinels Pion Advers
White Maximummer Ultra
(no wK)

Notify of

Newest Most Voted
Inline Feedbacks
View all comments
Luce Sebastien
Luce Sebastien
May 27, 2015 22:23

812 & 812.1 are C+ Winchloe but
811 has no solution

Kjell Widlert
Kjell Widlert
May 28, 2015 01:25
Reply to  Luce Sebastien

The difference is – of course – in the interpretation of the Ultra Maxi rule. It is clear that White can check only with his longest move, but what if his longest move leaves his K in check or puts it into check (=self-check)? Py, and the Composer, obviously reason that in the first solution of 811, White’s longest move is Qa8-f3, but this is an illegal self-check, and therefore White has no legal move and is stalemated. WinChloe agrees that Qa8-f3 is an illegal self-check, and therefore considers Qa8-a1 or Qa8-h8 to be White’s longest moves – so he is not stalemated.

This is not a case of different handling of the combination of two fairy forms, it is a case of different handling of a single condition.

peter harris
peter harris
May 28, 2015 08:59

The definition of Maxi Ultra is:

A move shorter than the longest is illegal.

The definition is not as per the definition of Maxi.

One definition disallows Qa1, the other allows it.

That is the difference.

[With Maxi Ultra and wKa1, bKc1 White would be stalemate].

WinChloe has a bug.

May 28, 2015 15:05

It doesn’t sound like a bug – WinChloe is using another definition of Ultra Maxi: a King is in check only when threatened by a move of maximal length (on the contrary checks remain normal in Maxi).

For WinChloe, “Maximal length” is defined only among legal moves and hence Ka1-b1 is of maximal length in the position wKa1 bKc1 and condition Ultra Maxi.

peter harris
peter harris
May 28, 2015 18:03

The problem has WHITEmaxi ultra and my wKa1, bKc1 example was in this context.

With WhiteMaxi Ultra [with no Maxi of any sort applicable to Black – as per my problem] would White be stalemate or not – according to WinChloe?

WinChloe has a bug – even with the definition you give for Maxi Ultra.

Nikola Predrag
Nikola Predrag
May 28, 2015 19:35

If the Winchloe-definition is indeed as it was mentioned, Ka1-b1 is not legal since only White is constrained by the condition.
But Ka1-a2 would be legal so, Ka1-b1 is indeed of maximal length 🙂

May 28, 2015 19:46

Not stalemated – According to WinChloe Ka2 is legal under the condition white ultra maxi.

It fits WinChloe’s own definition “White must play a max-length move among the legal ones”. Indeed the overall max-length move is Kb2 which is not legal, and hence Ka1 is free to move on square a2.

Your definition (the one coming from StrateGems I guess) is just “White must play a max-length move”. According to this definition white is stalemated as the unique max-length move is Kb2 which is an auto-check.

So I still think there is no bug here, but just 2 different definitions, according to whether or not the max length move has to be chosen between every possible moves (the definition used in your problem) or only between every legal moves (WinChloe’s definition).

peter harris
peter harris
May 29, 2015 01:30

According to WinChloe, when a King whose side is subject to the Maxi Ultra condition is in check, the position is THE SAME as if Maxi [only] applied.

Thus WinChloe DEGRADES the Ultra condition and LESSENS ITS CHARM – both undesirable.

[For example if a position was BlackMaxi Ultra bKa8, bPh7, wPb7 Popeye say it is mate].

May 29, 2015 02:50

For the moment I didn’t introduce any argument in favor of one or the other ultra-max definition – I just pointed-out the existence of such an alternative.

As you now seem to accept that WinChloe is not buggy but is working with another definition than yours, the next step is obviously to try to detect which is the best approach between WinChloe’s and Popeye’s.

You said that WinChloe degrades and lessens the charm of the condition. Maybe this is a correct opinion but I do prefer, to get mine, to see how ultra-max behaves regarding old problems.

So I looked at some of them dealing with this condition, and was unable to find one which is unsound using WinChloe’s definition, but I found some which become unsound using Popeye’s definition.

As an interesting example, consider the one published by Jean Oudot, The Problemist 1971, s#2. The skeleton of the final mating position for b) is: wKa8 wPa7 bKc8 bBh1 and the condition is black ultra-max.

This is not a mating position for Popeye as White is able to play Kb8#. Indeed this move is legal and moreover a checkmate as the next black move is Bh1-a8 or nothing! Of course Kb8 is a self-check for WinChloe as Black is no more obliged to play the overall max-length move Ba8.

The conclusion is that Popeye’s definition of black ultra-max (and thus of white ultra-max) doesn’t correctly fit the intentions of the precursors. To my eyes this is an enough strong argument to claim that WinChloe’s definition should be considered as the right one.

peter harris
peter harris
May 29, 2015 08:25

The consequences of the of WinChloe’s definition of Maxi Ultra are far more far-reaching than I realized when writing my last comment.

The essence of the Ultra concept is that moves shorter than the longest are illegal.

This is generally accepted and is used in the Popeye program.

The consequences of having WinChloe’s definition is that of the millions of h# problems composed with the Blackmaxi condition it would make no difference in 99% of them if the condition was made Blackmaxi Ultra – and vice versa.

This makes WinChloe’s definition absurd.

It is absurd because one might as well not have an Ultra condition at all.

peter harris
peter harris
May 29, 2015 10:14

“This is not a mating position for Popeye as White is able to play Kb8#. Indeed this move is legal and moreover a checkmate as the next black move is Bh1-a8 or nothing! Of course Kb8 is a self-check for WinChloe as Black is no more obliged to play the overall max-length move Ba8.”

wKb8 is not self-check with WinChloe. It is not mate with WinChloe

Nikola Predrag
Nikola Predrag
May 29, 2015 15:06

Nicolas wrote
“The skeleton of the final mating position…”
But the WiCh-legality depends on the block/attack of b7,d7,c7.
As long as wKb8 may move to c7, bKc8 is not in check and Bh1-a8 is the longest legal move.
And if wKb8-c7 is illegal, wK would be checked by bK only if bKc8-d7 was illegal.
The relevant information is missing about that “skeleton” or about the WiCh definition.
At least, I’m missing something.

Nikola Predrag
Nikola Predrag
May 29, 2015 15:22

Of course, I missed that only Black is affected by the condition.
So, the legality of bKc8-d7 is relevant for the legality of wKa8-b8.

But how wKa8-b8 could be legal for PY when Ka8-b7 is longer?

Nikola Predrag
Nikola Predrag
May 29, 2015 15:31
Reply to  Nikola Predrag

And again, about Ka8-b7, I forgot that White is not constrained.
But I hope, for WiCh, wKa8-b8+ bKc8-d7 is still relevant.

May 29, 2015 17:05

To summary, there are 3 kinds of black max, the difference being related to the possibility, for the black side, to play moves which are not necessarily of max length (in some special circumstances).

– Black max: 2 special circumstances – to capture the white King and to defend against a white check.

– Black max ultra WinChloe: 1 special circumstance – to defend against a white check.

– Black max ultra Popeye: 0 special circumstance.

So you’re right Peter, for problems such as h# where the capture of the white King is not relevant, there is no difference between black max and black max ultra WinChloe.

To my mind both WinChloe’s and Popeye’s definitions have their own merits – the only difficulty is that they are different albeit called in the same manner! So I suggest to keep “max ultra” for WinChloe’s setting (just because otherwise some old problems become unsound, as I noticed in my previous message), and to introduce a new name – for example “max strict” – to denote Popeye’s setting.

May 29, 2015 17:30

“But I hope, for WiCh, wKa8-b8+ bKc8-d7 is still relevant.”

True, thanks! This is my mistake – An important bPd7 must be added in the skeleton. Now white is really checkmated (for WinChloe but not for Popeye) as wKb8 is a self-check for WinChloe but not for Popeye.

peter harris
peter harris
May 30, 2015 01:03

There was an article of mine on the Maxi Ultra condition in the May 2012 Problemist in which I gave explanations together with 12 Original examples with various stipulations and additional conditions.

I show 3 of them below

(a) wMaxi (b) wMaxi Ultra
beg pie
whi qb8
bla qa8 kd8
stip h#5
cond sentinelles pionad whitemaxi
twin sentinelles pionad whitemaxi ultra

(a) bMaxi Ultra (b) wMaxi Ultra
beg pie
whi ka1 qa2
bla qa7
stip hs#4
cond sentinelles antiandernach blackmaxi ultra
cond sentinelles antiandernach whitemaxi ultra

(b) + transmutedkings
beg pie
whi pc7d6e5
bla kf4 pa5
stip ser-h#13
cond blackmaxi ultra sentinelles pionadvers
cond blackmaxi ultra sentinelles pionadvers transmutedkings

peter harris
peter harris
May 30, 2015 01:24

In the first example above I see I omitted “cond” from the Popeye input for the twin.

May 30, 2015 02:28

The following example is another good illustration of the fact that Popeye’s definition of Max Ultra is not correct regarding how this condition was handled historically:

Anders Olson, Springaren 1951
s#2 Black Max Ultra
White : Kh8 Rç8é8 Bg8 Pg7h3
Black : Kh6 Bç3 Sé4 Pç6é6

1.R×ç6! B×g7 2.Ré×é6+ Sf6‡

According to WinChloe, Bg7 threatens the white King as Bxh8 is a longest legal black move. Moreover it is impossible for white to unpin the Sf6 (undoing this way the check by allowing black to move its Sf6), hence white is checkmated.

According to Popeye, the white King is even not in-check as black has no other possibility than playing a longest move, i.e. a knight one. In particular black on-move would be stalemated.

Kjell Widlert
Kjell Widlert
June 1, 2015 02:22
Reply to  dupont

The Olson problem was not published as an “ultra maxi” but as a “logical maxi”, in which a check on the white king must be a maxi move, but a check on the black king must always be defended if possible (by the longest defending move, of course). The problem has an intentional set-play, by the way: 1. – Bxg7 2.h4 Ba1#; not 1.h4? Bxf6 2.h5 Ba1 3.Kh7+! (Kg5); that’s why the bS has to be forced to f6!

So with respect to the effects on the two kings, the “logical” maxi works much like e.g. BackHome.

Maybe we should reintroduce the name “logical maxi” for this variety?

(I remember there are also the “consistent maxi” of J. Jessurun Lobo, and the “exact maxi” of Branko Pavlovic, but I haven’t tried to find out if and how they differ from the “logical” and the “ultra” maxis.)

peter harris
peter harris
May 30, 2015 14:45

Good that you found a 1951 Maxi Ultra problem.

Our fairy chess problem world is best served by using the definition that gives the greatest scope for composing interesting problems. [I read this somewhere! It was said by some wise authority]. So Popeye’s divergence from the 1951 regime should be judged accordingly.

[Please let me know if you find a 1940 problem that follows the Popeye definition!]

Juraj Lörinc
May 30, 2015 17:10
Reply to  peter harris

I remember I have used the mentioned idea in the discussion (that the fairy rule with choices offering the best possibilities is preferable), I agree with it, but the idea is not mine. It was Bedrich Formanek (and he undoubtedly has more authority than myself), who has taught me so.

The problem is, that it is not advisable to have two fairy conditions with different rules, but generally used same name. Generally, the original definition has the first choice in drafting the names. Once Ultra Maximummer was defined, all subsequent definitions – even fairly close – should rather use other name.

Two examples. First: Anticirce. The Calvet form (canonical now) allows captures on the rebirth square, but Yves Cheylan has put a lot of effort in popularizing Anticirce and he preferred option with captures on rebirth square forbidden. Thus we have got Anticirce type Cheylan. For me both conditions, Anticirce and Anticirce type Cheylan are more or less equal as regards apparent popularity and possibilities.

Other example: Disparates. Popeye implementation differs from WinChloe and it was recently agreed right here on JF pages that there will be other form called Disparates PY, underlining the fact that this other form agrees with Popeye implementation.

Perhaps there is a space for Ultramaxi PY too?

Georgy Evseev
Georgy Evseev
May 30, 2015 18:08
Reply to  Juraj Lörinc

I think that historically first problem with this kind of fairy condition is the problem of Carl Eckhardt (WinChloe ID 8403) reprinted in AF 14-44/III as #805 with condition “Ultra Max.” without description.

Unfortunately, its solution goes the same way under WinChloe and Popeye, so it is probably impossible to determine author’s definition, unless somebody has access to Chemnitzer Wochenschach.

Also, by running through problems, marked by WinChloe as “incorrect”, one may find that that _one more_ definition of Ultra Maximummer was available, first used by W.Pauly in the same year 1928.

Georgy Evseev
Georgy Evseev
May 30, 2015 15:29

There was a good article on this issue by Branco Pavlovic in Problem 95-98 (fifty years ago!).

Would love your thoughts, please comment.x