Checkmate pattern tags

[richard]

I've wanted to add checkmate pattern tags for a while, but have been putting it off as I wanted a way to structure the tags in hierarchy first (as a way to avoid too much clutter at the top level). However as it seems it is going to be a while before I have time to work on hierarchical tagging structures, and checkmate patterns are probably a very worthwhile tagging (and custom set) target, I'm planning to add these now.

The plan is to cluster them together by prefixing each tag with 'Mate' so you'd have
Mate - Anastasia's
Mate - Arabian
Mate - Boden's

This will provide some structure to the tag list in that tags are sorted alphabetically so to find a particular checkmate pattern you can zero in on the 'Mate -' section of the tag list. Not as nice as a proper hierarchy, but something I can fit in without a lot of development time.

I'm planning to use the list of patterns in the following wikipedia article:
http://en.wikipedia.org/wiki/Checkmate_pattern

If people have some well known mate patterns that they don't see on this list, please let me know (and provide a description and example), and I'll add them to the list of additions.

Regards,
Richard.

[interlist]

A homework assignment.

I'll come back with examples and descriptions later, but off the top of my head (and without revisiting the wiki page) the following are missing or not properly explained:

(Note - using my naming conventions, not sure if there are better alternatives, or any convention, except as noted)

Railroad Mates - one of my favorites of course, Q+R vs. K.  Very useful, and common once you recognize it.

Q+B Escalator - not as common, maybe more fun though. High level, like knowing B+N checkmate.

V.V.Arabian - following Tisdell's convention. R+N, extremely common and useful.

Q+B Dovetail - Abstraction of dovetail mate (instead of pawns blocking the bishop does). Also common and useful.

Q+B Edge Mate - queen three squares off edge (e.g. on 3rd rank) - Q+B diagonals cross - common enough to be useful.

Chimney Mates - the name is well used, the mates somewhat rare, but oh so fun.

Suffocation Mates - I think wiki has this category. Don't know about wiki's distinction oif a bishop covering the extra escape square vs. some other piece. The main idea is to accommodate the mates that are oh-so-close to being smother mates.  


One cavaet about the Wiki categories that I've been meaning to ask here for some time concerns the directionality of some of the "dovetail" mates. I don't see much utility in subdividing a Dovetail into Cozio's Mate (I think that's what I remember). I think the distinction of whether the king is nearer the edge of the board vs. the queen(*) is irrelevant to the way the mate works. I'll have to look this up and review - I might be mistaken.   

Would it be useful for me to elaborate any of these?


(*) [ed- I originally mistakenly put bishop instead of queen here]

[richard]

Hi interlist,

Elaboration would be great, especially for the ones where you are using your own naming convention. I'm not sure what Tisdell's convention is , is this different to the Arabian example on the wiki page?

I agree with the directionality of the dovetail variants, in fact I'd probably not bother separating based on the blocking piece type either.

If you can find examples of any of the patterns, either your own additions or from the wiki page that would be great, as I'd rather not take their examples if I can avoid it.

Regards,
Richard.

[interlist]

Hi interlist,

Elaboration would be great, especially for the ones where you are using your own naming convention. I'm not sure what Tisdell's convention is , is this different to the Arabian example on the wiki page?

V.V. Arabian is named after V.Vukovic, who wrote the book where Tisdell first found it (Art of Attack in Chess). It has kRN in a line perpendicular to edge (somehow the rook is protected). The classic Arabian is different (e.g. king must be in corner or blocked in similar fashion).   

I agree with the directionality of the dovetail variants, in fact I'd probably not bother separating based on the blocking piece type either.

We're in total agreement here.

If you can find examples of any of the patterns, either your own additions or from the wiki page that would be great, as I'd rather not take their examples if I can avoid it.

OK, I'll try to work a few examples (using CT problems) by tomorrow or the next day.

[bogidorus]

Mate - Anastasia's
Mate - Arabian
Mate - Boden's

I am recommending pictorial descriptions. I don't know, what the mate with a knight on f6, a rook on h7 and the opposite king on h8 has to do with Arabia. Of course the mate itself is an essential pattern with R + N. Or Anastasia's Mate - honestly I don't know, what that is. And I am also not sure what Boden is, is it the mate, where you sac a queen on c6 in order to mate with a bishop on a6 in the middlegame? Or the one,where you sac a queen on d8 in order to double-check with Bg5 and mate with Rd8 like in a famous game of Reti?

I can look it up in Wikipedia, but tomorrow I will have forgotten it.

For instance interlists picture of the "Railroad Mate" is excellent, because the name itself describes the method, not only the mate itself. Of course such an universal name is not always easy to find.

Maybe for the purpose of learning it is an idea to cluster the mate patterns according to the participating pieces:

Queen Mates: Q + Q, Q + R, Q + N, Q + B, Q + P , Q + K
Rook Mates: R + R, R + N , R + B , and - probably less important - R + P , R + K

and so forth (I would classify a Railroad Mate as a Q + R also, if the rook executes the mate)

[richard]

Hi bogidorus,

I'm probably going to stick with the traditional names where they are available. "railroad" mate is a nice pictorial name, but someone also has zero chance of knowing what it is without reading a definition, as is the case with Boden's etc. Of course it might aid in remembering it once they have seen it.

I'm also reluctant to go with a classification based on participating pieces, just because at the moment there is not an easy way to do the clustering. I'm clustering mate patterns by prefixing them with "Mate -", but it starts to get pretty messy if that is extended further.  They also don't really solve the non-pictorial issue by replacing the need for specific names, as a lot of the "named" mates are about more than just the pieces involved, it is often a particular configuration or action of the pieces that defines them rather than just the pieces themselves.

Regards,
Richard.

[rombelstielz]

To be honest: the Rombelstielz is a bit sceptical if this would be a valuable feature.
It feels "too much" to learn / too time-consuming to look it up each time we tag,
but maybe this "problem" disappears after some 100 tagged exercises. 

[richard]

To be honest: the Rombelstielz is a bit sceptical if this would be a valuable feature.
It feels "too much" to learn / too time-consuming to look it up each time we tag,
but maybe this "problem" disappears after some 100 tagged exercises. 

It only takes a minority of users to get familiar with the tags and use them for them to be useful for all premium members. Once there has been a reasonable number tagged, premium members can then use their per-motif stats to see which mate patterns they are doing well on , and which appear to need work (and then create sets targeting the ones which need work).  I suspect that for many of the very specific mating patterns, per mating pattern custom sets will be very effective in improving the recognition of those particular mates due to the high level of specificity compared to other tactical motifs that are much more general in theme.

Regards,
Richard.

[richard]

I've been working on this over the last few days, and have most of the wikipedia patterns done (nothing ready for upload yet). I have a few questions, and request for help in tracking down some missing examples.

Firstly I'm a bit confused about the differences between the Opera Mate and Anderssen's mate. According to Wikipedia Anderssen's mate is "the rook or queen is supported by a diagonally-attacking piece such as a pawn or bishop as it checkmates the opposing king along the eighth rank."  The Opera Mate is: "It works by attacking the king on the back rank with a rook using a bishop to protect it.". The opera mate goes on to say "A pawn or other piece other than a knight of the enemy king's is used to restrict its movement.", but an extra piece, blocking or protecting an escape square is also required in Anderssen's mate. Has wikipedia perhaps got a little too generic on Anderssen's, and it should refer to the specific case of a pawn supported rook?

I started to add some of what Wikipedia calls "the four major checkmates mates", i.e. Rook mate, Queen mate, king and two bishops checkmate, and bishop and knight checkmate (all with bare king). However I'm wondering how useful these are in the context of CT tactical problems where the positions are usually never simple enough to get down to such basic mates. I could remove the requirement for bare king, but then these tags might become the mating equivalent of "Exposed King" where anything involving a rook for example gets tagged "Rook Mate". What do other's think?

I've avoided 'h' file mate for now for similar reasons, i.e. it seems very generic, and it might be better to define specific pattern names for typical 'h' file mates not covered by existing tags.

I've made a few of the wikipedia definitions a little more generic in the CT versions where I think the pattern remained essentially the same. For example in some cases where a bishop might be specified on the wikipedia definition, I allowed for a bishop or queen, where the non-diagonal movement of the queen in the situation isn't relevant. Similar widening in scope were made from rook to rook or queen in some definitions.

Definitions I haven't been able to find nice problem example positions for so far are (see wikipedia page for definitions):
- Corner Mate.
- Bishop and Knight Mate (one of the "Major" mates I'm thinking might be better off removed for now).
- Max Lange's mate.
- Réti's mate (lots of people seem to get this confused with the Opera mate, which if I understand correctly is the opposite, Reti is bishop giving mate with rook protecting, Opera is rook giving mate with bishop protecting).
- Rook and King (another "Major" mate I may remove).

I've been going through problem comments to extract some of the examples, by searching for mate names, and there is already a fair bit of confusion on some of the mate patterns. I had been thinking that pictorial descriptions would be best where traditional names are not yet available, but after looking through some of these, I'm not as convinced. The issue with a pictorial description is that while it often nicely evokes the mate it refers to, it also tends to be something that can be more easily overgeneralised. For example "Corner mate" is quite specific in the wikipedia definition, but I can see a lot of people tagging any mate that happens in the corner with that tag (either the king or the mating piece). Named mates don't have that problem, in that you really do need to look them up (or have heard of them before) to sensibly tag them, as you can't be tricked into thinking you've guessed the correct motif from the pictorial name.

I'll have a look at some of interlists suggestions once I've got the existing the wikipedia entries nailed down. I've noticed for things like the Q+B Dovetail that interlists suggested, that users are already calling these dovetails in the comments, but I tend to agree that the pattern is different enough to warrant a specific variant (the bishop is attacking, while in the traditional dovetail, the pawns (or other pieces) are blocking).

Anyway, I'm quite enthusiastic about this, as I mentioned to rombelstielz, I think these reasonably common, but quite specific patterns have a fairly big bang for back value in terms of improved tactical performance by remembering them.

Regards,
Richard.

[interlist]

In my notes I have a general section for mates where I adopted Bogidorus' idea of categorizing by participating pieces. This is akin to the challenge zoologists and botanists have in creating their classification schemes. It's a daunting task, given the generality.

But I hope some of the following mates I intend to discuss have patterns that are specific enough to be both useful and memorable. In that sense I'm hoping to help Richard's side of the argument vs. (the) Romelstielz.

I must say that I do find some of the following very useful during my problem solving.

(Eg. - doing blitz I find I am very competitive when Railroad Mates are involved - often beating the blitz avg solve times. Which, for a slow-poke beginner like myself, is a good indication of progress here on CT)

I do agree  with Rombelstielz that some practice is key - especially if the patterns are unfamiliar or involved. For instance, I really love the full-blown Q+B escalator mates (involving several of the "staircases"), but they are so rarely encountered that the utility of their study is small. Until you're playing a game which you must absolutely win and an escalator is waiting to take you to triumph or defeat, that is!

OK, enough banter, let's start with a simple division of mates into open board, edge and corner. This refers to the final position and not the king hunt. Of course it's an obvious division, but I find useful. The open board mates put the king in a 3x3 square, the edge in a 3x2 square, and the corner in a 2x2 square.

Let's start with two of the most useful open board patterns - the queen checking a king from the side, or from a diagonal.


I've shows the lines of attack - but when I scan a board I just see the two escape squares lit up.


Dovetail and Swallowtail Mates

If the Black king were blocked in by his own pieces (e.g. pawns), you get two of the more common and useful mates on CT - the Swallowtail Mate (aka Gueridon Mate) and the Dovetail Mate.


How to remember the difference in names? I just remember that the swallow has the wider tail => non-adjacent escape squares.

In a future post I will return to these mates, giving a few examples from CT problems, and possibility discussing variants. At present I want to discuss just one variant - the Q+B Dovetail Mate.


Q+B Dovetail Mate

This mate is very analogous to a normal Dovetail, except that the two escape squares not covered by the queen, are covered by the bishop. As such, it is a fully dynamic mate, and can occur anywhere on the board, an important consideration. And while it isn't the most common pattern on CT, it is very simple conceptually, and can come in very handy if you know to look for it.

Here is a list of example problems I found in my notes containing Q+B Dovetails:

Code:

#---------------------
# Q+B Dovetail Mates
#---------------------

31750   M-2  
34235   M-2  
35114   M-2
61828   M-2
62653   M-2
73635   M-3 (perhaps the best example)
90573   M-2
93783   M-1 (perhaps the best example)

10391   M-3 (1.Rb8+ Ke7 ( 1...Kd7 2.Qd8# Q+B swallowtail ) 2.Qd8# M-2 Q+B dovetail)
100433  M-3 (1...Qd1+ 2.Kg2 Qg1+ 3.Kf3 Qg4# M-3 Q+B dovetail mate)

40058   M-2 (Almost looks like a Q+B dovetail with 1...Qg4+, but g3-pawn is blocking, allowing king to escape to g2)

And here are a couple of them I think might be good examples.

( ) vs ( )

Date:

Event:

Site:

Round:

Result:

[Event "CT Blitz"] [Site "ChessTempo"] [Date "2013.07.25"] [Round "73635"] [White "interlist"] [Black "chesstempo"] [Result "*"] [WhiteElo "1939"] [BlackElo "1454"] [EventDate "2013.07.25"] [Blitz-stats "1454.1 00:30 301 73.75%"] [Standard-stats "1192 01:18 945 69.63%"] [User-stats "1934.1 (0.00) 0.00"] [Tags "Sacrifice, Exposed King, Attraction"] [FEN "2br3k/4bR1p/p5p1/1p2q1P1/1P1NP3/1BP5/P4Q1P/6K1 w - - 1 1"]  1.Rxh7+ Kxh7 2.Qh4+ Kg7 3.Qh6# *  [Event "CT Blitz"] [Site "ChessTempo"] [Date "2013.07.25"] [Round "93783"] [White "interlist"] [Black "chesstempo"] [Result "*"] [WhiteElo "1939"] [BlackElo "1475"] [EventDate "2013.07.25"] [Blitz-stats "1475.4 00:19 175 68.57%"] [Standard-stats "1260.8 00:58 480 68.54%"] [User-stats "1939 (0.0) 0:00"] [Tags "Exposed King"] [FEN "rR6/5k2/2p3q1/4Qpb1/2PB1Pb1/4P3/r5R1/6K1 w - - 0 1"]  1.Qe8# *

Well, that's enough for now, but I'll be back later with some more.

--interlist

[richard]

Thanks interlist!

Those diagrams are excellent, I wish I could use them for the motif description page, but for consistency reasons, I'll probably stick with the current pgn viewer board methods. The new pgn viewer has some board annotation features ,and I might be able to extend them to allow the richness of your depictions in the future.

After I posted yesterday, I've added descriptions and examples of Escalator, Railroad and Vukovic variations of the Arabian (which to be accurate I think should probably be considered a whole new pattern in its own right, although for now, I've called it a variation, really the mechanics are quite a bit different).

Your examples for the Q+B dovetail are well timed, as I was going to look at that one next.

Regards,
Richard.

[interlist]

Thanks interlist!

Those diagrams are excellent, I wish I could use them for the motif's description page, but for consistency reasons, I'll probably stick with the current pgn viewer board methods. The new pgn viewer has some board annotation features ,and I might be able to extend them to allow the richness of your depictions in the future.

Of course I'm glad you find them useful. They were done with SCID's graphics, with some MS Paint post-production (I sometimes use Photoshop LE too). The SCID features (arrows, colored squares, circles and X's) are pretty good, and show up as pgn comments. Here's a quick recap for the interested:

Coloring squares:      [%mark square color]
Drawing arrows:        [%arrow fromSquare toSquare color]
Numbering squares:   [%draw {0-9},square,color]
Filling squares:         [%draw full,square,color]
Marking square:        [%draw {circle,disk,x,+,-,=,?,!},square,color]

Are you thinking along similar lines?

After I posted yesterday, I've added descriptions and examples of Escalator, Railroad and Vukovic variations of the Arabian (which to be accurate I think should probably be considered a whole new pattern in its own right, although for now, I've called it a variation, really the mechanics are quite a bit different).

That's great - you're quite productive. I'll have to check them out at some point, are they available for public perusal already?

As concerns the V.V. Arabian, I agree - it's a pattern that stands on its own two feet. I consider a more powerful mate, than the corner Arabian mate. I also have the impression that it's much more common. I still intend to do my own write up on it at some point. Right now I want to cover the Q+B edge mate -- but that requires a little setup to properly introduce.

Your examples for the Q+B dovetail are well timed, as I was going to look at that one next.

It was pretty funny that I was actually typing in my post when I got the message advising me to review a new post. 

I had some delays getting started, partly because I needed to collate my notes with all my CT comments. That's what I get for my tendency towards verbosity!

Well, gotta get back to it...

[interlist]

Here's a puzzle with one of my favorite pattern. I think it may be the longest mate utilizing the pattern -

(Black to pre-move -- White to mate)

( ) vs ( )

Date:

Event:

Site:

Round:

Result:

[Event "?"] [Site "?"] [Date "????.??.??"] [Round "?"] [White "?"] [Black "?"] [Result "*"] [FEN "8/8/8/8/8/4kR2/8/K2Q4 w - - 0 1"]  1...Ke4 *

One word of warning - there are a lot of alts that can also be played here. The question is, can you beat Uri's 5 second solve time on this one!


Here's another position - which I cooked up to illustrate a different pattern. This position is a little interesting because White can mate with the pattern no matter who has the first move. If White has the move it's a simple M-1, whereas if Black has the move it's an M-4.

(White to move M-1, Black to move M-4)

( ) vs ( )

Date:

Event:

Site:

Round:

Result:

[Event "?"] [Site "?"] [Date "????.??.??"] [Round "?"] [White "?"] [Black "?"] [Result "*"] [FEN "4k3/8/3Q4/8/8/8/4B3/K7 w - - 0 1"]  1...Kf7  

[richard]

The SCID features (arrows, colored squares, circles and X's) are pretty good, and show up as pgn comments. Here's a quick recap for the interested:

Coloring squares:      [%mark square color]
Drawing arrows:        [%arrow fromSquare toSquare color]
Numbering squares:   [%draw {0-9},square,color]
Filling squares:         [%draw full,square,color]
Marking square:        [%draw {circle,disk,x,+,-,=,?,!},square,color]

Are you thinking along similar lines?

Yes, the new viewer already has similar pgn comment commands for setting the colour of a square, and drawing arrows between two squares. Missing the others though. I try to use chessbase format for these types of PGN extensions where possible, just because it is a bit closer to a quasi-standard than scid, but in theory options for reading and writing scid format would be possible in the future if the demand was high enough. One downside of the new viewer, and one reason why I haven't updated the existing motifs page to use it is that all the new features come at a performance cost, and creating a board for each motif examples (multiple examples in some cases) would take quite a lot of time with the new viewer, and I'd probably need to switch to a 'show board' button if I was to use the new board in its current performance state.

After I posted yesterday, I've added descriptions and examples of Escalator, Railroad and Vukovic variations of the Arabian (which to be accurate I think should probably be considered a whole new pattern in its own right, although for now, I've called it a variation, really the mechanics are quite a bit different).

That's great - you're quite productive. I'll have to check them out at some point, are they available for public perusal already?

Not yet, I'll try to put up the descriptions on a test server so you can look at them before they get pushed live. I don't want to push a partially complete job live, as I wouldn't like poor or incomplete descriptions to cause an initial set of poor tags.

As concerns the V.V. Arabian, I agree - it's a pattern that stands on its own two feet. I consider a more powerful mate, than the corner Arabian mate.

Yes, the fact that it is not restricted to the corner makes it a more useful pattern I think.

I had some delays getting started, partly because I needed to collate my notes with all my CT comments. That's what I get for my tendency towards verbosity!

The volunteer translators have complained more than once about my own verbosity :-) I find it hard to sufficiently describe the mate patterns without excessive verbiage , but hopefully the text and diagrams together will get the idea across.

I'll let you know when I have a URL you can look at the work-in-progress. A few of your problem comments has already come in handy when drafting some of the descriptions. Especially the railroad and elevator.

Regards,
Richard.

[interlist]

(Work in Progress - please check back for semi-final edit)

Edge Mates

Having introduced the topic of mating patterns with an open board mate, I'd like to next discuss a couple of edge mates. Edge mates are much more familiar and common than open board mates - and I'm sure most people know well the two mating positions shown on the right in the following diagram.



I'd like to discuss the mating position on the left, which I call the Q+B Edge Mate, in detail later in this post. As far as I know there is no other standard name for it, or at least, nobody has replied to request for such a name in the various comments on CT where I've asked. So, lacking a better alternative, and since it seems a sensible enough name, let's adopt this name from here on.


Railroad Mate (RRM) - a brief introduction

RRM-diag

RRM-ortho

I've also adopted my naming convention for the other two edge mates shown, where RRM = Railroad Mate is a name for both the technique of driving the king, and the mating position into which it's driven. The position on the left is RRM-diag, that on the right is RRM-ortho.

Originally, I found both of these positions in problem 74862, where the mating pattern alternates between the two configurations on an open board. The naming is a bit bulky, but descriptive. Recognizing the two positions are related becomes quite apparent when examining the king hunt on the open board. The naming convention is then based on the alignment of the queen and rook - if across from each other vs. diagonally aligned - hence ortho and diag (for orthogonal and diagonal).

aka Q+R box mate

Now, there isn't a universal adoption of names here either. Zwisch, for example, called the king hunt a "Ladder Mate" iirc. It's also been adeptly called the "Z-maneuver" by amateur_player. And what I call RRM-diag, I've seen others call Q+R Box Mate (see diagram on left). But, since these are my posts, I'll stick to my terminology from here on. No matter what you call them, both of these mating final positions are extremely common on CT and impossible not to know well. But probably less well known is the full blown Railroad Mate pattern, though probably more common than you might think. At least that was my experience, once I learned to recognize it. But a proper discussion of the Railroad Mate will have to wait, my main purpose of introducing these final positions was to compare the RRM-diag to the Q+B edge mate.

Q+B Edge Mate

Q+B Edge Mate

RRM-diag

The Q+B Edge Mate is shown on the left, in comparison to the RRM-diag on the right. It can be seen that the queen plays the exact same role in each, whereas the rook and bishop each play their analogous role of delivering check and attacking the f7 square. Both are edge mates where only two pieces are involved and no blocking pieces are needed.

But unlike the RRM mate, the Q+B edge mate must be delivered from afar. In fact, the bishop can be anywhere on the diagonal. This mate is much more common than you might expect, partially because the bishop often swoops in unexpectedly in the middle of much traffic on the board. The criss-crossing diagonals of the queen and bishop should be noted, giving this mate a little Boden-like flavor.

Of course the mate can occur on any edge of the board, but it's almost always see on the opponent's home edge. One of the signatures to help spot this mate is when your queen is on the 3rd row from the edge.

This is a first draft of the post - I certainly need to edit the above. And I'll come back with some examples. But I want to save this work now and will be back soon (I hope!).

Examples:

Code:

504     M-1  1526.1 00:19 720 71.25% (Exposed King, Coercion)
9231    M-2  1567.9 00:30 914 67.40% (Exposed King, Coercion)
12773   M-2  1519.2 00:24 663 66.06% (Exposed King)
41952   M-2  1514.5 00:29 565 67.08% (Sacrifice, Clearance, Exposed King)
42524   M-2  1499.6 00:32 519 69.17% (Exposed King, Coercion (+3, -0))
45455   M-2  1507.8 00:34 803 70.61% (Exposed King)
46786   M-2  1520.9 00:23 524 66.98% (Exposed King, Clearance, Sacrifice)
54971   M-4  1773.1 01:24 724 63.81% (Exposed King, Coercion)  [Good example]
54974   M-2  1595.3 00:28 842 67.93% (Exposed King, Coercion)
56436   M-3  1554.2 00:47 745 72.48% (Exposed King, Coercion)
59335   M-2  1555.2 00:30 633 67.61% (Distraction, Advanced Pawn, Exposed King, Sacrifice (+3, -2))
59709   M-2  1522.0 00:20 671 70.94% (Sacrifice, Exposed King, Clearance, Distraction)
61387   M-2  1507.3 00:27 702 71.51% (Coercion, Exposed King)
61939   M-3  1393.9 00:25 372 77.96% (Exposed King)
63352   M-2  1598.7 00:19 696 67.53% (Sacrifice, Clearance, Exposed King, Attraction)
72992   M-3  1627.5 00:29 613 73.08% (Exposed King, Coercion)
75630   M-3  1611.2 00:38 340 64.41% (Sacrifice, Attraction, Exposed King) Q+B edge mate w R sac
76234   M-3  1716.0 00:50 402 68.41% (Advanced Pawn, Exposed King, Attraction)
76391   M-3  1674.0 00:44 597 71.19% (Exposed King)
77067   M-5  1736.2 01:13 341 70.97% (Sacrifice, Exposed King, Attraction) Q+B edge mate w R sac
77078   M-4  1472.5 01:04 265 83.77% (Exposed King, Coercion (+2, -0))(agree with drahacikfm's comments "Nice maneuvering with queen and bishop.")
78078   M-3  1520.8 00:46 559 76.92% (Sacrifice, Exposed King)
80357   M-5  1620.0 00:56 330 73.64% (Sacrifice, Exposed King, Coercion) (A modified Q+B edge mate)
93276   M-2  1500.9 00:21 195 75.38% (Exposed King)