Duo to the Death

Blokus has taken over my life

Archive for June, 2009

Studies in backstepped F #5

Posted by nand on June 23, 2009

Continuing our examination of orange X-up responses (which, again, I think aren’t all that good, but can still be very tricky for violet to win if orange is savvy)…. here’s a look at the main line for orange and violet.

Tricky but not hopeless for orange

Tricky but not hopeless for orange

The main problem with this particular opening is that it’s susceptible to “quasi-mirroring”: viz., orange keeps playing roughly “equivalent” moves (if violet plays a long hanging piece, then orange plays a long hanging piece; if violet plays a kiss move, then orange plays a kiss move), and even if they aren’t exact mirrors the results are usually nearly a tie. So violet must find a way to (1) create a decisive advantage and (2) do it before orange gets within striking distance of the vulnerable left bottom corner of violet’s N (if orange is permitted to cut that off, then orange obtains most or all of the left bottom corner, which probably means an orange win).

This game shows one way of trying to stop the circling & create a violet advantage by pushing orange away from the vulnarable N while maintaining violet attacks on both fronts. Mike Yosuke first showed me the violet L5 move here. Its power is that orange cannot BOTH block the right AND attack, since the only move that will cut off the violet L5 is orange I5 (and to do so would be to permit violet to attack on the left on the following turn).

That said, violet has a very precarious position on the bottom, as the nicer long pieces (Y, N, L5) are now all gone and violet has few felicitous ways to defend the bottom while maintaining control over the bottom left. As you can see from the following game, simply ceding the bottom left corner results in a tie:

64-64 tie. Violet I4 is obviously too cautious.

64-64 tie. Violet I4 is obviously too cautious.

Moves:
1. F – X
2. X – W
3. N – Y
4. L5 – N
5. Y – L5

Violet Y may be turned with the knob on the left rather than the right. If this is done, then I think orange must play I4 or I5–the key is to NOT kiss, as it’s important not to give violet ANY easy way to grab the bottom left corner.

6. W – Z5

Orange Z5 here is the maximum headache I could think of for violet: it makes a credible threat (to block off the bottom left corner & get a paw on violet’s N) and cannot be blocked by any single move with the exception of violet I4. (Violet I5 would give orange a 1-T4-I3 series of holes: disastrous.) — All that said, it’s not entirely clear to me that violet need block it, rather than blocking off the top. But for the sake of this analysis let’s look at what happens if violet blocks….

7. I4 – F

NOT T5 or anything else that could be blocked with a piece that would also simultaneously dispose of the leak off orange X. Orange needs to keep this leak in play….

8. T5 – 1
9. Z5 – I2
10. I3 – V5
11. P, and the rest is packing.

The result: a 64-64 tie. Quite nice for orange. So, the question is: can violet do better by ignoring the orange Z5 & proceeding with the attack on top? Tune in to the next episode to find out….

Advertisements

Posted in Oldmanc backstepped F, openings | 2 Comments »

Studies in backstepped F #4

Posted by nand on June 23, 2009

I’ve been trying to be thorough with this opening, as I wish to leave no avenue unexplored. So I’m starting with what I consider to be the weakest lines first, where orange plays X up and then plays below rather than wrapping above. (In general I think that all the wrap-below options are bad for orange.) Here is what happens if the orange line I discussed in the previous post in this series — X-Y-N — is not immediately blocked by violet.

Violet 61 - Orange 56. This is cautious play for violet yet it's a 5-point win.

Violet 61 - Orange 56. This is cautious play for violet yet it's a 5-point win.

Moves:

1. F – X
2. X – Y
3. N – N
4. Z5 – W

Two notes here. 1) I picked orange Z5 because it’s the most difficult move for violet to block (because in the following move, as we’ll see, violet decides to block, so I wanted to make life difficult for him) and because I think orange F is more useful for defense on the bottom. If someone can show me a better move here or in the following steps I’d be interested to know.

2) Violet W is key: violet Z5 looks as good or better, but it’s actually a terrible mistake. As we’ll see, we need the Z5 for the block on top, & violet W on the bottom also keeps vulnerable corners away from orange.

5. Z5 – L5

I wanted to see what happens if violet plays defensively from this point onwards. Surprisingly, the answer is that — even though it looks like the board is split nearly evenly — violet has a quite decent win in his pocket. The key from this point is for violet to hang on to the Y, as the Z5-Y combo is what seals the top (with L4, T4, or V3 equally good for capping it off)

6. L4 – F
7. T5 – I2
8. Y – I1
9. V3 – W and the rest is packing.

Final score 61-56 for violet — pretty good showing. I may have missed a slightly better packing.

Unless someone can show me a variant line I will assume that the X-Y below line for orange is a washout, & move on to other lines.

Posted in Oldmanc backstepped F, openings | Tagged: | 1 Comment »

Studies in backstepped F #3

Posted by nand on June 19, 2009

In my previous post I’d started to explore what happens if orange plays X-up then circles beneath rather than above. My general feeling is that this is bad for orange but it cannot be instantly ruled out. Orange L5 on move 3 is a bad response, as we saw, since violet can block it with impunity & win handsomely. Orange N, on the other hand, cannot be blocked without violet getting into serious trouble. With ideal packing on both sides, I believe violet still wins by 1 point: here’s my packing:

Violet 61, Orange 60

Violet 61, Orange 60


If there’s an improved packing for either side I’d like to know, as it obviously makes a huge difference in a game this close to a tie.

Moves:

1. F – X
2. X – Y
3. N – N
4. W – Z5
5. P – 1
6. L5 – Z4
7. 1 – V3
8. I2 – L4 and the rest is packing

Obviously on move 4 violet must ignore orange N & instead wrap around the right with W or Z5. I think those lines do not obviously favour either side, but need to investigate further.

The other question here is whether violet N on move 3 is good. I can’t see any other obvious move, but I may be wrong.

Posted in Oldmanc backstepped F, openings | Tagged: | Leave a Comment »

Studies in backstepped F #2

Posted by nand on June 18, 2009

This is the opening I’m currently exploring a lot, as I find it very slippery. My basic feeling is that it’s not a very good opening once analyzed carefully, but it’s got enough complexity to it that it can really throw some players for a loop (including me).

My general feeling is that orange’s worst option is X up, & that trying to go underneath violet X on move 2 rather than over the top is also bad. But, again, it’s never exactly straightforward with this opening. Anyway, I’m going to pick away at it slowly, so let’s start with lines that are obviously bad for orange, before I get to ones that may favour orange.

Final score violet 65, orange 58

Final score violet 65, orange 58

Moves:

1. F – X
2. X – Y
3. N – L5

I’m not sure that violet N is good, but I’m very sure that orange L5 is a bad response to it. At this point violet has no particular reason to continue circling: he can take a stand & count on grabbing enough points to win.

4. Z5 – Z5

Violet Z5 rather than T5: normally I’d want the T5 because of the extra corner in orange’s area, but in this case (1) this is going to be a packing game and violet needs all the extra space and corners within his own area that he can get, and (2) the Z5 is plenty threatening on its own, as orange has no clean way to block it without creating further leaks.

5. P – 1
6. L5 – P

This is my best solution to the problem of how to staunch the leak off violet Z5. If violet plays I3 or I4, then orange I4 can cut across it easily.

7. I5 – U … and the rest is packing.

Possibly there is a better solution to the packing for orange; but I’m sure that the violet packing is optimal.

So: a 7-point loss. Clearly a bad line for orange. Next question is whether it’s stronger if orange L5 is replaced with a different move (like Y or N or W)…

Posted in Oldmanc backstepped F, openings | Tagged: | 1 Comment »

Achilles heels #3

Posted by nand on June 10, 2009

One of the most popular openings is the XF kiss. People often call it the “TT” (Tetris-Time) after a player who developed it a lot (not sure if it’s original to him). Most recently it has been reexamined fruitfully by mike_yosuke. Some people tell me it’s only a 1-point win if orange plays perfectly, but I have to inspect the lines myself before I’ll vouch for that.

Fairly early on it became known that the best reply by orange is to block off the more dangerous (hanging) corner of the X with the N. On the whole I think this should be played with the 3 side DOWN. Violet’s best response is to hang an N around the N in return (a line I’ll discuss in a different series of posts, as it’s actually a pretty solid opening), but a lot of beginning and mid-level players are too cautious and play a W to kiss it, which is pretty weak if orange knows what to do. Here’s a typical game with an orange victory.

Violet 57 - Orange 68. One example of why move 3 violet W is bad in this opening.

Violet 57 - Orange 68. One example of why move 3 violet W is bad in this opening.

This isn’t the best line for violet, but I’m more interested here in showing “typical” play. If you’re a beginning player you can learn a lot from this game about how to close up apparently dangerous leaks: the I3 side on orange’s N is far less vulnerable than it appears, as violet learns to his cost.

The moves:

1. X – X
2. F – N
3. W (bad!) – Y
4. Y – Z5

note: NOT F, it’s important to save it for the bottom. And not T5, as it’s important that orange have two corners in violet’s area which cannot be simultaneously blocked by a single violet move.

Violet has by now let orange secure the bottom right corner; & orange is also now impossible to stop from grabbing some of the top right. (Violet’s Y should be turned with the knob on the left, but I’m not sure this is sufficient to save the game for violet.) From here on out, various lines are possible, none of them favouring violet.

5. Z5 – F
6. N – L5

VERY important to create the one-square “pocket” here (4 up, 4 across from the bottom corner) which will eventually stop the leak around orange N.

7. T4 – P
8. 1 – I3
9. I3 – W … and the rest of the game is just packing.

Note that the beauty of the orange N in this orientation is that if violet decides to plunge ahead with an attack rather than defend, the only way he can get very far is by wrapping L5 around it…. giving orange lots of nice long sides to work with (& meanwhile orange’s smaller area is eminently defensible with all its jagged corners).

Posted in flawed openings, openings, XF Tetris-Time | Tagged: | 3 Comments »

Studies in turned F #1

Posted by nand on June 9, 2009

Anton (Gerenuk) requested that I post something on the turned-F opener. I haven’t got a handy nickname for it; the players who currently use it most are BIGWOOD and Bobby_Bob, but I’m not sure whether either of them can be said to have first dibs on it. (Any suggestions for a name are welcome!)

UPDATE: Bobby_Bob tells me it’s BIGWOOD who laid down most of the groundwork for this opening. So perhaps it should be indeed called the Bigwood F!

There are two varieties of this opener, depending on whether the F’s 3-bar (the row of 3 squares) is horizontal or vertical. Here is a typical game with a horizontal orientation for the 3-bar, where orange replies with X-up. I’m not sure that that is necessarily a bad response, but certainly orange must be wary. The T5 below strikes me as a very poor response.

Violet 74 - Orange 61

Violet 74 - Orange 61

I didn’t witness this game but here is the order of moves I infer:

1. F – X
2. Z5 – T5 (this is a mistake: U or P would be better)
3. Y – Y
4. L5 – F
5. T5 – V5
6. W – L5
7. N – W
8. P – U
9. I5 – I5 (very bad move)
10. L4 and the rest of the game is obvious

Needless to say, this is a risky opener, not least because the other lines if orange DOESN’T play X-up are far more ambiguous.

Posted in Bigwood F, openings | Tagged: | 1 Comment »

Half-open vs open

Posted by nand on June 6, 2009

I like developing terminology for games. (It’s always a kick when people start using terms you devised. I was pleased to see Will Shortz referring to “cages” in a puzzle recently — it’s a term I coined for killer sudoku a few years ago, & seems to have stuck.) & more importantly, having terms for things can advance your understanding of a game. (I will post at some point about the idea of the “semi-mirror” for instance.) So: I’d like to refine the usual classification of openings into “open” and “closed” into a three-way distinction.

  1. “closed”: the most popular kind: violet’s 2nd move creates a 1-square hole (what I call an “eye”). XF Rubik and Yasu, discussed in previous posts, are examples. (POTENTIAL VARIANT: violet creates an i2 hole or larger. I have yet to see an opening that does this that is very effective.)
  2. “open”: violet creates no eye, offering orange the chance to wrap around on move 2 and actually pass through on move 3.
  3. “half-open”: violet creates no eye, but positions the 2nd piece in such a fashion that orange cannot immediately penetrate; it will take until move 4 to actually pass through violet’s piece-formation.

Here are four examples of half-open openings.

Note that orange cannot play directly into the central hole on move 2; it will take 3 moves to reach it, and (if orange wishes to do so, & violet does not obstruct him) a 4th move to actually pass through the hole. — Also note that in all these openings it is very difficult for violet BOTH to attack AND to defend the hole, because it’s very easy for violet to end up in a situation where the 1 piece is the only way to block the central leak.

My general instinct is that this is a fairly dodgy way for violet to open, but having lost plenty of games to these as orange I cannot say this for sure yet. The basic question with all of these is whether orange should play on the left or right on move 2. Typically on the right the response is Y upright, threatening to either wrap round the top or play Z5 into the central hole; typically on the left the response is W or F wrapping around, planning to put a piece in the central hole on move 3. Further investigations pending…

Posted in half-open openings, openings | Leave a Comment »

Studies in XF Yasu #1

Posted by nand on June 5, 2009

This is another XF opening (again, named after a player who developed it) that’s effective but a little weaker (in my experience) than the Rubik XF. Here’s one good line for orange — not a guarantee of a win but a guarantee of a tough game for violet. It is the handiwork of Bobby_Bob.

In my view violet N is a mistake. Probably blocking the orange Y is best for move 3 (with L5, usually).

In my view violet N is a mistake. Probably blocking the orange Y is best for move 3 (with L5, usually).

The key here is that there is no single move by violet that will block the attack off orange W. So orange is free to block on the next turn if violet continues circling off the Z5.

Best tactics for violet, I think, are blocking orange W with L5, which forces orange to loop L5 around the top of it in return; then violet has a golden opportunity to kill all violet’s corners in top left & forge ahead with an attack on the top side of the board.

Posted in openings, XF Yasu | Tagged: | 4 Comments »

Studies in XF Rubik #1

Posted by nand on June 2, 2009

This is one of the most boring, because most effective, of openings: violet can usually win, though not usually by a lot if orange knows the lines. It’s often referred to as the Rubik after Rubik87, a player who spent a lot of time working this one out (with me frequently as orange!). — Anyway, here’s one line.

I think this is perfect play on both sides... improvements welcome

Violet 65 - Orange 63. I think this is perfect (but cautious) play on both sides... improvements welcome


Moves:

1. X – X
2. F – W
3. Y – Y
4. P – I5
5. L5 – T5
6. I4 – L5
7. T4 – O4
8. I5 – P and the rest is packing

[I originally misordered the moves, hence Rubik’s comment on I3 below: the above is the correct version]

Note that it is essential that orange play W, not F or Z5 (or W flipped around). Otherwise violet is not forced to play I4 to prevent the potential leak off W — instead he can play U, & the packing favours violet by a wide margin.

The main variants here come with violet’s P, which is also sometimes W (hanging) or N (kissing). I’ll post more about these later.

The great thing about this line, by the way, is that as far as I can see violet’s moves are virtually all forced, with the exception of the P. Once the P is played, there is nothing violet can do. A 2-point win is still a win, but it’s so small (given that violet has a huge advantage by going first) that most experienced players consider it a tie.

Posted in openings, XF Rubik | Tagged: | 5 Comments »

Achilles heels #2

Posted by nand on June 1, 2009

This is a favourite of newbies, but it’s not very strong versus a player who knows the trick. Basically, despite the obvious temptation for orange to play on the bottom on move 2, DON’T DO IT. Be more patient, & set up the nice 1-3-2 series of holes off the central N. This is my invariable response when violet tries to circle round the bottom, & if it’s played right then it usually wins.

A weak opening... note the 1-3-2 series of sides available to orange in the centre

A weak opening... note the 1-3-2 series of sides available to orange in the centre

Note that orange’s main concern here is that if violet’s attack persists below, then if orange keeps circling above he will start to mess up his own trap laid for violet (the 1-3-2 series) by moving into that area. It’s essential to prevent this from happening by keeping a distance between the attacks & by blocking violet’s progress below before it gets too serious. And it’s important not to give violet the opportunity to block the leak in the centre before orange can play the 1.

If violet blocks on top rather than playing the Z5 this can be trickier for orange, but usually at worst it’s a 1-point loss for orange.

Posted in flawed openings, openings, XN in centre | Leave a Comment »