Catchup: Wrong result recorded, game rules not 100% correct.

Game #1656720 was recorded as a win for my opponent (molnar), when it actually should have gone to me (it went down to the fifth tiebreaker). Since this was part of a large tournament I'd like the result to be fixed, if possible.

Additionally, Catchup's rules allow for players to place 1 or 2 stones per turn (1, 2, or 3 on "catchup" turns), but there's no way to do that here. I haven't encountered a situation where placing fewer stones makes sense, but it might be possible in very rare endgame situations.

I hope someone in charge reads this :P

But what a game, wow!  ;)

Indeed, a prototype position is

<br/> . .<br/>. . .<br/> . B<br/>

where the two winning moves for O require to not put two stones:

<br/> O .<br/>. . .<br/> . B<br/>

<br/> . .<br/>. O .<br/> . B<br/>

These positions don't seem like winning for O, with a next move like this:

 . B<br/>B O .<br/> . B

 O B<br/>B . .<br/> . B

Rather, O should play this:

 O .<br/>. . .<br/> O B

Thanks for pocking this up. You counter examples require B to put less than 3 stones despite O increased the size of the largest group wich is not allowed afaik.

For example we have

<br/> B .       B .<br/>. O B  => O O B<br/> B .       B O<br/>

and B can not move because there is only one free field but B is supposed to put three stones. (Or is not?? Is B allowed to put just one?)

The same applies to the other example:

<br/> O .       O .<br/>. . .  => B B .<br/> O B       O B<br/>

and O can not move because there are only two free fields left.

Players are never forced to place more than one stone per turn.

@Luis: Is this mentioned in the rules somewhere?

The rules then would read:

In every turn a player is allowed to put _at least_ one and _at most_ $n$ stones where $n$ is one in the very first move and either two or three in subsequent moves, depending on whether or not the size of the largest group changed in the previous move.

@Richard: How is it implemented on LG?

E.g. the rules seems to be like

<br/>let new_board = puts fields board<br/>let csize = maximum (size_of_components new_board fields)<br/>available_stones = if csize > high_water board then [1,2,3] else [1,2]<br/>

<br/><br/>but on LG it seems to be like<br/>

<br/>available_stones = if csize > high_water board then [3] else [2]<br/>

<br/>

Well, I had assumed, as Luis, that a player is never compelled to play more than one stone. I think it is explicitly stated in the rules... I just checked them on Nick Bentley's blog:

'(...) must drop 1 or 2 stones on her turn (...) your opponent may drop up to 3 stones on her next turn (...)'

Additionally, the iOS app, which Nick was heavily involved with, lets you choose not to play all your stones, though it kindly warns you that is what you're about to do.

Playing 1-2 stones less on a turn isn't going to come up except in some very rare scenarios (although it could be the difference between victory and defeat).

The bigger issue here is that there's a bug in the code that gives a victory to the losing player if you go far enough down the tiebreak chart.

You are right, that one is the real issue and needs a fix.

By playing I found that on LG you are forced to put two or three stones. However, I had no chance to see what happens when the number of free fields is less than the number of stones a player is supposed to put. Is the move possible _then_?

Yes, you complete the game by playing a single stone: http://littlegolem.net/jsp/game/game.jsp?gid=1655741.

Coincidentally I was just wanting to place a single stone in a game (or at least considering doing so, until I found that the user interface wouldn't let me), in order to avoid being the first  to create a 2-stone group. So I came to the forum to see if this bug was already reported. :)

One can actually play 1 stone (or 1 or 2 when 3 are allowed) as can be seen in this game, move 2: http://www.littlegolem.net/jsp/game/game.jsp?gid=1658767&nmove=2

Not sure if Richard did this intentionally but hasn't implemented the interface yet or it's a side effect.

If anyone wants to place less than 2 or 3 moves, the procedure is: click on the place you want the stone to be placed and click on another random empty cell. In the next page do not click Send move but manually edit the url.

If it ends with "...move=eaff" and you want to move only into the E1 (ea) cell, edit the url to "...move=eaea" and click enter.

In the following page, the wanted move will be displayed and the Send move button will work as expected. If it isn't what you want, click Cancel move and repeat the procedure.

With Luis' rules (optional at mindsports.nl: http://mindsports.nl/index.php/the-pit/620-catchup) the above confusion is a non-issue.

@hyper: Cool. (And obvious in some sense.) I wonder whether or not it will work in the other direction too, to put _more_ stones?

No, it will not.

Christian: I like those rules, but it's not really Catchup that way, simply because the game is slower. This isn't even addressing the fact that it encourages sandwiches (probably the strongest formation) even more because 2:1 is a bigger material ratio than 3:2.

I don't understand what "really" adds to your statement and in any case, I'm not addressing the question whether or not it is "really Catchup". Luis rules are a simpler implementation of introducing negative feedback in the move protocol, as was Nick's original goal. It makes for a different game, not because the nature of the dilemma changes, but rather the timing of the decision point. So it immediately has a different feel. Now if one is very used to the current implementation, this might be felt as an unpleasant surprise. The same might be the case if Luis' rules had been the standard and someone had come up with Nick's. In that case, an additional argument to reject it might have been that it complicates matters and is less aesthetically pleasing.

Christian: not disagreeing, just saying it's definitely not the same game. I'd be willing to give it a shot at some point.

You can challenge me at mindsports.nl (and at the risk of stating the obvious, I prefer Luis' rules :)

If I can weigh in with my opinion here.... without any expertise with either rule set — the original rules by Nick Bentley provide greater opportunity for placing more than one stone on a move, as I understand it, than Luis' alternative rules. In my opinion, the trend for new games to permit multiple moves on a turn is a good one — for example, Tzaar, Yodd, and Arimaa.

Kerry: There are usually around seven 3-stone turns in a game of Catchup, and the average number of turns per game is 13-14 per player (including dead turns). The variant rule does make the game longer, probably 17-18 per player, and tactics are less dynamic. There are still typically two stones placed per turn,

@Kerry: Where Kris certainly considers trends and marketing issues (and thus is handicapped by the implicit need for non-generic material), I don't. Nor does Luis, so far as I can tell. If I recall correctly, Nick doesn't particularly like multiple moves in games, giving it as the reason that he wasn't particularly drawn to Symple. Also, Nick considers no implementation as final because there's "always room for improvement". So if improvement suggests itself, I'm sure he will consider it.Nick's original idea regarding Catchup was to introduce negative feedback into the placement protocol. Combined with the object of having the largest group, "cascading" down to the next largest in case of equal scores, this suggests a simple protocol: give negative feedback to placements that are explicitly aimed at this goal. Such placements are those that result in a group becoming larger than the largest group on the board was, at the outset of the turn.---Here's Nick's implementation of such a protocol:On a turn, a player must  perform exactly one of the following actions:

• Place 1 stone if the board is empty.
  
• Place 1 or 2 stones on empty cells of the board.
  
• Drop 1, 2 or 3 stones on empty cells of the board if, and only if, his opponent on his last turn has made a move that resulted in a group larger than the largest group that existed at the outset of that turn.
  

Here's Luis' implementation of such a protocol:On a turn, a player must  perform exactly one of the following actions:

• Place one stone on an empty cell of the board.
  
• Place two stones on empty cells of the board, in such a way that the size of the largest group on the board at the end of the turn is the same as it was at the outset of the turn.
  

---Note that in Nick's version the negative feedback follows the act of enlarging, while in Luis' version it is build into the act itself.Note that Nick's version requires an extra rule for white's first move, while in Luis' version the restiction to 1 placement is implicit.Note that in Luis' version the board position is always a complete description of the current game state.---@Metz: I'm not convinced that the option to place more stones inherently leads to more dynamic tactics. Go's tactics for instance are pretty dynamic with one placement per turn.

@Christian, thank you for the feedback and explanation. I understand and agree with your reasoning about the elegance of the rule set in Luis' variant. Whether it actually leads to a better game, I am unequipped to judge. I am not so clear about your reference to Tzaar with respect to trends and marketing. I see the option to make multiple moves in a turn as an interesting  theme in abstract game design.

@Kerry: So do I, actually, and may I suggest  you try Symple to let me substantiate this view?

"Whether it actually leads to a better game, I am unequipped to judge."  @Kerry: In terms of experience with gameplay I might say the same. But I deeply trust concepts where object and mechanics embrace each other in such a perfect fit. And I never quite trusted Catchup despite the concensus that it is a great game. That's why it currently resides in the Pit rather than in the ArenA. That might change if Nick reconsiders and comes to the same conclusion. And in terms of gameplay, where could it go wrong? It certainly doesn't feel as if turn order balance could be a decisive factor.

We did it again.

The same happende in game 1656703

 Game #1656703

I mentioned this to Richard and he has corrected the result of this game.

A game of catchup where it was of advantage to place 1 instead of 2 stones: game 1658767

nice!

@ypercube, you placed a single stone twice, including the second move. Why did you feel it was an advantage that early in the game?

Indeed, I see why the single stone placement was useful on turn 26, but not on turn 2. Why the turn 2 singleton?

Based on ybercube's comment at 2014-09-10, I'd guess that it was an experiment with the UI to determine whether it was possible.

I'm impressed that he was able to win the game despite this "handicap".

Of course, I could be wrong and it could be that ypercube was playing a very deep zugzwang....

What if Klaashaas played 23.A9A9? (lower left corner 1-stone move)

Dang, Carroll, now you're making all of us think!

I don't know if/when I'll have time to analyze that, but it sounds like fun....

Sorry for the confusion. The zugzwang move was in 26 (and could have been klaashas' win if he had played 23.A1 as @Carroll noticed.

The move 2 was just a test to see if such a move was possible / implemented here. It was a  that it came helpful in the same game.

(Dang, of course I meant 23.A9 (not A1) or any othe single stone move that would not extend the 16-stone snake.

Bump

the mistake in our game still stands. ( metzgerism will win the tournament regardless the mistake)