Rules for start

Chinese rules and all the many variations.

Moderator: Shirluban

iandstanley
Senior Reacher
Senior Reacher
Posts: 138
Joined: Mon May 18, 2009 11:20 am

Re:Rules for start

Post by iandstanley » Fri Jun 12, 2009 2:11 pm

iandstanley wrote:With MCR is has a set of 81 fans (scoring elements which define:

1. the complete hand (e.g. 13 orphans)
2. part of the hand (e.g. 3 kongs)
3. attributes of the tiles in the hand without determining the type of set (e.g. all green; upper tiles; Upper Four)
I nice quick reference I\'ve come across for MCR is here http://www.ninedragons.com/mahjong/COMJquicksheet.pdf

iandstanley
Senior Reacher
Senior Reacher
Posts: 138
Joined: Mon May 18, 2009 11:20 am

Re:Rules for start

Post by iandstanley » Fri Jun 12, 2009 4:37 pm

Tom Sloper wrote:
iandstanley wrote:MCR ... hands are slightly imbalanced but not as bad as WSoM ...
I\'d be very interested in hearing how you arrived at that conclusion, Ian. The concept of balance is very interesting to me.
I only think it is slightly imbalanced ... the rules after all have a lot of tournament level play testing. I quite like the rules overall and they do come across as very well thought out

A number of other players have mentioned on this board that they consider one hand slightly easier to get than another.

My main qualm is over the knitted hands. One of the most confused features of MCR is the knitted sequences 147-258-369 which seems to me very overrated in it\'s points value twelve points! the logic behind having the hand in the available combinations seems to be this guy has got such a bad starting hand we\'ve got to give him a chance of winning.

Compare the 12 point knitted sequence of 1-bam 2-crak 3-dot 4-bam 5-crak 6-dot 7-bam 8-crak 9-dot to the following elements:

two dragons (6 pts) green-dragreen-dragreen-dra
big three winds (12 pts) :east:east:east :south:south:south
Mixed Triple Chows (8 pts) 3-bam4-bam5-bam 3-crak4-crak5-crak 3-dot4-dot5-dot
2 concealed kongs (6 pts) 3-bam3-bam3-bam3-bam 6-dot6-dot6-dot6-dot
Mixed shited pungs (8 pts) 3-bam3-bam3-bam 4-dot4-dot4-dot 5-crak5-crak5-crak
Upper & Lower fours (12 points) limiting your hand to the upper and lower tiles


I realise that it has had a lot of playtesting at tournament level but still the knitted hands which score 12 or 24 points seem a little overrated. Yes allow the hand element but it ought to be closer to 8 points as it is barely above the chicken hand.

User avatar
Tom Sloper
Moderator
Moderator
Posts: 297
Joined: Sat Jan 17, 2009 8:01 pm
Location: Los Angeles, CA
Contact:

Re:Rules for start

Post by Tom Sloper » Fri Jun 12, 2009 4:45 pm

iandstanley wrote:
Tom Sloper wrote:
iandstanley wrote:MCR ... hands are slightly imbalanced but not as bad as WSoM ...
I\'d be very interested in hearing how you arrived at that conclusion, Ian. The concept of balance is very interesting to me.
I only think it is slightly imbalanced
WSOM? You said WSOM was more imbalanced than MCR. I\'m very interested to hear how WSOM is so badly imbalanced.
4649おねがいします。

iandstanley
Senior Reacher
Senior Reacher
Posts: 138
Joined: Mon May 18, 2009 11:20 am

Re:Rules for start

Post by iandstanley » Fri Jun 12, 2009 10:39 pm

WSoM Balancing

Compare the following hands:

1. Thirteen Terminals 160 pts (reach thirteen orphans - limit)
2. Big three dragons 130+30 pts (reach - limit)
3. Mixed Triple Run 120 pts (reach 1 han)
4. 7 pairs 30 pts (reach 2 han)
5. small three dragons 40+20 pts (reach 2 han)

These hands appear to be roughly equal if we consider a limit somewhere between 120-160pts (if we compare with 13 terminals as being limit). But the limit in WSoM is 320 which highlights some oddities

And yet consider pure greater terminals 400 pts (reach all terminals = 1x limit) which is apparently harder to achieve in WSoM than in Riichi whereas other similarly difficult hands in riichi are similarly difficult in WSoM

All terminals in WSoM is 3x harder to achieve than 13 orphans and yet they are comparable in riichi!




Consider these hands

3-dot3-dot3-dot 4-dot4-dot4-dot 5-dot5-dot5-dot
with some other non scoring parts


You can take this hand as
3-dot4-dot5-dot 3-dot4-dot5-dot 3-dot4-dot5-dot
as Three Identical Sequences for 120 points OR as 3-dot3-dot3-dot 4-dot4-dot4-dot 5-dot5-dot5-dot for Three Consecutive Triplets for 100 pts

You may say that is down to the difficulty of drawing the hands the drawing identical sequences being 20% harder because you can only get your tiles self-drawn or fromt he player on your left whereas the 3 consecutive triplets can be self-drawn or ponned from all the other players. That is fair enough and I would not argue that either as this is the basis of all mahjong scoring mechanisms.



A similar part hand of

3-dot3-dot3-dot3-dot 4-dot4-dot4-dot4-dot 5-dot5-dot5-dot5-dot

does not exhibit the same problem .



You can take this hand as as

3-dot4-dot5-dot 3-dot4-dot5-dot 3-dot4-dot5-dot 3-dot4-dot5-dot

as four Identical Sequences for 480 points or as

3-dot3-dot3-dot3-dot 4-dot4-dot4-dot4-dot 5-dot5-dot5-dot5-dot

for Four Kongs for 480 pts



If the rationale for the relative difference of obtaining the sets (i.e. the basis on which all mahjong scoring exists) three identical sequences vs three consecutive sequences is the difference in drawing tiles restricted (ie self-draw/left-player vs all players/self-draw) then the same rationale has to apply to how the fourth identical set of tiles and therefore the latter two hands should differ in scoring!

It is clear here that the scoring mechanism is unbalanced ... if 9 tiles drawn using method #1 is harder than 9 tiles drawn using method #2 then it must follow (as the methods remain unchanged) that 12 tiles drawn using method #1 is harder than 12 tiles drawn using method #2

WSoM fails on this basic statistical issue. I understand that the scores are additive and the limits would mitigate some of this effect but the basic components seem slightly off.


The above example was a rather contrived scoring on part hands

In my opinion Riichi is better balanced by being less prescriptive in the scoring mechanisms by scoring hand points against defined doubles. The removal of the limit mechanism in MCR smooths out what I see as flaws in the relative points that they prescribe. WSoM takes a different approach by implementing a compound limit of 320 limits the manner in which you can add components

Yes I know you are going to say but so does riichi. But riichi counts the arithmetically progressive minipoints and then doubles up .. the fan tables are constructed such that these anomolies are less pronounced.

In summary, all versions of Mahjong have their problems but Riichi and MCR (taking different approaches) have solved the problem of artificial limits. MCR by removing the limit and reach by carefully constructing a arithmetic progression of limits.

User avatar
Tom Sloper
Moderator
Moderator
Posts: 297
Joined: Sat Jan 17, 2009 8:01 pm
Location: Los Angeles, CA
Contact:

Re:Rules for start

Post by Tom Sloper » Sat Jun 13, 2009 1:29 am

iandstanley wrote:in the above second example you cannot claim 480+480 but only 480 (a listed limit hand only the highest listed limit counts) whereas you can score 100+120 for the first example
No, you can\'t. They\'re either chows or they\'re pungs. You can\'t claim the tiles as both types of sets.
Most of the time, someone who makes one of these hands would have exposures. The exposures would determine whether the sets are chows or pungs.
Let\'s assume there are no exposures. You still have to declare which types of sets they are. Another example:

2-crak3-crak4-crak2-crak3-crak4-crak2-dot3-dot4-dot2-dot3-dot4-dot4-bam4-bam

2-crak2-crak3-crak3-crak4-crak4-crak2-dot2-dot3-dot3-dot4-dot4-dot4-bam4-bam

This hand could be either seven pairs or all chows. You can\'t declare it as both, and actually collect the points for both. It\'s either a seven-pair hand OR an all-chow hand; not both simultaneously.

If I am wrong about this, please tell me precisely which web page and which paragraph and which sentence of the WSOM website I should go to, to see the rule for myself.
4649おねがいします。

iandstanley
Senior Reacher
Senior Reacher
Posts: 138
Joined: Mon May 18, 2009 11:20 am

Re:Rules for start

Post by iandstanley » Sat Jun 13, 2009 9:28 pm

Touche!

I originally had the additive rules explained to me wrong a while ago, and being quite satisfied with Riichi & MCR I never bothered to check (what was that X-Files phrase ... trust no-one) them with the WSoM site when I made my original assertion.

Serves me right for trying to construct a complex answer in the early hours of the morning. Even more so when I then tried to create a \'more convincing example\' based on my flawed understanding of the rules.


Reviewing the rules you are precisely right in that it works in a very similar if not identical way to MCR. I have removed that part from my original post above (not to hide my shame or I wouldn\'t have made this reply ... but so as not to confuse anybody).

However, I believe the rest of my point over balancing still has some merit.

Looking at the earlier examples of hands particuarly the 7 pairs and the mixed triple pung, the 7pairs hand seems to be penalised for little reason in WSoM compared to MCR and Riichi reducing it down for some unknown reason.

In each set of rules you expect different favouritism to come through to varying extents and modify the statistical expectation. But the end result is that a particular kind of hand approximately performs at a similar kind of level compared to other hands of the same statistical probability no matter which major ruleset you use but WSoM seems to have a few oddities. Maybe they were trying to discourage certain hands from being used in such a major stakes competition .... then why not just drop that scoring element rather than penalise it.

User avatar
Tom Sloper
Moderator
Moderator
Posts: 297
Joined: Sat Jan 17, 2009 8:01 pm
Location: Los Angeles, CA
Contact:

Re:Rules for start

Post by Tom Sloper » Sat Jun 13, 2009 9:43 pm

iandstanley wrote:I believe the rest of my point over balancing still has some merit.

Looking at the earlier examples of hands particuarly the 7 pairs and the mixed triple pung, the 7pairs hand seems to be penalised for little reason in WSoM compared to MCR and Riichi reducing it down for some unknown reason.
I may be misremembering, but I seem to recall Alan Kwan (the creator of the WSoM rules) saying on the mahjong newsgroup once that he wasn\'t particularly fond of 7 pairs hands. I\'m too lazy to go digging around to see if I can find anything like that.
There might be something about that on his website, though. It\'s at http://home.netvigator.com/~tarot/Mahjong/Mahjong.html (Edit: don\'t click it - the forum\'s URL-handling is flawed - copy and paste it to go to the address).

I haven\'t done any analysis of Zung Jung and WSoM - pretty sure WSoM either is ZJ or is based on ZJ.

Anyway, thanks for the additional explanation re balance. I\'m clearer on that now.
Cheers,
Tom
4649おねがいします。

iandstanley
Senior Reacher
Senior Reacher
Posts: 138
Joined: Mon May 18, 2009 11:20 am

Re:Rules for start

Post by iandstanley » Sat Jun 13, 2009 11:05 pm

I\'m too lazy to go digging around to see if I can find anything like that.
And why should you when it\'s my assertion :)

As usual Tom you\'re spot on

It is buried at least more than one layer down as the first page does not link to it (I found it using a google search with the site:.... option)

For those following the conversation it is at

http://home.netvigator.com/~tarot/Mahjong/mj_7pair.txt


For those who haven\'t read the article ...

Alan says regarding 7 pairs \"Clearly, 7 pairs is not as difficult or rare a pattern as many players believe, and it would be inappropriate to assign to it too high a scoring value.\"



He mentions earlier that \"You need \"lucky draws\" to match loose tiles, but you need lucky draws to match loose tiles for All Triplets either. \"

But the lucky draws for a 7 pairs hand are required to be self drawn from the wall (final pair excepting) whilst the All Triplets can be taken from any discard or the wall i.e. 100% of the wall potentially (if nobody else is collecting a chow including them)

** WARNING LATE NIGHT CALCULATION **

Start off with a set of 144 tiles remove the 4 seasons and the extra four tiles (either all red fives, flowers, or four regular fives if playing with red fives) leaves you with a starting set of 136 tiles.

From those 136 tiles you need to remove 14 tiles for the kong box and 4*13 tiles for the other players hands leaving 70 tiles.

For the pung hand this leaves you 70 tiles to potentially pung from the wall (either through self draw or from unwanted discards) plus potentially a further 8 from the players hand if they are held in the original deal .... this leaves you wanting 4 tiles of 16 tiles with 70 to 78 draws/discards

Consider 7 pairs ... you would not consider going for it without at least 4 pairs (You often see advice saying 5 pairs or some very early matches). you are looking to match off those pairs in the 70/4 draws left (yes I know the last one may be ronned) this leaves you with a 7 unmatched pairs. Drawing a tile to match 7 (of a total of 16 types of tiles) .... sorry didn\'t mean to scare anyone ... but to simplify the maths a little lets say that about half the time you are not going to match the tile for the first tile reducing to a tile to match 3 tiles then a single wait with only a quarter of the available

It\'s far too late to derive the maths involved but they do seem very in favour of the pungs having a lot more draws to draw similar tiles.

Alan\'s statement of
marginally more available tiles
for the pung and describing the 7 pairs\' final single wait (3 max available tiles) as an advantage over the two wait options of the all pungs\':

(i) two remaining pairs (4 available tiles) ... most players would obviously choose to be in the situation of having two pairs remaining than a pung and single tile
(ii) single tile awaiting (3 max available tiles) same as the 7 pairs last wait

Obviously this is a very simplistic view... it does not take into account other players collecting pairs, chows, keeping dangerous tiles and so on.

But it does seem to support the view of a harder 7 pairs hand as rewarded by MCR and other rules.


One consideration that Alan is silent upon is that most experienced players would not rule out either hand at the beginning of the game preferring to wait and see if the pung or more pairs appear first. If your existing pairs are certain premium tiles (e.g. winds, dragons and fives if playing red fives in Riichi) then your opponents are less likely to ditch these tiles in your favour preferring to wait a few draws to see if they can use them. In this situation you are more likely to end up matching a 5th pair than a 1st pung which would distort the amount of times 7 pairs comes up in tournament play. According to Alan (and I have no data to argue) 7 pairs comes up around 7% of the time in Japanese tourneys .. there are other considerations here as well as in Japanese rules 7 pairs may also pick up Menzen Tsumo, Tanyoo chuu amoungst other doubles bringing the right 7 pairs to a good total. Also 7 pairs, as you are limited to a single wait, most will declare riichi on a concealed hand as they have to keep their hand concealed until the winning tile anyway. You will avoid declaring riichi if some of the tiles have gone and you wish to keep open whilst you reorganise you last tile to one with the most remaining tiles left (but then you are likely to riich then instead).

iandstanley
Senior Reacher
Senior Reacher
Posts: 138
Joined: Mon May 18, 2009 11:20 am

Re:Rules for start

Post by iandstanley » Sat Jun 13, 2009 11:21 pm

For those interested in WSoM/ZJ I don\'t know if Alan frequents these forums but he does post in the forums on boardgame geek (BGG) and is very willing to answer questions in english on his forms of Mahjong there either by forum or by BGG mail. See also his non-english forum at http://tree.atbbs.jp/zjmj/
pretty sure WSoM either is ZJ or is based on ZJ.
According to Alan\'s own website .... http://home.netvigator.com/~tarot/Mahjo ... index.html
The Zung Jung Mahjong Scoring System is being adopted by the World Series of Mahjong competition, as the \"World Series of Mahjong Scoring System\".
It appears that WSoM maybe Version 4 of ZJ. The page refers to a version 4 although v3.2 is described there.

iandstanley
Senior Reacher
Senior Reacher
Posts: 138
Joined: Mon May 18, 2009 11:20 am

Re:Rules for start

Post by iandstanley » Sat Jun 13, 2009 11:25 pm

Tom Sloper wrote:
It\'s at home.netvigator.com/~tarot/Mahjong/Mahjong.html (Edit: don\'t click it - the forum\'s URL-handling is flawed - copy and paste it to go to the address).
Try typing your urls within [ u r l ] ...... [ / u r l ] (without the spaces) and it works fine

User avatar
Tom Sloper
Moderator
Moderator
Posts: 297
Joined: Sat Jan 17, 2009 8:01 pm
Location: Los Angeles, CA
Contact:

Re:Rules for start

Post by Tom Sloper » Sat Jun 13, 2009 11:59 pm

iandstanley wrote:Tom Sloper wrote:
It\'s at home.netvigator.com/~tarot/Mahjong/Mahjong.html (Edit: don\'t click it - the forum\'s URL-handling is flawed - copy and paste it to go to the address).
Try typing your urls within [ u r l ] ...... [ / u r l ] (without the spaces) and it works fine
I don\'t bother putting myself to all that effort. When some forums handle URLs seamlessly and transparently, I like that best. When you have to do extra typing to make URLs work, I don\'t bother.
4649おねがいします。

HotelFSR
Expert Reacher
Expert Reacher
Posts: 328
Joined: Sat Jan 17, 2009 11:39 am
Location: Federated States of Micronesia

Re:Rules for start

Post by HotelFSR » Sun Jun 14, 2009 10:23 am

I have to say I strongly disagree with Alan Kwan\'s philosophy in general as well as with many of his hand valuations.

It seems to me that one of his primary concerns is creating a \'historically correct\' version of Mahjong, such that he likes to penalize or remove many of the more modern aspects of the game.

He is obsessed with Mahjong being a game of patterns (and by extension a game of all out offense and little defensive skill), where the \'aesthetics\' end up being more important than the gameplay. This is not how he would put it but I think it amounts to a fair description.

Until he starts prioritizing gameplay that maximizes skill without consideration for aesthetics or pretty patterns, I don\'t think the WSoM rules will ever be authoritative or widely popular outside the tournament.

To say that patterns are the most important aspect of the game is like saying that poker is all about patterns. What the patterns (hands) are in Poker or how they rank has little fundamental bearing on what makes Poker a great game.

Benjamin
Senior Reacher
Senior Reacher
Posts: 157
Joined: Sun Jan 18, 2009 1:49 am
Location: Tokyo, Japan

Re:Rules for start

Post by Benjamin » Sun Jun 14, 2009 11:21 am

HotelFSR wrote: He is obsessed with Mahjong being a game of patterns (and by extension a game of all out offense and little defensive skill), where the \'aesthetics\' end up being more important than the gameplay. This is not how he would put it but I think it amounts to a fair description.
I think you\'re dead on here. I don\'t know about Alan\'s personal ambitions in designing the Zung Jung rules, but their major flaw is that they don\'t take defensive play into account.

iandstanley
Senior Reacher
Senior Reacher
Posts: 138
Joined: Mon May 18, 2009 11:20 am

Re:Rules for start

Post by iandstanley » Tue Jun 16, 2009 2:59 pm

Tom Sloper wrote: I don\'t bother putting myself to all that effort. When some forums handle URLs seamlessly and transparently, I like that best. When you have to do extra typing to make URLs work, I don\'t bother.
I wouldn\'t either but you can paste in your url, highlight it and click the URL button in the message composer and the board puts in the [ u r l ] [ / u r l ] for you

Tina
Fresh Reacher
Fresh Reacher
Posts: 14
Joined: Mon May 18, 2009 12:29 pm
Location: Copenhagen, Denmark

Re:Rules for start

Post by Tina » Tue Aug 11, 2009 11:38 am

iandstanley wrote:
MCR follows Riichi with penalising the discarder more but you still get stung:
* on discard (extra points + basic points) x3 (non-discarders still pay extra points)
* on self draw: extra-points x3 + basic points
This should be:
* on discard (extra points)x3 + basic points (non-discarders still pay extra points, discarder pays both basic and extra points)
* on self draw: (extra-points + basic points)x3
where
extra points = 8
basic points min. 8
Despite all this it is becoming very popular in China and is the official standard there. The MCR rules where written by a committee that included China, Japan, USA, Germany, France, Holland, and Hungary and was a complete overhaul of the traditional chinese rules. It has a distinct flavour with hints of the japanese game popping up in the odd rule.
While MCR indeed has an official status in China, claiming that is it very popular is an exaggeration. The truth is that the vast majority of Chinese play with their local/regional rules which differ a lot. I doubt the players of MCR in China amount to more than a few thousand, but it\'s hard to get a clear picture.

The MCR rules were finalized in 1998 by a committe of Chinese and Japanese. The Americans and Europeans only contributed to the translation.

Post Reply