Thread Rating:
  • 0 Vote(s) - 0 Average
  • 1
  • 2
  • 3
  • 4
  • 5
A notion for WoCG tournaments
This is some work, to be sure, but it might be a good idea.  

The basic tournament structure is total scoring, with rounds of 4 hands.  This allows all rounds to be played in the same time frame, and allows players/partnerships to play against a wide variety of opponents.  The total points scored over all the rounds is used to determine the winner.  

For PAIRS, partnership play:

After each round, each pair is ranked based on total score so far.  The highest scoring pair plays against the 2nd highest scoring pair *that they have not yet played.*  And this pattern gets repeated for all pairs.  This is what's commonly called Swiss format.  Note that win/loss doesn't matter;  if pair A gets a double run on one hand, and pair B gets double aces on another, they're both in line for high total scores for the round.  So be it.  The overall winner isn't based on win/loss, so there's no reason to base the matchups on win/loss.

So...say there's 32 pairs...there will be 16 tables in play.  In a typical Swiss format with win/loss scoring, 5 rounds would amount to a straight single elimination.  Swiss usually adds an additional round to avoid's kinda like a double elimination.  If we postulate 20 minutes per round and a 5 minute break between rounds...5 rounds would be over 2 hours.  Might be unwise to go much longer than that.

These are the basics, largely following NPA, but also very similar to ACBL teams play.  NOW, the somewhat messy parts.  This is based on ACBL pairs tournaments.

a)  A massive flaw in total scoring is the sickening luck factor of the random deal.  Eliminate this:  in each round, every table plays the SAME 4 hands.
b)  If you implement a), then you can also implement a more complex, but MUCH more comprehensive, scoring system, as is done in duplicate bridge...matchpointing.  In matchpointing, your result is compared against all the other pairs that had your hands.  You get points for each pair who fared less well.  

Note that this makes good play *essential*...and bidding to get the hand directed properly is critical.

A quick example of how it works.  4 tables.  On the first hand:
Table 1:  E/W  90  N/S  22    E/W +68
Table 2:  E/W  87  N/S 0       E/W +87
Table 3:  E/W  86  N/S 0       E/W +86
Table 4:  E/W  91  N/S 21     E/W +70

E/W at table 2 had the best result;  they get 3 points.  Table 3 E/W gets 2, Table 4 gets 1, and Table 1 E/W gets 0.  The N/S pairs are the converse;  here, it's the *least negative* that did the best.  

(BTW:  the difference?  Say E/W's best trump suit is a NON run.  They pull considerably more...enough to deny N/S 20.

A final potential refinement:  Tables 2 and 3 might be considered a tie;  each N/S and E/W would get 1/2 for each tie.  Let's face it:  the meld scores will generally be the same.  Is 1 or 2 points in the play...UNLESS we're talking save/no save or make/set...significant?  That is something for discussion WAY down the line.

At the end of each round, pairs move.  With not too many pairs, the movement is arranged such that each pair plays against each other pair at the table...and tries to balance out how often each pair compares against the other pair A and pair B sit E/W or N/S about the same number of times as pair 9 and pair 11 do so.  This isn't entirely simple...but it's also been worked out LONG since.

Oh, and as a side note:  you don't necessarily play 4 hands per round.  You can play 2 hands per round just fine.  Probably even 3 hands per round, for a small tournament.

So...with 8 pairs, you would have 7 rounds.  IIRC, there might be *one* pair that stays in one pair of seats throughout;  this is the 'anchor' around which the movement works.  I'd go with 3 hands per round...a total of 21 hands.    If you have a much larger field, like 18 pairs...9 won't have time to do a full 'round robin.'  The most common notion is that the E/W pairs always sit E/W (and move from table to table) while the N/S pairs always sit N/S (and don't move.)  So you have 2 winners, one for each group.  This works best when the event actually comprises 2 full *sessions* like this...because you redistribute the N/S and E/W pairs for the second session.

I know of NO online site that runs tournaments like this...particularly if you implement "duplicate pinochle."  And you'd have a good, solid ladder system based on tournament results.  That would mean a LOT more to me than the BS rankings systems.

Nuff said, it's late and I'm going to bed.  This is pretty dense, I'll grant, and for me it's largely second nature...I played tournament bridge for many years, and directed at our local club semi-regularly.  This *sounds* complex...but it's really not that bad.  Doing everything on a computer would make things actually a WHOLE lot easier.
I think this is a fantastic idea, there is certainly nothing like it for Pinochle on the web.
This could be implement by any existing Pinochle host or by a newcomer to online Pinochle hosting.

There is a correlation to online Poker here.

Back when I played on PokerStars and FullTilt, you would see a whole range of different variants of Poker and different event types.
Specifically you would see:
  • Cash-games -- players could sit down for as little as 1 hand and then leave with no penalty; maximum freedom with minimum table stability.
  • Sit-n-Go's -- players would register for a game that would start as soon as all of the seats were filled.  This is effectively what happens at all Pinochle sites.
  • Tournaments -- players register (with a ridiculously high ceiling of maximum players) for a game that officially starts at a preordained time.  I don't know of any existing Pinochle software that allows a massive pool of players to all sit for a single tournament -- at least not in the style of online Poker or in Swiss format.
I'd love to see something like Duplicate Bridge for Pinochle!
This is certainly the angle I was promoting with our recent Bot Competition discussions.

When you get Apples:Apples comparisons, you get really clear results with a large portion of luck stripped away.

This is format is just begging to get picked up by a forward thinking developer/host.
I'll keep my fingers crossed for a while.
It's unbelievable how much you don't know about the game you've been playing all your life. -- Mickey Mantle
This is an interesting idea.

I have it on my to-do list to add a tournament system to the site. But it's not a high priority item. Why? It appears to me that any such system will be rarely if ever used.

About 1 year ago, I started a Spades "league" called Jungle Gym Spades. Since I didn't have time to work out all the details of running tournaments "in-house", I decided to use a 3rd-party system. There's a website called MyLeague that manages leagues. You can run the actual games on just about any software that you want, and MyLeague is used for reporting results and managing the player ratings. So people will play World of Warcraft, Scrabble, etc within their leagues and use the MyLeague site to schedule tournaments and report results.

As the main (currently only) tournament admin at World of Card Games, I find the system to be cumbersome. It's not much of a chore to actually schedule a tournament, but sending out reminders is a pain, and sometimes people forget that they registered. In those cases I have to find someone to substitute, or I have to substitute myself. I like to play a game, when I have time, but I'm often trying to get some work done while a tournament is running, and it can be inconvenient.

Further, the MyLeague system has very little automation. So I actually have to pull the trigger to start the tournament, and I have to schedule the rounds between players. It's just very hands-on, and in this day and age I find it ridiculous that I can't just schedule a tournament and let the rest happen automatically.

I developed a crude user interface for seating players at tournament tables, but it's not something that I'd want to see in wide use. If I wanted to make it available for the average user, I'd need to put some work into it.

However, based on experience so far, there is not much demand for this kind of system. I've been told that people love to play in tournaments. But our Spades league has never had more than about 35 players at one time. And of the 35 players, it's usually only 10 or so that will register for a tourney. There are about 8-10 players who seem to really enjoy the tournaments and can be counted on to register whenever I schedule one. But this number of players is not enough to make it worthwhile to build out an entire tournament system.

We do have the problem that I can't schedule enough tournaments, I suppose. Originally I'd try to schedule 2-3 per week. Ideally they'd be running all day long. To do this you'd need a dedicated team of tournament directors to arrange things. I did have one volunteer who helped out for quite a while, but eventually she quit. And as the site developer, I don't have time to run tournaments all day long (although I'd do it if I thought this would bring many more players to the site).

If I could run tournaments in an automated way, so that they could be run with frequency, it might improve participation.

Currently I've seen a max of about 20 pinochle players at the site at one time. In contrast, Hearts usually has about 60-70 during peak hours, with Spades getting about 45 or so during the peak.

One question is: would enough people have the time to set aside the 1-2 hours required to play in a tournament? I just don't know; I get the impression that the answer is "no" at World of Card Games. It seems to be just a small fraction of players who have this time, at odd hours. At times I've scheduled a tournament and had to cancel because we didn't get enough players.

At WoCG, I've already started adding a rating system (Elo ratings) to games, since some people seem to like that part of the tournament system. Ranked tables are strict: you can't start a game with bots, and you must be registered to play. If you quit and don't return in 2 minutes, you'll get a loss. These games take substantially longer to start. I'm guessing people don't like having to start a game without bots, since there's a risk that someone will quit, and people don't like waiting for a replacement (understandably).

My conclusion, so far, is that tournaments sound nice, but without enough players, they just don't work. And I don't have the time to build something into the site that is used by only a small fraction of players. Until I have strong evidence that enough people would use an in-house tourney system, and that it would help to bring players to the site, I'm inclined to avoid that project.
Play Pinochle at World of Card Games!
Fair enough.
I agree, the project needs critical mass to be successful.
Best to put this one on the shelf for the moment.
It's unbelievable how much you don't know about the game you've been playing all your life. -- Mickey Mantle
Definitely need more players on a regular basis, but then again, poker websites have been setting up what they called sit and gos, which are tournaments that have a set number of players, and a set structure and they only start when enough players have registered. Sometimes you end up sitting at a table for 20 minutes before it fills up, so i wonder if small pinochle tourneys could be set up like that? What about an 8 player tournament? When 8 people finally register it would automatically start. Then you can have rules/structure similar to what TE recommended, but just on a small scale? That would give you 4 teams. Just a thought.
Hmm. It could be done. shows the 2 table movement. 24 hands should be'd be *long* by the standards of online play, but that's always going to be the case.

One distinction, tho: in a poker SNG, if you get disconnected, or if you choose to give up (not common in SNGs, but I've seen it happen in low-stakes or freeroll tournaments) big deal. You have a set time window within which to act. If you don't do so, your hand is dead, and you can't win that hand. For pace of play, if you do time out, you're marked as "sitting out" until you click a button saying "yes, I'm back." The event can continue just fine without you.

The essential difference is, of course, the fluid nature of power play. A sit-out is no different from a fold. Things don't work as well in bridge or pino. At a live tournament, it *can* happen that a player becomes ill, or some external emergency arises such that they have to leave. The usual procedure, as I recall, is to recruit a substitute from among the kibitzers, or other players not participating, or sometimes the event caddies...generally kids roped in to help with purely mechanical aspects. I've never played a bridge tournament online; I don't know what they do. Their rules should supply some guidance.

This might work as a testbed for quite a few things. Note that it can be run as a pairs tournament or an now have 7 rounds of 4 boards. You can refine timekeeping rules, and scoring systems. (HOWEVER, scoring for 2 tables probably has a *significantly* different basis.) You can start getting people used to playing for several hours...or find they just don't care for it.

Unfortunately, tho: I'd estimate that setting up the 2-table SNG would involve probably 90% of the effort needed to set up for a 24-pair, 12-table event. As rak points out, tho, it does accomodate a smaller player base. WoCG might not be big enough...but some of that might be their policies. (Just over there...I have 2 choices, play against the computer, or against people but with bots allowed. FORGET IT. Bots are an abomination. Plus, no lobby.) PlayOK probably would have enough of a player base to pull this off, tho.
(05-18-2016, 02:03 PM)ToreadorElder Wrote:  ... over there [WoCG]...I have 2 choices, play against the computer, or against people but with bots allowed.  FORGET IT.  Bots are an abomination.  Plus, no lobby.)  

For the record, at WoCG, you can configure a private table to exclude the use of bots.
And the "lobby" is a feature accessible by clicking "list of tables" text in the lower left corner of the page (it shows all active public card games).

BAD bots are an abomination -- a reality that a handful of PP members are actively trying to correct.

I'll agree that the WoCG lobby is not what I was initially looking for.  I was so used to the Yahoo Pinochle, PokerStars/FullTilt Poker, and PlayOK Pinochle lobby interfaces.
After I asked for directions, I was able to see what I wanted.
It's unbelievable how much you don't know about the game you've been playing all your life. -- Mickey Mantle
I jumped the gun on that comment; my apologies. I was playing without a login, and it was a cursory check. Unfair to criticize on that basis.

I will disagree about bots tho. ALL bots are an abomination, IMO, on principle...and it doesn't help that I've never seen any bot I'd rate above weak. To be sure, lots of players are as well...played a game over on PlayOK earlier, and my partner's play had some big holes.
I'll agree that all known existing pinochle bots are bad bots and bad bots are abominations (hindrances to the game).

This reality does not need to be permanent.

Marya will be the first to admit that her bots are imperfect.
She is also committed to improving them as new, intelligent guidance is provided to her (mostly in these forums).
I wish to offer sincere thanks to Marya for her spirit and dedication to grow her software and provide modern options for online play.

Across the world of games and bots, not all bots are abominations (think Chess, if not others).
There is a growing list of games where bot development has become so strong that it can regularly defeat human competition.
This is a grand compliment to what the human mind can create -- computational intelligence superior to our own!
It's unbelievable how much you don't know about the game you've been playing all your life. -- Mickey Mantle
Yes, ALL bots are abominations in this scenario, mixing human players and bots. Sure, there are chess bots that play pretty well...certainly better than me! Don't even count Deep Thought, or the best computer go algorithms; those are elephants compared to a chihuahua that Marya could get running...not a matter of skill, but time and computer resources.

When it's 1 human versus 1 bot, then a decent bot is not unreasonable. In ANY other situation, it's awful in my book. Even 3 bots with one player is pretty bad, IMO, even with decent bots. It's not just the quality of the bot, it's the nature of the game. I can understand, tho, as a teaching exercise, 3 bots and 1 player. It might not be great long-term, but as a starting approach, bots have the advantage that they never yell at their partner. Smile

BTW: go and chess are much easier, in some ways, to code, compared to card games. AFAIK, there are bridge bots that can bid...pretty well. At least with a lot of hands. They fall down BADLY on play, and especially on defense...which in part mirrors human experience. Defense is much harder than declarer play, at any level. Pino has more look-ahead than bridge, and less available info. I think a decent pino bot is harder to write, in terms of the play...probably easier to write for bidding, simply because the bidding is so much simpler...but the factors that are involved in play decisions are, I think, more numerous, AND there's less information. It's nothing like poker, but I know of no game that is. It's actually non-trivial to identify all the factors that come into play in choosing an action; but the deciding factors between plan A and plan B in pino can be quite subtle.

Forum Jump:

Users browsing this thread: 2 Guest(s)