Slitherine

hammy wrote: It only makes a difference really in the mid table where a player might get 2 wins and 2 losses while another gets 1 win and 3 draws. It is likey that the one with one win and 3 draws will have more points on the 25-0 system but not in a win loss system.

It might be the middle where it is most needed though. Consider the LH discussion which is in some cases about people playing for draws and not finishing on the top, but finishing higher than people that "go for it."

A tense draw on the top couple of tables is one thing, a boring non-game when you are in the middle of the pack at the end of the comp is another. Also worth noting that a general feeling that the middle table has a better chance of pushing someone up to the top table might encourage more aggression all around...

ethan wrote:

hammy wrote: It only makes a difference really in the mid table where a player might get 2 wins and 2 losses while another gets 1 win and 3 draws. It is likey that the one with one win and 3 draws will have more points on the 25-0 system but not in a win loss system.

It might be the middle where it is most needed though. Consider the LH discussion which is in some cases about people playing for draws and not finishing on the top, but finishing higher than people that "go for it."

A tense draw on the top couple of tables is one thing, a boring non-game when you are in the middle of the pack at the end of the comp is another. Also worth noting that a general feeling that the middle table has a better chance of pushing someone up to the top table might encourage more aggression all around...

True but the current system will give a player that wins two games and loses two games more points (more likely than not) than one that draws four.

wildone wrote:Why not a simple 5 - 1 - 0 system where you get 5pts for the win and 1pt for a draw. Use the 0-25 points to sort out rankings.
So in a 4 game competition, a player who wins 3 games and draws 1 gets 16 pts. A player who wins 3 games and loses 1 gets 15ps.
If you have 2 players who both win 2 games they get 10 points each and their respective rankings are determined by the 0-25 points system.

I define a draw as either running out of time or both armies mutually breaking.

regards
Brent

The 3-1-0 system used in France for DBM worked just fine. I played on the French circuit for nearly two years and the system very quickly produced the desired results. Most players soon realised than one draw or loss would see them out of the race for the top position in a four game competition. No matter what army you took you had to get stuck in to win and playing for draws virtually disappeared.

What I found interesting at the time was that it was the Brits who then proceeded to emasculate the system by giving 2 points for winning draws.

Julian

The problem with a 3/1/0 or 5/1/0 system is that if you have more than 16 players you can have 2 or more players with 4 wins, which means that you need something more granular to separate them. Whilst I do agree that it would emphasise the need to win to do well, I also like the fact that the current system gives some reward if you don't succeed, proportionate to the success that you have had. Maybe increasing the win bonus in the current system to 10 would make the necessary point, but would retain the granularity.

The actual problem is, it does not solve the problem with players that are only out not to lose. That type does not care as much about his final placement in the tourney. Instead what is important to that type is being able to claim that he wasn't defeated at all. So no matter if a Draw counts 0, 1 or 100 it will solve at best part of the problem.
The only thing you can do with a scoring system is to encourage the players that care about their final placement (already there) and not hand advantages to those trying to avoid defeat (that is where the current system falls slightly short as certain high BG armies find it easier to avoid defeat). How much a draw or victory is worth is IMO unlikely to make much difference.
We will try a slightly different scoring system in a tournament in July (you can find the details in the invitation HERE) that might deal with the second part.

Hammy is right. A stale no loss game gets 10-10. A really bloody fight where you each are one Attrition Point away from losing can end up with a score something like 10.9-9.1. If one player wins by breaking the other it end up 15.9-9.1 or 16.3-8.7. That is quite a big swing. Makes going for it very worthwhile.

Is now the moment to open up the thread about non-zero sum games and the risk of colusion vs. systems that encourage corner sitters?

Oh well, thought not.

kevinj wrote:The problem with a 3/1/0 or 5/1/0 system is that if you have more than 16 players you can have 2 or more players with 4 wins, which means that you need something more granular to separate them.

In theory yes, in practice no. 12 points was hard to get. But they did use a tie-break system in France, which IIRC was percentage inflicted less percentage lost (Julian may remember better than myself). Which is essentially the 20-0 score.

kevinj wrote: Whilst I do agree that it would emphasise the need to win to do well, I also like the fact that the current system gives some reward if you don't succeed, proportionate to the success that you have had. Maybe increasing the win bonus in the current system to 10 would make the necessary point, but would retain the granularity.

Isn't giving some reward for losing just emasculating it again? To make a winner takes all scoring system work, you need to emphasis the winning part. 20-0 +30 would work.

Running it as a one-off doesn't work either, it takes a while for mind-sets to change.

timmy1 wrote: Is now the moment to open up the thread about non-zero sum games and the risk of colusion vs. systems that encourage corner sitters?

Risk of collusion seems to be a particular UK concern, no idea why.

How about a system, where blindly after each match a player votes.
"My opponent was fun and played to win"
"My opponent played not to lose"

Then whomever get the most "Played not to lose" secret votes gets announced.

"Shoot one to encourage the rest"

Isn't giving some reward for losing just emasculating it again? To make a winner takes all scoring system work, you need to emphasis the winning part.

I'm not totally sold on a winner takes all system. I think there needs to be a balance where it is clear that in order to get anywhere you need to look for wins, but retain an element of reward in other games so that people who are not winning can see some return. I'm particularly thinking of this in the context of newer players, who may be put off if they play their games and don't see even any slight improvement from one competition to the next. Also, I think that if winning the game becomes the only way of retaining an interest in the competition, that may erode some of the pleasure of playing. I wouldn't want to see a return to the way competitions were 10-15 years ago.

Risk of collusion seems to be a particular UK concern, no idea why.

Experience. Under the old 7th edition scoring system where players could score anything from 0-1250 points for the winner and 0-600 for the loser, there were a surprising number of games that finished at the top end of the range.

Ghaznavid wrote:The actual problem is, it does not solve the problem with players that are only out not to lose. That type does not care as much about his final placement in the tourney. Instead what is important to that type is being able to claim that he wasn't defeated at all. So no matter if a Draw counts 0, 1 or 100 it will solve at best part of the problem.
The only thing you can do with a scoring system is to encourage the players that care about their final placement (already there) and not hand advantages to those trying to avoid defeat (that is where the current system falls slightly short as certain high BG armies find it easier to avoid defeat). How much a draw or victory is worth is IMO unlikely to make much difference.
We will try a slightly different scoring system in a tournament in July (you can find the details in the invitation HERE) that might deal with the second part.

Actually, the "I don't want to lose no matter what" player isn't a problem I've considered because, as you say, there is nothing you can do about them. What does happen with annoying regularity with the current system is that in the last game both players have won three games but one has enough of a lead that a draw will ensure him first place. With a 3-1-0 system you can't do that because both players would be have 9 points and more than likely the people on the next table could also be contenders.

One system that's beem suggested in Spain is to have the seeded players each other in the first round. This usually ensures the top players play at least two other good players instead of racking up the points against newbies before finding that a draw will be enough in the last round.

Julian

Nyaaaargh...! I just typed a long response and the system ate it

To summarise my points:

1) in 3-1-0 a player on 3 wins after 3 games can only be pipped by another player on 3 wins after 3 games or by the tie break

2) in 25-0 if you can win the comp with a 10-10 draw you have to be 16 points ahead of third place which is not easy

3) using a system with no win bonus while it might reduce the complaining about the lost win opportunities that draw minded players present actually encourages playing for a draw.

4) all games have the issue of players playing not to lose against better players. This is something that the better playrs have to find a way to deal with and being able to do this is a key to being a top player.

5) fixing the draw so top players play other top players in round 1 just leads to round 2 being a total lottery where some winning top players get easy games and some losing ones get really tough ones from other losing top players. IMO any form of seeding is bad and anything like accelerated pairings is doubly so

hammy wrote: 4) all games have the issue of players playing not to lose against better players. This is something that the better playrs have to find a way to deal with and being able to do this is a key to being a top player.

I don't think it's a question of players playing not to lose, this is normal. It's a question of the high BG numbers and the percentage-based scoring system combining to encourage armies that are to all intents and purposes unbreakable in a normal competition game.

hammy wrote: 5) fixing the draw so top players play other top players in round 1 just leads to round 2 being a total lottery where some winning top players get easy games and some losing ones get really tough ones from other losing top players. IMO any form of seeding is bad and anything like accelerated pairings is doubly so

Conversely without seeding a "top player" can pick up an easy first round and a second from another weaker player who won. They can then effectively cruise to victory. I've seen it happen many times. So there are arguments for and against.

Personally I prefer challenging games and given the choice I'd rather play the best four players at a (4 game) tournament than win it. But there are many who prefer 4 easy games and the usual garish plastic cup at the end. If anyone can design a draw system that allows me to always play the best, I'm all for it

kevinj wrote: Experience. Under the old 7th edition scoring system where players could score anything from 0-1250 points for the winner and 0-600 for the loser, there were a surprising number of games that finished at the top end of the range.

That sounds like a badly designed scoring-system as there is an incentive to give away points. In 3/1/0 (or 5/1/0) there is no incentive; giving an opponent 2 means you lose 1.

peterrjohnston wrote: Conversely without seeding a "top player" can pick up an easy first round and a second from another weaker player who won. They can then effectively cruise to victory. I've seen it happen many times. So there are arguments for and against.

Not being a top player or likely to be following on what you have said here, why not change the seeding on its head.

At most events its only the top 20 or so that are seeded and thus missing each other in the first round why not have the top 20 seeds on the day play each other and the remainder of the players do as they do now.

People have explained to me many reasons for seedings but if your a top player you'll more than likely play a top player in the second round any way so why not play them in the first round.

Just a little thought

hammy wrote:
4) all games have the issue of players playing not to lose against better players. This is something that the better playrs have to find a way to deal with and being able to do this is a key to being a top player.

5) fixing the draw so top players play other top players in round 1 just leads to round 2 being a total lottery where some winning top players get easy games and some losing ones get really tough ones from other losing top players. IMO any form of seeding is bad and anything like accelerated pairings is doubly so

Here in lies your problem.

You are focused on top players. Thank you. Since I am rated better than you Hammy. Gotta love this ELO.

I would prefer that you focus on making the game enjoyable for the vast middle swath of players. Like those that don't really worry about themselves in the winner circle, but hope to have fun and break past the mid-point.

Too much of the tournament world is focused on only solutions and pleasure for the contenders.

david53 wrote:
Not being a top player or likely to be following on what you have said here, why not change the seeding on its head.

People have explained to me many reasons for seedings but if your a top player you'll more than likely play a top player in the second round any way so why not play them in the first round.

Just a little thought

Yep its called Accelerated Pairings. It was used in the end stages of the DBm tournament circuit. I thought it provided more better games to more people.
The elite english tournament tigers objected with the following critique:
The seeding is imperfect, so a quality non-seed gets easy first rounds while quality players get draws and bloody hard fights early on. Also you can have a good middlin player pop past all the deserving quality players and may win. Finally the belief that weaker players get stronger playing better players.

I prefer Accelerated Pairings in spite of the critique.

Sadly I consider accelerated pairings to be total b*&*("$

The only AP system I have played under went like this:

Split your field into 4 equal chunks based on a ranking system
In the first found the top quarter play people from the second quarter and the third quarter play people from the fourth.
In round two winners from the top half of the draw play other top half winners, losers from the top half play winners from the bottom half and losers from the bottom half play other losers from the bottom half.

This is fine in theory but unless it is tied to an accurate ranking system the results are total hogwash. With an accurate ranking system there is still a significant risk that you end up in round two with players from the bottom half of the draw in the top places and as you have an accurate system this means you end up with bunnies at the top of the draw. The sharks who did well will probably be just below these lucky bunnies and then it is pot luck for the sharks as to if they get drawn a bunny and lots of nice nearly free points or yet another shark.

The end result is that it is actually quite common for players at the top of the field to find their games getting easier as things progress which is not right.

hazelbark wrote: Here in lies your problem.

You are focused on top players. Thank you. Since I am rated better than you Hammy. Gotta love this ELO.

I would prefer that you focus on making the game enjoyable for the vast middle swath of players. Like those that don't really worry about themselves in the winner circle, but hope to have fun and break past the mid-point.

Too much of the tournament world is focused on only solutions and pleasure for the contenders.

Actually much as it may seem odd I am focussed on getting more people playing and on players all through the field having a good time.

Sadly there are a number of mid field players who's primary driver at tournaments seems to be not losing. They don't care if they placed in the bottom quarter of the field, what they care about is that they didn't lose a game. There is no change to any scoring system or draw process that will change the way these players play. Heck when I was a new player my first focus was to stop losing lots of games so I fully understand the sentiment.

As for ELO rankings and international validity thereof I would be more than willng to play anyone from the States for a hefty wager related to the difference in our ELO rating And I don't play many singles comps so I must be an easy victim

hammy wrote:The end result is that it is actually quite common for players at the top of the field to find their games getting easier as things progress which is not right.

Why does that matter? Sure it is different from what we might expect without accelerated pairings but it doesn't have to be "wrong." Imagine I play four games two easy and two hard. Does it really matter if it is Hard, hard, easy, easy vs. Easy, Easy, Hard, Hard?

If we think a tournament winner should have two easy and two hard games (or whatever). What we want to eliminate is the chance that they might get an Easy, Easy, Easy, Hard draw while everyone else at the top is Easy, Easy, Hard, Hard.

What system best does that is an open question, but as long as everyone at the top has the same number of quality games ("hard" games" it doesn't seem to me that it matters what order they come in.