Slitherine

Hi, I won both battles on the last round (3rd) but on the one we just finished on the results (after receiving the winning message) it reads lost cause of time out. Is there a way to fix it?

Tks!

Could we also get a clarification on how scoring works and how time out impacts upon this. I've used the formula provided on the previous page but can't get it to work for every score. Also all but 1 of my matches say timed out though I won them all with ample time to spare. Why don't you guys just change the scoring system to a win/lose style heedless of scores as the current way sucks. Players leap up the ladder due to resignations or players losing interest, the match-ups are all important as skilled players deny points to each other while someone just below plays a newbie and gets a monty haul etc. So many faults with the current way and if it is just win/loss based then it won't matter if you got to the top through byes as it provides no advantage in score.

Cheers,

Steve

We'll look in to it- thanks!

I asked about leaving the competition on the 26th.

I have just begun another game and that shouldn't have happened. Noticed 4 other people have left the comp.

Could something be done otherwise I am going to resign again and wait for the bugs to be zapped.

According to the Swiss system, I should not meet the same opponent twice in the same tournament. But I'm playing with Iandavidsmith on round 3 and 4.

Well as it is only fair to Peter I will carry on.

How are you and the family Iain?

Another problem with your scoring system is that I played for a draw in Hidde's match but this is meaningless under the present scoring system. The game finished with no victor but the scoreboard shows a 25/5 split in points. In the in game results it says victory for me but draw in the victory screen but this is an old flaw. I look at the completed games menu and see that it says victory to me 63% losses to about 20% for him bu the game was timed out. On the way to the end I was observing the results table as we played the turns and noted the score stopped changing at a 18/12 split in my favor even though I took further casualties. I had ample time left to play pass the completion time so I assumed that my opponent must have run out of time before we had actually finished. This is a really annoying system and denies a whole element of the game where you can play for a draw with an inferior force which in itself is a victory. I noted that two evenly matched players lost as the Indo and as the Romans in the previous scenario. If you can't play for a draw then don't have imbalanced scenarios in the tournie as it just means those players who get the lucky match-ups score big or leap ahead with negligent opponents. If you guys want to keep the present system please explain why you think it is so good so that I can pinpoint the flaws in your reasoning.

appologies to opponents have been away unexpectedly, could not access net.

pantherboy, the game didn't time out. It ended after my last turn were completed, turn 15.
I agree with all pantherboy write about the scoring. Although the biggest damper on my enjoyment of tournaments like this is absent players and unfinished games the scoring does little to help. The impact of byes are to big and the reason for having the Swiss chess scoring (pairing players based on skill) goes out the window.
I will not join the next tourny (not only because my dislike of the scoring) and I doubt I will play any in the future if the scoring remain as is.

some suggestions
- use chess clock based timing where both players are given a time allocation say 72 hours. if one persons time runs out they automatically lose with maximum points to their opponent. the incentive is to finish games and all games will result in a conclusion within a guaranteed period i.e max game length is 2 times each players time allocation.
- explore a simple win 2, draw 1, loss 0 scoring with count back to split players if they have run into each other during the tourney, where players are still tied then battle group destroyed to lost differential, still equal then highest number of destroyed bg's is the order.
- i think 6 rounds needed for the best player to prevail using swiss chess match-up, maybe 5 would be ok but 4 i think is too short
- a ratings system like in chess to determine initial match ups based on past tourney outcomes i.e. novice beats a grand master scores high rating increase... this will become an ongoing side affair with regular players wanting to increase their ratings


regards CheAhn - the wrath of god

hidde wrote:pantherboy, the game didn't time out. It ended after my last turn were completed, turn 15.
I agree with all pantherboy write about the scoring. Although the biggest damper on my enjoyment of tournaments like this is absent players and unfinished games the scoring does little to help. The impact of byes are to big and the reason for having the Swiss chess scoring (pairing players based on skill) goes out the window.
I will not join the next tourny (not only because my dislike of the scoring) and I doubt I will play any in the future if the scoring remain as is.

Hidde the reason I thought you may of timed out was because I had more time left than what was left on the clock. If we both have 2.5 days and the game is 5 days long than it should of been impossible for me to have more time left after the clock expires unless your clock had run out. But I'm with you and will withdraw from the next tournie even though I like scenarios and refuse to play unless a major overhaul is done to the scoring system. I feel rather dissapointed that my draw was scored 25/5. Too many problems with the system and we've expressed these concerns too many times for me to persist.

What scoring system would people prefer. It must be one that differentiates between small and large wins so win/lose/draw is not what we are looking for. We need a formula that works for all situations of completed games, time out games etc. Rouight ideas are nice but not something you can implement and often have logical impossibilities so real equations are needed to make it work better.

On the byes - what would you prefer. It is not fair to reduce a players score because their opponent did not play. If you score them average points how do you place them for the draw. Any solution needs to cover all the bases.

On the drop outs how can we make it better?

We do not need a system that means the best player always wins. We need a system where the good players are at the top and the weaker ones at the bottom.

If someone wants to manage the disucssion and bring us some ideas we'd be happy to listen but we havent really got time to manage the disucssions ourselves right now - just too much going on.

iainmcneil wrote:It must be one that differentiates between small and large wins so win/lose/draw is not what we are looking for.

Lets just take this one point before addressing any others. I ask why do you need a difference for large or small wins? You state your desire without any explanation why. By applying such varied scores it doesn't improve the results in any way nor have any true meaning. It adds a level of complexity to the calculations and creates issues in a variety of situations. If you are dogmatic about this point then that will make it difficult to improve the system.

iainmcneil wrote:We need a formula that works for all situations of completed games, time out games etc. Rouight ideas are nice but not something you can implement and often have logical impossibilities so real equations are needed to make it work better.

Thus the request for a simplistic scoring system which can easily address said issues. If you recieve a bye you'll recieve points which are no more or less than every other player who wins that round. If you time out the opponent wins with no greater or worse advantage than other winners. If the scenario is terribly imbalanced then you can play for a draw in one and a win in the other and grab the points for the overall win (this addresses the luck factor of who you draw to play and avoids eschewing the scores). And so on for other issues.

iainmcneil wrote:On the byes - what would you prefer. It is not fair to reduce a players score because their opponent did not play. If you score them average points how do you place them for the draw. Any solution needs to cover all the bases.

On the drop outs how can we make it better? .

Answered by a simpler system as stated above.

iainmcneil wrote:We do not need a system that means the best player always wins. We need a system where the good players are at the top and the weaker ones at the bottom

Well your system certainly doesn't achieve this as weaker players get to the top through byes and give easy wins to better players while others have to contend with tougher match-ups. Any system where people can bye or win through default will have players of lower skill getting to higher positions but the issue here is by using a formula as you do now then that creates imbalanced scores. By a simpler scoring method all the players high on the ladder will have the same score and by playing through to the 4th or later round you'll end up with a clear winner of the player who wasn't defeated otherwise tied players with a defeat each with BG casualties being used to differentiate though I'd just avoid this and go for 5 rounds or have tied positions which I think is fine.

My 2 cents worth...

How about a scoring system as follows
4 points for a win
2 for a draw
1 bonus point for a narrow loss or a big win

Then use the current numbers as tie breakers for those on equal points.

A win by dropout would be a standard 4 point win.

cheers,

Take a look at this for example. From Lysimachos "Ist FoG Challenger’s Tournament".

8 pts. 2 victories
7 pts. 1 victory 1 draw
5 pts. 1 victory 1 defeat (with better BG difference)
4 pts. 2 draws / 1 victory 1 defeat (with equal BG difference)
3 pts. 1 defeat 1 victory (with worst BG difference)
1 pts. 1 draw 1 defeat
0 pts. 2 defeats
The BG difference will be comprehensively calculated counting in positive for the winner the number of BG lacking to reach the loss threshold and counting in negative for the looser the number of BG overpassing it.
Example:
Crusaders (35/39) = +4
Ayyubids (47/41) = -6
Total BG difference = Crusaders +10, Ayyubids -10

Maybe changing 7pts to 6 and 1pts to 2 might be better. Anyhow it makes draws mean something and it differentiate between players result to quite a degree.
Especially if you count the BP. In the tournament after two matches I and Morbio both have 16pts. Morbio is +62 and I'm +104.
Plenty of room to separate players with equal points.
Byes is still a problem, of course. Perhaps give 6pts to the non-offender and make the absent player feel some guilt and peer pressure.

Lets say we have 100 players - a win/lose/draw does not differentiate enough. Its as simple as that! I struggle to see why you would not want to reward someone more for winning big than a close fought game. I do not want a system where multiple players are on the same points at the top with a goal difference separating them. Its very unrewarding and the only way around it is many many rounds to differentiate players which is not what we want. That is why this is a requirement of the system and I completely disagree that this stops teh system working. I also completely disagree that a win/lose/draw scoring system solves any issues and for me it just adds new ones unfortunately.

Weaker players do not get top the top through byes - they only get to jump up the table a number of places. The score for a bye could be reduced but what if you are winning the comp, your opponent cant play and someone sneaks past you because you did not get the opportunity to get a bigger win? No solution is going to deal with all situations so you have to find the best compromise. To me it looks like there are only 2 players in the top 10 who have had a bye so neither are in teh top 3 so its not changing the result of the compeition so I don't really see the big deal? How do you feel it is causing large issues?

I do not want a system where multiple players are on the same points at the top with a goal difference separating them.

Ok, fair enough. I wouldn't mind, though. Take the example I gave; 62 vs 104.
62BP: one game of four=big win
104BP: three games of four=big win.
The big wins are rewarded in this way as I see it.

To me it looks like there are only 2 players in the top 10 who have had a bye so neither are in teh top 3 so its not changing the result of the compeition

I count six byes(five players), including me. The margin down to second place is nine pts...
But I'll agree there is no perfect solution to the problem with byes. At least I don't have one.

iainmcneil wrote:Lets say we have 100 players - a win/lose/draw does not differentiate enough. Its as simple as that! I struggle to see why you would not want to reward someone more for winning big than a close fought game. I do not want a system where multiple players are on the same points at the top with a goal difference separating them. Its very unrewarding and the only way around it is many many rounds to differentiate players which is not what we want. That is why this is a requirement of the system and I completely disagree that this stops teh system working. I also completely disagree that a win/lose/draw scoring system solves any issues and for me it just adds new ones unfortunately.

Weaker players do not get top the top through byes - they only get to jump up the table a number of places. The score for a bye could be reduced but what if you are winning the comp, your opponent cant play and someone sneaks past you because you did not get the opportunity to get a bigger win? No solution is going to deal with all situations so you have to find the best compromise. To me it looks like there are only 2 players in the top 10 who have had a bye so neither are in teh top 3 so its not changing the result of the compeition so I don't really see the big deal? How do you feel it is causing large issues?

I absolutely disagree with you. How would you consider a game where a beginner plays a veteran in an imbalanced scenario and the veteran wins big in both games worth more than two veterans playing each other in the same scenarios with a win and a draw. To draw is an achievement in itself when fighting against the odds but slaughtering a beginner certainly isn't. You believe that a big win deserves a greater reward rather than the achievement itself and the personal gratification that you will recieve. Rather than post meaningless totals in your current table which seem to never truly reflect the state of the match just post the actual battle numbers. I think all the top players are the ones that don't care so much about the points being weighted one way or the other.

How about forwarding me the results of the mirror matches from this latest tournament (not the useless numbers the tables show) so that I can examine them and then mock up what an alternative would look like. And only the raw data in the case of you awarding full casualties to players who timed out so that I can take into regard how a program would need to handle it. Maybe if I illustrate it to you then you'll believe.

Yes. You guys obviously can't read stats. As hidde points out the top of the table is full of byes except for 2nd and 3rd who played every game. In my case versus hidde I won and drew but the draw is a 5/25 split. I won and lost to davouthojo in the following game for a 12/18 split in the loss. This doesn't make sense. I scored higher for a loss because I fought to a bloody end while I mustered a draw versus hidde with an outclassed army in terrible terrain and was punished for not just killing as much as I can in futile attacks. Also when examining the results you've got to look at who matched up with who and was it because they had byes. Also examine the top loading of other rounds. I'm stunned that you guys can't see these things but then I said to you prior to the patch things relating to the secuirty program that are true but you said I was wrong because your programmers assured you it was so. I didn't even need to see the program to figure those things out and also I'm not a programmer so it was simple logic that I relied upon. I've also predicted other things to only be told wait and see and then you guys have done an about turn. Please consider that you have an excellent pool of skills amongst the players and that we maybe able to shed some light on issues that would require you to adjust your positions.

Once again I agree with Pantherboy. Winning a battle, or not losing it, is generally all the generals of history cared about in the vast majority of cases (I'm guessing here). Win and you get the rewards; gold, slaves, women, prestige, etc. Lose and you are dead or a slave at best.

So a simple 2 for a win, 1 for a draw would work fine. Or 3 for a 1 and 1 for a draw if you want to reward winners relatively more than non-losers. Alternatively, the Lysimachos table of scoring (which is a variant of the Pantherboy orignal for paired matches LOEG(season 1)) is great too. Especially since I helped contribute to the creation of the scoring system i.e.
8 pts. 2 victories
7 pts. 1 victory 1 draw
5 pts. 1 victory 1 defeat (with better BG difference)
4 pts. 2 draws / 1 victory 1 defeat (with equal BG difference)
3 pts. 1 defeat 1 victory (with worst BG difference)
1 pts. 1 draw 1 defeat
0 pts. 2 defeats

Iain's concern is not being able to differentiate and having a mass of players on the same points, well lets address that issue.

Is a 100 players a realistic number? Let's assume so...

I read earlier that any swiss competition needs a minimum of 5 rounds to be effective. So let's assume a minimum of 5.

Assuming no draws. Gives 50 people on 2 points and 50 on 0. After 2 rounds, 25 on maximum points. 3 rounds, 13. 4 rounds, 7, 5 rounds, 4.

With draws then the number of players with maximum points could only reduce.

So, the maximum number of players that could have maximum points is 4... and that assumes that all win all their games. If only 64 players are in the tournament then the maxium number of players on maximum points is 2.

Of course, all of this assumes that there's always someone, or some people, that win all games. But, based on the competitions I've been involved in so far tends to be true.

If there are a greater spread of wins and losses then I accept that there may be a greater number of people on the same points. This is where the table of points above will help because big wins and close losses matter in an imbalanced scenario (i.e. 5 points rather than a neutral 4) and if you are good enough, or lucky enough to play a weak player, then 2 wins gives 8 points. 8 points! Which is 60% more than the paltry 5 points.

Addressing the issue of not over or under rewarding people receiving byes is going to be a problem, especially with any system that rewards based on size of win. If you get max points for a bye or incomplete match it will overpromote non-players. If they get low or average points then a good player is penalised through no fault of his own. The basic problem is 'what's a win worth?' Well, it's just a win, so why score it differently?

I recommend one of my main rules I use at work when talking to the developers. Keep it simple!

I also recommend Pantherboy's offer of using the raw data to model using different scoring systems. This will help determine if issues are real.

Finally, I expect that the table above would be very easy to implement. Code it and trial it for a couple of tournaments,,,, RoR please, it's all I have