Slitherine

Although the March tournament is still on going we've opened up applications for the April tournament. This one is going to be Swords & Scimitars themed.

This is an open tournament with no player limit.

The tournament starts April 1st and each round lasts 7 days. The later batles are really quite large so we've added 2 more days to complete them and this is the biggest tournemant yet!

http://www.slitherine.com/tournaments/fog_pc

You need Swords & Scimitars to take part in this tournament.

Enjoy!

Hi Iain,

As there will be at least one bye in each round in this April tournament have you been giving any more thought to moderating scores to take out the skewing effect of byes and time-outs with no bp loss?

The link below applies the moderation I suggested in the LT Tournament thread.
After two rounds it only alters the positions of the players with the byes - one hardly at all, one considerably but does reflect their game performance thus far.

I do think it is a fair solution and as it makes absolutely no change to the fundamental scoring system and is very easy to script is well worth considering for future tournaments.

https://docs.google.com/viewer?a=v&pid= ... y=CKOD0LQN

Btw - As this is the second time this method puts ianiow top I wish to clearly state I do not owe him any money nor am I his agent nor am I ianiow in disguise!!

Though loosing in this way the 1st place,

I agree with Triarius that this kind of calculation is much more correct

hoping to regain swiftly the top rank in the last enconters!

I think the general opnion is that a BYE should be worth less points. I'm thinking we should offer around 25 points instead of 30 for now and see how that feels. If that does not help we can reduce it further.

In my humble opinion this looks to be a very little step forward but the real change, with minimal programming work, would really be that of adopting the system envisaged by Triarius. In fact, giving each one the average score of his total matches, without counting the BYE, should reflect each player's real strength, avoiding those spectacular jump in the standings given by the actual system to players benefitting of a BYE. Moreover, structured as it is now, a BYE counts much more in the final stage of a tournament then at its begininng, given the fact that a medium player may gain 60 points in the last turn jumping on the top while someone who got it at the beginning, if always winning his game, has to meet all the best players having little choice of greatly enhancing his scores (also if you win the points won't be 30+30 but generally no more then 17+18 or something like this, wich is 25 points less than a BYE).

The problem with that is if the the drraw will be incorrect for players with BYE's as it has no points to award them meaning they will be getting a much easier draw than they should be getting for those points. It could lead to people with BYE's winning more tournaments and not less. It's not a programming issue - its an issue that we dont like the design as there is no gaurantee it will resolve the issues.

iainmcneil wrote:The problem with that is if the the drraw will be incorrect for players with BYE's as it has no points to award them meaning they will be getting a much easier draw than they should be getting for those points. It could lead to people with BYE's winning more tournaments and not less. It's not a programming issue - its an issue that we dont like the design as there is no gaurantee it will resolve the issues.

Sorry Iain but I do not understand that.
In the alternative system what you suggest can only happen once in the first round to one player.

This is because the player with the bye gets zero points and by default in round 2 plays whoever got the worst drubbing in round 1.
If a second player gets a scratch win they would also get zero points and your scenario does not apply as the two zeros face off in round 2.
If a third player gets a scratch win then one of those three plays the worst score the other two play each other and so on.
And all this can only apply in round 1 as thereafter each player has an average score per game.

I would also argue that the odd person getting a bye in round one then only plays one potentially (and only potentially) weaker player once and thereafter has a position based on their performance. This in effect extends the luck of the round 1 draw into round 2 for just one player or pairs of players with bye/scratched opponents.

This is very different to the current situation where every bye/scratch gets 60 points in any round. Therefore there is a potential skewing effect in every round and furthermore the ripple effect is enormous as players are consistently over-promoted and then get stuffed in their next game consequently over-promoting their opponent and playing havoc with the whole premis of swiss chess matching.

So in the alternative system you may get one mismatched pairing for the odd player in round two only but in the current system there is mismatch in every round with knock on effect creating more secondary mismatches.

Triarius wrote:

iainmcneil wrote:The problem with that is if the the drraw will be incorrect for players with BYE's as it has no points to award them meaning they will be getting a much easier draw than they should be getting for those points. It could lead to people with BYE's winning more tournaments and not less. It's not a programming issue - its an issue that we dont like the design as there is no gaurantee it will resolve the issues.

Sorry Iain but I do not understand that.
In the alternative system what you suggest can only happen once in the first round to one player.

This is because the player with the bye gets zero points and by default in round 2 plays whoever got the worst drubbing in round 1.
If a second player gets a scratch win they would also get zero points and your scenario does not apply as the two zeros face off in round 2.
If a third player gets a scratch win then one of those three plays the worst score the other two play each other and so on.
And all this can only apply in round 1 as thereafter each player has an average score per game.

I would also argue that the odd person getting a bye in round one then only plays one potentially (and only potentially) weaker player once and thereafter has a position based on their performance. This in effect extends the luck of the round 1 draw into round 2 for just one player or pairs of players with bye/scratched opponents.

This is very different to the current situation where every bye/scratch gets 60 points in any round. Therefore there is a potential skewing effect in every round and furthermore the ripple effect is enormous as players are consistently over-promoted and then get stuffed in their next game consequently over-promoting their opponent and playing havoc with the whole premis of swiss chess matching.

So in the alternative system you may get one mismatched pairing for the odd player in round two only but in the current system there is mismatch in every round with knock on effect creating more secondary mismatches.

Agree

I agree with Triarius.

We've tried the system where you allocate scores at the end for BYEs and it ends up with very odd results. People from half way can jump to first place on the last result. E.g. a 3 round comp, I get a BYE, then I get an easy game - 60 nil, then I get a mid level game, say 40-20. I have 50 point average giving me a BYE scopre of 50. I go from 60 points after round 2 to 150, so 90 points in my last game. It makes any score board meaningless. This system means a BYE for a goodplayer is actually the best way to win a tournament as they'll avoid any hard games.

The 60 point bye also has issues. Thats why I think reducing the BYE to around 45-50 points would be the best solution.

I want a solution where the BYE score is allocated instantly or you end up with people getting a much easuier draw than they should do and everyone knows where they stand.

Can you you create a holding score from a bye or a no show of say 30 (15 for each game) in the early games to count toward compiling the draw and then when they start playing actual games that holding score is replaced by their average over the remaining games.

iainmcneil wrote:We've tried the system where you allocate scores at the end for BYEs and it ends up with very odd results. People from half way can jump to first place on the last result. E.g. a 3 round comp, I get a BYE, then I get an easy game - 60 nil, then I get a mid level game, say 40-20. I have 50 point average giving me a BYE scopre of 50. I go from 60 points after round 2 to 150, so 90 points in my last game. It makes any score board meaningless. This system means a BYE for a goodplayer is actually the best way to win a tournament as they'll avoid any hard games.

The 60 point bye also has issues. Thats why I think reducing the BYE to around 45-50 points would be the best solution.

I want a solution where the BYE score is allocated instantly or you end up with people getting a much easuier draw than they should do and everyone knows where they stand.

Iain

I do not think I explained my suggestion well enough.

I am not suggesting allocating points for byes at the end I am suggesting using no points for byes. This can be done at any point in time in any round.
Not just for byes but also for scratched opponents.

This is the situation now with some round three games still playing out - this took me 10 minutes to script and upload from scratch. It gives a current moderated ranking.
https://docs.google.com/viewer?a=v&pid= ... y=CMm_zrcC

My first post has the situation at end round two - I could have done the same at end round 1 or at any point in the tournament.

The key point is that score per game is the 'moderated' ranking score - this changes your system not one whit except that it moderates out byes. Byes do not contribute a score and do not contribute a game. and therefore confer no advantage or disadvantage at all to either a good player or a less experienced player.

Hope this is clearer.

For the first time ianiow is not top so it seems fairer to me too

I think Iain's issue with your system is that by allocating zero points for a bye it could enable an easy game for them in the next round. My suggestion to Iain is to use a dummy score of 30 for byes in the first round when allocating games and then to use your method for allocating actual points (which are based on actual games played). It's not unduly complicated and addresses both issues...ie the draw and a fair awarding of points.

dazzam wrote:I think Iain's issue with your system is that by allocating zero points for a bye it could enable an easy game for them in the next round. My suggestion to Iain is to use a dummy score of 30 for byes in the first round when allocating games and then to use your method for allocating actual points (which are based on actual games played). It's not unduly complicated and addresses both issues...ie the draw and a fair awarding of points.

Hi Dazzam,

I follow that but I do not understand the point - as I tried to explain above with the alternative suggested scoring this scenario (goodplayer gets bye gets easy game) could only happen in round two for a single good player if that good player received a bye in round 1 - in rounds 3 onwards a player gets a game according to their average points per game. This is very different to the current system or any system that gives a bye any points and so skews results in every single round in which byes or scratches occur, for every single player who gets an arbitary score and then has a knock on effect in each following round. This is a cascade of unintended consequences.

So what Iain suggests may happen once in round 2 for a single player.
Furthermore round 1 is 'luck of the draw' matches/pairings anyway - it must be.
Therefore a good player getting a bye and a possiibly easier round 2 is no more or less disadvantaged than he (or she) would have been with an 'easy' draw in round 1 - you could argue that that is more fair.

Let me change the debate - go on convince me that a bye should be given any arbitary score at all and count as a game played and then explain how you justify that score and its affects in every round for every player to whom it is applied?

Actually Dazzam and Iain I will add what I think is the weakness of the alternative system but one which remains a much more limited weakness than scoring byes and will possibly apply to one player in round 2.

It is this scenario :-

An experienced player (A) gets a draw against an inexperienced player (X) in round 1 and scores the biggest win.
A second experienced player (B) recieves the round 1 bye.
Because going into round 2 B has no score yet and X has the lowest round 1 score - swiss chess pairs them.
Likely result is that X gets a second thumping in a row.

The experienced players are not disadvantaged (A is playing in the top of the swiss chess draw for round 2 and in round 3 B is up at a level which reflects the scale of his victory over X) and carry on with their tournament. Player B has been no more or less lucky than A in his opening game pairing. However the inexperienced X, potentially a newcomer, has had two thumpings in a row and has had two bad pairings rather than 1 unlucky round 1 pairing and is disadvantaged.

It will of course only happen if a good player gets a round 1 bye. This likelihood could be mitigated by only drawing newcomers for round 1 byes - and it is only a game

The only issue I have with your example Triarius is that the player B is a little advantaged as by having the bye his subsequent matches count for 1/3rd of his aggregate score while those without (incl Player A in your example) each match-up counts 1/4 of their aggregate score. So in the drubbing player B will inflict in the 2nd round actually will have a greater impact on his aggregate score in the end.

I agree totally with your way of allocating points. I'm merely trying to deal with Iain's issue of it impacting on the draw. I think my suggested add-on to your methodology deals with Iain's concern yet retains your fair system for allocation of points.

Just a clarification about the rules.
If a player isn't able to finish a game for a shortage of time how much point is he going to get (0) and how many will his opponent (30)?
Or will it be taken care, in some ways, of the score reached until then?
Thank's

Hi,

I don't understand anything. Probably I'll quit to play this tournaments.
Before the tournament was over, in my last games the score was 27+25 vs 3+5 for me and it was my opponent turns.
The tournament finished and I won the two games (timeout victory) but the score now is 16+27 vs 14+3 for me.
I don't understand why I have to loose 9 pts.

Some brief considerations about my first (enjoyable) experience in a Slitherine tournament:

1 – The 7 days format is, in my opinion, the minimum required if the idea is to attract more people as possible. In the past I didn’t join the previous tournaments just because the time to complete each match was too short and not compatible with my work rythm (at least 3 days a week I’m out of home from 8.00 to 20.00 and in this way you begin the game already loosing 1 ½ day). Moreover this format works better with players situated in a totally different world zone where every gamer looses a lot of time waiting the opponent’s move (like happened in my last game vs massina-nz, when I was fearing of running out of time) and also if playing more complex scenarios.

2 – The score system works quite well, though it could be improved. As I stated earlier on this thread, the system envisaged by Triarius seems more balanced (expecially when adopting the additional suggestion made by dazzam) but the complex results of the two methods don’t differ so much at the end. If the guys at Slitherine prefer to maintain the actual system it should be anyway stressed that probably a score of 20 points for a Bye or a scratched game should be truly enough.

3 – Probably it would be wise to clarify in a concise text the exact way the system calculates the score.

And also one question:

1 – Would it be possible to use the Slitherine Tournament engine for mastering player organized competitions like the “League of Extraordinary Gentlemen”, the “RoR Companions Cup” or the “Challenger’s Tournament”?

tofman04 wrote:Hi,

I don't understand anything. Probably I'll quit to play this tournaments.
Before the tournament was over, in my last games the score was 27+25 vs 3+5 for me and it was my opponent turns.
The tournament finished and I won the two games (timeout victory) but the score now is 16+27 vs 14+3 for me.
I don't understand why I have to loose 9 pts.

I am puzzled too - The points in my last game with ianiow reflect the message I got at 3pm Friday - defeat (timed out) in both games.
However it was ianiow's turn in both games fromm late on the 27th.
How does that work?