Seeding unrated players for Swiss tournaments

[Kevin Bonham]

Quite a few people not aware of proper practice for seeding unrated players so I'm making a sticky thread to provide for information and discussion on the rules and possible approaches.

When a player has no rating within the system you are using, it is frequently not correct to put them in at zero. This is covered in the FIDE General Handling Rules For Swiss Tournaments:

https://www.fide.com/fide/handbook.h...4&view=article

Before the start of the tournament, a measure of the player’s strength is assigned to each player. The strength is usually represented by rating lists of the players. If one rating list is available for all participating players, then this rating list should be used.
It is advisable to check all ratings supplied by players. If no reliable rating is known for a player, the arbiters should make an estimation of it as accurately as possible.

(my bold)

The 2016 Arbiters Handbook has some advice about doing this for FIDE-rated tournaments:

The estimated rating of an unknown player can be determined on the basis of a national rating, if available, using the appropriate conversion formulas

When a player has a national rating, but no FIDE rating, we can convert the first to an equivalent value* ‐* in some cases directly, in others by using appropriate formulas. When a player has no rating at all, we shall usually need to estimate its [sic] strength according to current practices and national regulations.

As well as using a national rating for a FIDE-unrated player in a FIDE-rated event, the reverse is extremely common (an ACF-unrated player with a FIDE rating should be seeded with their FIDE rating for an event seeded in ACF rating order).

For more tricky cases here are some other possible solutions:

* Use a performance rating based on all the results of games the player has played that you can find. Pax's calculator http://www.paxmans.net/performance_calc.php is excellent for this purpose.
* Use a player's rating from a different time control within the same or a comparable rating system.
* Use a player's online rating - with caution as some online rating systems are inflated, I tend to deduct about 200 points in these cases.
* Get an estimate of the player's strength from other players they have played casual games against, if you know any.

When unrated players are weak juniors or adults who have never played in tournaments at all, no harm is done by putting them in at zero. But when they are strong adults it can make a serious mess of the pairings, especially in small tournaments.

Of course when an unrated player enters a 100-player tournament five minutes before the start of round 1 an arbiter may not have time for any of this stuff. But where an arbiter does have the time they should try to seed the player based on a proper estimate and not zero.

[ElevatorEscapee]

One variation I've sometimes seen in touranment entry conditions is that the arbiter may revise the estimated provisional rating after a certain number of rounds (eg 3) based on the unrated player's performance in the current tournament.

So if a player is given an estimated provisional rating of 1400 and beats three 1900+ rated players in a row, the provisional rating can be revised upward for future pairings.

[Ian Rout]

One variation I've sometimes seen in touranment entry conditions is that the arbiter may revise the estimated provisional rating after a certain number of rounds (eg 3) based on the unrated player's performance in the current tournament.

So if a player is given an estimated provisional rating of 1400 and beats three 1900+ rated players in a row, the provisional rating can be revised upward for future pairings.

I probably wouldn't do this, on the basis that you can't undo what has already happened but you can make it worse. Changing seeding order mid-stream can cause more instability than a wrong seeding order.

For instance somebody may have had a couple of easy (or hard) pairing) on the basis of a dodgy rating, but just when they are about to get one that evens it up the change of rating gives them another in the same direction. Or others in the same score group are shuffled to other positions in the group which again causes further distortion.

So my suggestion is to keep it as is, so everybody is paired on a consistent basis throughout the event (though it may be appropriate to reclassify somebody for prize eligibility purposes).

[MichaelBaron]

I believe as long as they are eligible for unrated prize only (in case there is one) rather than rating group prizes - organizers should be able to seed them as they ''feel''.

[ElevatorEscapee]

^^^ Well, the issue may well be that the unseeded player either actually wins the tournament, or finishes very high after playing some very strong opponents....

^^^ In which case, surely it is not the players' fault that the arbiter initially gave him a rating that was obviously below the player's actual strength?

Why wouldn't an unrated player who wins all his/her games be not eligible for the grand prize?

[Craig_Hall]

^^^ Well, the issue may well be that the unseeded player either actually wins the tournament, or finishes very high after playing some very strong opponents....

^^^ In which case, surely it is not the players' fault that the arbiter initially gave him a rating that was obviously below the player's actual strength?

Why wouldn't an unrated player who wins all his/her games be not eligible for the grand prize?

Open prizes aren't rating group prizes.

[MichaelBaron]

^^^ Well, the issue may well be that the unseeded player either actually wins the tournament, or finishes very high after playing some very strong opponents....

^^^ In which case, surely it is not the players' fault that the arbiter initially gave him a rating that was obviously below the player's actual strength?

Why wouldn't an unrated player who wins all his/her games be not eligible for the grand prize?

Eligible for grand..but not for group prizes!

[DontAcceptDraw]

In my tournaments, when unrated players are juniors I create a virtual ELO for those player as follows:

• I add his age to his zero rating FIDE.
• If unrated player has FIDE ID then I add 10 to his rating fide

Example: http://chess-results.com/tnr318716.aspx?lan=2

[strollingrhino]

Can I get some advice regarding an unrated player in our current event please... They are from the Philippines and after some round 2 research, being 2/2, I found that they were listed on the NCFP list, the Philippines national list, as rated 1693 and has a chess.com of 1998 (high of 2126 April this year), no FIDE ID.
Our event is an 8 round swiss - ACF rated with 21 entrants and he is currently seed 20.

Q1 - Should I assign him a rating for the next 6 rounds and reseed the event from round 3?
Q2 - What rating should I assign him? Does anyone know an NCFP to ACF conversion?

[Garvinator]

Can I get some advice regarding an unrated player in our current event please... They are from the Philippines and after some round 2 research, being 2/2, I found that they were listed on the NCFP list, the Philippines national list, as rated 1693 and has a chess.com of 1998 (high of 2126 April this year), no FIDE ID.
Our event is an 8 round swiss - ACF rated with 21 entrants and he is currently seed 20.

Q1 - Should I assign him a rating for the next 6 rounds and reseed the event from round 3?
Q2 - What rating should I assign him? Does anyone know an NCFP to ACF conversion?

I would be re-seeding the player, but it is difficult to give you a specific answer without the following information.

The name of the player - So I, or other on here, could do our own research to see if we can find any other information on the player. We might be able to find some more information on the player for you to use.

The list of players in the event and also the current cross table in score order. This will give us information on the ratings of the other players and their current standings.

[ER]

...
The list of players in the event and also the current cross table in score order. This will give us information on the ratings of the other players and their current standings.

So will this be calculated on
I) individual rating of other players?
II) average rating of other players?
III) both?

[Patrick Byrom]

This website should help. The player's rating (effectively) of 1700 on the NCFP list converts to 1950 USCF which converts to 1800 FIDE, which should be equivalent to 1800 ACF - so his 2/2 is not unexpected!

[strollingrhino]

I would be re-seeding the player, but it is difficult to give you a specific answer without the following information.

The name of the player - So I, or other on here, could do our own research to see if we can find any other information on the player. We might be able to find some more information on the player for you to use.

The list of players in the event and also the current cross table in score order. This will give us information on the ratings of the other players and their current standings.

Ok - The event is the GCCC Steiner Cup and standings are posted on chesschat.
Player is Simon Dolosa

[strollingrhino]

I would be re-seeding the player, but it is difficult to give you a specific answer without the following information.

The name of the player - So I, or other on here, could do our own research to see if we can find any other information on the player. We might be able to find some more information on the player for you to use.

The list of players in the event and also the current cross table in score order. This will give us information on the ratings of the other players and their current standings.

Here's the Round 2 Crosstable, Thanks for assisting I'm a new NA:

2021 GCCC Steiner Cup
 Cross Table at round 2 sorted by score

Pos NAME                           Rtg   T  Fed  Pts |   1     2    |  Buc1    BucT    DirE  
---------------------------------------------------------------------------------------------
  1 Ford,Daniel                    1590     QLD  2.0 | +B11  +W8    |    1.0     2.0     0.0 
  2 Dolosa,Simon                      0     QLD  2.0 | +W16  +W10   |    1.0     1.5     0.0 
  3 Morris,Byron                   1731     QLD  2.0 | +W18  +B7    |    1.0     1.0     0.0 
  4 Croucher,Toshio                1462     QLD  2.0 | +B19  +W9    |    1.0     1.0     0.0 
  5 Van Der Byl,Orsen              1252     QLD  1.5 | +F20  =W6    |    1.5     2.0     0.0 
  6 Kitikov,Oleg                   1511     QLD  1.5 | +W12  =B5    |    1.0     1.5     0.0 
  7 Barkley,Christian              1429     QLD  1.0 | +B14  -W3    |    2.0     3.0     0.0 
  8 Lucas,Peter                    1153     TAS  1.0 | +W15  -B1    |    2.0     3.0     0.0 
  9 Smith,Tristan K                1064     QLD  1.0 | +W13  -B4    |    2.0     3.0     0.0 
 10 Ming,Kenny                     1447     QLD  1.0 | +W21  -B2    |    2.0     2.0     0.0 
 11 Smith,Mathew                    797     QLD  1.0 | -W1   +B21   |    2.0     2.0     0.0 
 12 Inukai,Rentaro                  753     QLD  1.0 | -B6   +F20   |    1.5     1.5     0.0 
 13 Luey,Ivan                      1905     NSW  1.0 | -B9   +W17   |    1.0     1.5     0.0 
 14 Rawlley,Aarav                   100     QLD  1.0 | -W7   +B18   |    1.0     1.0     0.0 
 15 Mortoni,Alan                      0     QLD  1.0 | -B8   +W19   |    1.0     1.0     0.0 
 16 Breeden,Marek                  1155     QLD  0.5 | -B2   =BYE   |    2.0     2.5     0.0 
 17 Tong,Justin                     663     QLD  0.5 | =BYE  -B13   |    1.0     2.0     0.0 
 18 Consiglio-Cockle,Jared          920     QLD  0.0 | -B3   -W14   |    2.0     3.0     0.0 
 19 Inukai,Kotaro                   676     QLD  0.0 | -W4   -B15   |    2.0     3.0     0.0 
 20 Benapuria,Awadesh                 0     QLD  0.0 | -F5   -F12   |    1.5     2.5     0.0 
 21 Qu,Harrison                     502     QLD  0.0 | -B10  -W11   |    1.0     2.0     0.0 

Generated by Vega - www.vegachess.com
(copy licensed to: Gold Coast Chess Club - AUS)

[Garvinator]

Looking at that cross table and Pat's reply, I think anything from 1600 to 1800 is ok based on his two results. I was hoping that in his second game, Simon had defeated at a player rated around 1600, then I would say 1800 with more confidence.