September’s Just the Rules looks at Split Results:
1 Like
Split results can be rated. The TD simply uses special codes when reporting those games to US Chess: N for a Win, R for a Draw and S for a Loss.
Split results are not required to be rated. Rather, it is up to the TD.
I’ve always appreciated the way this is addressed in the TD/Affiliate Support Area FAQ:
How do I enter split (or inconsistent) results such as a win for one player and a draw for the other player?
First of all, is this really necessary? This topic has come up several times in the US Chess Forums. While there may be reasons to do this for prize purposes, it is harder to justify why that should affect the ratings from that event.
However, the way to do this is to edit the event after uploading it (but before submitting it for rating) to use a different result code for the two players to indicate what the result should be for each player for rating purposes. This will prevent an inconsistent results error. Use this table to determine the result code:
1 Like
In WinTD, you can directly do a “split results” operation which creates separate game records for each player. The US Chess validation will warn about the inconsistent results, but it will go through.
1 Like
Split results (I prefer to call them inconsistent results) are for ratings purposes only, and they can cause confusion among players looking at the crosstable for an event and wondering why a player’s point total varies from what was used at the event for prize purposes. They also interfere with post-hoc activities like trying to compute tiebreaks from a crosstable or look into round-by-round pairing decisions. (It is exceedingly infrequent these days, but pairing decisions could also be impacted by adjourned games, it is probably more common for a result to be corrected later in the event or after the event was rated, and that can also disrupt trying to recreate the tiebreaks that were used for prize purposes.)
My recommendation for the Leago system is to include at least 3 sets of tiebreaks in the reporting structure, probably by creating a new upload format.
This may be a philosophical point, but in at least one instance the inconsistent results are better thought of as for prize purposes only:
22C5. Cancellation of irrevocable byes.
If the director agrees, a player may cancel an irrevocable half-point bye under the condition that if the player wins, the result will be treated as a draw for prize purposes.
I actually suggested the possibility of keeping 3 separate results fields, but that might be TMI. In order of usefulness in post-event crosstables they are:
- Result for ratings purposes
- Result for prize purposes
- Result for pairing purposes
I’m not sure I see the point of #3. (Other than to make a TD’s life miserable ex post). After all, the result for pairing purposes for R2 may be different from the result for pairing purposes for R3. (R2 paired with incorrect result for R1, fixed before R3).
Your point is quite valid, but at the time there was discussion of having Leago check for pairings issues, which I never understood the point of including. But if you’re gonna do it, you probably need to make sure that the record reflects what the TD knew at the time pairings were made, which can change from round to round.
Not only can it change from round to round, but you have “look-ahead” issues such as not giving a bye to someone who is scheduled to miss a later round. Which can only apply if the bye is already “scheduled”—someone who notifies the TD in advance that they will be missing R5 won’t be given an earlier bye; someone who gets a bye and then drops out after R4 would show exactly the same -U- in R5 but the forced bye would be handed out under different circumstances.