--- In tsh-users@yahoogroups.com, "G. Vincent Castellano" <gvc@...> wrote:

> Are there any tricks to dealing with drop-outs?  If I pair someone
> with player 0 for the remainder of an event, will I have trouble
> with generating pairings?

In general, the only times you should have to manually pair someone
with 0 are when they have been included in computed pairings that
won't be modified by their absence, and it's only the affected rounds
where the pairings need to be adjusted.

Otherwise, quoting from http://tsh.poslfit.com/doc/trouble.html :

> If a player gives you notice that he will be missing one or more
> games, or you end up determining that he is a no-show after the
> tournament has begun, you can prevent them from being considered
> for pairings using the "EditScore" command. If player 4 in division
> A should no longer be paired, enter "es a 4" to begin editing the
> player's scoresheet, then enter "off -50" give them forfeit losses
> whenever pairings are computed. ("off 50" will give them a 50-point
> bye, and "off 0" will record a missed game without assigning a win
> or loss). If the player subsequently returns to the tournament, you
> can re-enable their pairings by entering "on" while editing their
> scorecard.

John
--
John Chew (poslfit on MD) * jjchew@... * http://www.poslfit.com

Are there any tricks to dealing with drop-outs?  If I pair someone with player 0
for the remainder of an event, will I have trouble with generating pairings?

In particular, as an extreme case, I gave a player a bye for the first three
rounds and then did 'smp 1-3' for that division.  tsh then complained that
"Division D already has pairings. [ehaspair]" even though the only pairing was
that one player's byes.

Clearly this is an edge condition and would not be a real problem, as at the
start I would just give Byes to whoever was matched with Ms. Missing.  Are there
any other cases where setting up Byes in advances will give the pairing commands
indigestion?

Relevant config segments or the condition above:

config pairing_system = 'chew'
config max_rounds = 7
config initial_schedule = 3
config prize_bands{'a'} = [1,2,3]
config prize_bands{'b'} = [1,2,3]
config prize_bands{'c'} = [1,2,3]
config prize_bands{'d'} = [1,2,3]
config director_name = "Mary-Lou Goetz"
config event_date = "May 9, 2009"
config event_name = "Annapolis Scrabble Tour 2009"
config force_koth = 1
# config session_breaks = [12]
config no_boards = 1
config gibson = 1
config pair_page_break = 25

autopair d 0 1 if 3 d
autopair d 2 4 ns 0 2 d
autopair d 4 5 ns 0 4 d
autopair d 5 6 koth 0 5 d
autopair d 6 7 koth 1 6 d

--- In tsh-users@yahoogroups.com, "G. Vincent Castellano" <gvc@...> wrote:
> Now, if I want to use these Chew pairings for the first 5, and two KotH at the
> end, is there a way to do this without putting in autopair commands for
> everything?

I would put in autopair commands for everything. There's no
easy way to do this, because I don't think it's a good idea to
have more than one KOTH at the end of an event.

John
--
John Chew (poslfit on MD) * jjchew@... * http://www.poslfit.com

John Chew wrote:
> --- In tsh-users@yahoogroups.com <mailto:tsh-users%40yahoogroups.com>,
> "G. Vincent Castellano" <gvc@...> wrote:

>  > If I remove the autopair directives, it appears that Chew pairings
> refuses to
>  > pair more than the first three rounds, even if initial_schedule is
> set to 5.
>
> It works fine for me. Do you have a recent copy of tsh, and if
> so, can you send me the offending config.tsh file again?

If I see this problem again I'll package everything up for you, but now I can't
make it fail in this way, so never mind.

Now, if I want to use these Chew pairings for the first 5, and two KotH at the
end, is there a way to do this without putting in autopair commands for
everything?  I suppose I could add the autopair commands to the config.tsh after
I've generated the first 5 rounds, right?
--gvc

--- In tsh-users@yahoogroups.com, "G. Vincent Castellano" <gvc@...> wrote:
>
> John Chew wrote:
> >
> > config initial_schedule only works to modify default (Chew)
> > pairings, and what you have here are what I confusingly call
> > automatic pairings, using the autopair configuration command.
>
> So the presence of an 'autopair' directive means I also have to designate how
> the initial rounds are paired, right?

Correct.

> If I remove the autopair directives, it appears that Chew pairings refuses to
> pair more than the first three rounds, even if initial_schedule is set to 5.

It works fine for me. Do you have a recent copy of tsh, and if
so, can you send me the offending config.tsh file again?

> I tried using InitFontes for the first 5 rounds, but one thing I don't like
> about the resulting pairings is that they fail the "sum of opponents ratings"
> test--I don't have the exact results handy, but this disparity was evident in
> the case of two players at top and bottom of the field who shared 4 of the
same
> opponents, and whose 5th was mismatched, which would seem to greatly favour
the
> higher-rated player.

Hard to comment without seeing the specific pairings data,
but in general in any sort of round robin scenario, high rated
players will face an easier field than low rated ones.

John
--
John Chew (poslfit on MD) * jjchew@... * http://www.poslfit.com

John Chew wrote:
>
> config initial_schedule only works to modify default (Chew)
> pairings, and what you have here are what I confusingly call
> automatic pairings, using the autopair configuration command.

So the presence of an 'autopair' directive means I also have to designate how
the initial rounds are paired, right?

My confusion stemmed from the fact that in past I had added 'autopair'
directives after the event was underway.

If I remove the autopair directives, it appears that Chew pairings refuses to
pair more than the first three rounds, even if initial_schedule is set to 5.

I tried using InitFontes for the first 5 rounds, but one thing I don't like
about the resulting pairings is that they fail the "sum of opponents ratings"
test--I don't have the exact results handy, but this disparity was evident in
the case of two players at top and bottom of the field who shared 4 of the same
opponents, and whose 5th was mismatched, which would seem to greatly favour the
higher-rated player.
--gvc

Looks fine to me. I agree with you that your pairing_system and
initial_schedule lines are now redundant. I'd say the same goes for
force_koth too since in your autopair lines you have specified KOTH
for the final rounds anyway.

Stewart



2009/5/5 G. Vincent Castellano <gvc@...>:
> Thanks to the DRYrun command, I think I've got the config.tsh shaken out.  Now
> I'm looking for a sanity-check on my whole program.
>
> We have 84 registrants in 4 divisions for 7 rounds.  I'm doing the data entry
on
> my own, and nothing but computer operations.  I'll use Fontes for 1-3, and
> NewSwiss for Round 4 (based on results from 2).  There is a lunch break after
4,
> so I will do NewSwiss for 5 based on 4, then 2 KotH (no-repeat followed by one
> allowed) for 6 and 7.
>
> We have the hall from 8 to 6, so are planning on starting at 9 and ending at 5
> sharp.
>
> Is this all workable?  Any hidden glitxhes with the config below?  (autopair
> commands for b, c, and d divisions are identical to a.)  I believe the
> 'pairing_system' and 'initial_schedule' are obviated by the autopair commands.
>
>
>
> config pairing_system = 'chew'
> config max_rounds = 7
> config initial_schedule = 3
> config prize_bands{'a'} = [1,2,3]
> config prize_bands{'b'} = [1,2,3]
> config prize_bands{'c'} = [1,2,3]
> config prize_bands{'d'} = [1,2,3]
> config director_name = "Mary-Lou Goetz"
> config event_date = "May 9, 2009"
> config event_name = "Annapolis Scrabble Tour 2009"
> config force_koth = 1
> # config session_breaks = [12]
> config no_boards = 1
> config gibson = 1
> config pair_page_break = 25
>
> autopair a 0 1 if 3 a
> autopair a 2 4 ns 0 2 a
> autopair a 4 5 ns 0 4 a
> autopair a 5 6 koth 0 5 a
> autopair a 6 7 koth 1 6 a
>
> --gvc
>
>
> ------------------------------------
>
> Yahoo! Groups Links
>
>
>
>

Thanks to the DRYrun command, I think I've got the config.tsh shaken out.  Now
I'm looking for a sanity-check on my whole program.

We have 84 registrants in 4 divisions for 7 rounds.  I'm doing the data entry on
my own, and nothing but computer operations.  I'll use Fontes for 1-3, and
NewSwiss for Round 4 (based on results from 2).  There is a lunch break after 4,
so I will do NewSwiss for 5 based on 4, then 2 KotH (no-repeat followed by one
allowed) for 6 and 7.

We have the hall from 8 to 6, so are planning on starting at 9 and ending at 5
sharp.

Is this all workable?  Any hidden glitxhes with the config below?  (autopair
commands for b, c, and d divisions are identical to a.)  I believe the
'pairing_system' and 'initial_schedule' are obviated by the autopair commands.



config pairing_system = 'chew'
config max_rounds = 7
config initial_schedule = 3
config prize_bands{'a'} = [1,2,3]
config prize_bands{'b'} = [1,2,3]
config prize_bands{'c'} = [1,2,3]
config prize_bands{'d'} = [1,2,3]
config director_name = "Mary-Lou Goetz"
config event_date = "May 9, 2009"
config event_name = "Annapolis Scrabble Tour 2009"
config force_koth = 1
# config session_breaks = [12]
config no_boards = 1
config gibson = 1
config pair_page_break = 25

autopair a 0 1 if 3 a
autopair a 2 4 ns 0 2 a
autopair a 4 5 ns 0 4 a
autopair a 5 6 koth 0 5 a
autopair a 6 7 koth 1 6 a

--gvc

Stewart, for some reason, even though your answer was exactly right, I just
cruised right over it.

Stewart Holden wrote:
>
> If you are using autopair lines for some of the rounds then you need to
> specify pairings for ALL the rounds, even if you just want the default
> Chew pairings. So you'll need some more autopair lines in your config..
> I'm not sure exactly what you mean by doing 5 pre-set rounds.

I meant I want to generate pairings for the first 5 rounds at the start of the
event.

> Incidentally, I've never seen anyone use division numbers rather than
> letters. In the UK at least it is always Div A, Div B etc. and the

I was using the division "names" in CrossTables, but I'm switching to lettered
divisions because otherwise it makes the tsh command-lines too confusing.
--gvc

--- In tsh-users@yahoogroups.com, "G. Vincent Castellano" <gvc@...> wrote:
>
> We have a 7-round one-day with 4 rather large divisions.  I plan to do 5
pre-set
> rounds followed by 2 KotH (0 repeats, then 1 allowed).
>
> I have 2 questions:
>
> 1) Why am I getting "Error: No pairings yet for Division 1 Round 1.
[enopryet]"
> when I use "sp 1 1" to generate a single round's pairings?

config initial_schedule only works to modify default (Chew)
pairings, and what you have here are what I confusingly call
automatic pairings, using the autopair configuration command.

To get the effect you want, use autopair with the initfontes
command.

> 2) Do I need to do anything to enforce Gibsonization in the KotH rounds?

config gibson = 1

John
--
John Chew (poslfit on MD) * jjchew@... * http://www.poslfit.com

We have a 7-round one-day with 4 rather large divisions.  I plan to do 5 pre-set
rounds followed by 2 KotH (0 repeats, then 1 allowed).

I have 2 questions:

1) Why am I getting "Error: No pairings yet for Division 1 Round 1. [enopryet]"
when I use "sp 1 1" to generate a single round's pairings?

2) Do I need to do anything to enforce Gibsonization in the KotH rounds?

Transcript of my command-line session, including a dump of the config.tsh:

[gvc@localhost tsh]$ ./tsh.pl
Using most recent config.tsh in 2009-05-09-AnnapolisMD.
Loading configuration file 2009-05-09-AnnapolisMD/config.tsh. [iloadcfg]
Welcome to tsh version 3.280. [iwelcome]
Loading Division 1. [iloaddiv]
Loading Division 2. [iloaddiv]
Loading Division 3. [iloaddiv]
Loading Division 4. [iloaddiv]
Odd number of active players (21) in Division 4. [iodddiv]
Event loaded: Annapolis Scrabble Tour 2009, May 9, 2009. [ievtname]
tsh> sp 1 1
Error: No pairings yet for Division 1 Round 1. [enopryet]
tsh> q
tsh quits. [iquit]
[gvc@localhost tsh]$ cat 2009-05-09-AnnapolisMD/config.tsh
division 1 1.t
division 2 2.t
division 3 3.t
division 4 4.t

config max_rounds = 7
config initial_schedule = 5
config prize_bands{'1'} = [1,2,3]
config prize_bands{'2'} = [1,2,3]
config prize_bands{'3'} = [1,2,3]
config prize_bands{'4'} = [1,2,3]
config director_name = "Mary-Lou Goetz"
config event_date = "May 9, 2009"
config event_name = "Annapolis Scrabble Tour 2009"
config force_koth = 1
# config session_breaks = [12]
config no_boards = 1

autopair 1 5 6 koth 0 6 1
autopair 1 6 7 koth 1 7 1

autopair 2 5 6 koth 0 6 2
autopair 2 6 7 koth 1 7 2

autopair 3 5 6 koth 0 6 3
autopair 3 6 7 koth 1 7 3

autopair 4 5 6 koth 0 6 4
autopair 4 6 7 koth 1 7 4

It's been way too long since I've formally released a new version of tsh,
as I've been swamped setting up NASPA.  I'd like first of all to thank
Ross Brown for contributing code to the ScoreBoard command (a first time
ever, I think, for user-contributed code), enhancing it to display players
who are out of contention in blue.

I am a little behind on replying to user requests for new features and
bug reports; please feel free to remind me of them if I don't reply within
a week.  As far as the Vista SUBMIT issue is concerned, Bryan Pepper was
able to submit tournament results from Vista without trouble, from his
first TSH tournament, so it must be a user-specific configuration issue
(ugh).  If you have trouble with a Vista SUBMIT, please contact me privately.

So here is what is 'new' for 3.270, though I imagine most users have already
seen these features in regular UPDATEs.

* util/mirror-ftp, which keeps a copy of your TSH files on an FTP server,
   and is a cheap way to provide event coverage, is ready for public release.
   It's easy to set up and use, and has been tested on OS/X, Windows and LInux.
* automatable tweeting to twitter.com
* multiple enhancements to ScoreBoard
* new pairing systems: bracket, Guelph, five-round NAST
* new rating systems: aus, elo, nsa2008 (new default for nsa realm), wespa
* new rating list: deu (Germany) - if you want support for your own rating list
   please contact me.
* post-tournament ratings are stored under extention field 'newr'
* config gibson = 'strict' ignores repeats when finding a victim
* config html_index_bottom and html_bottom correspond to ..._top
* config custom_stylesheet lets you override values in tsh.css
* new report commands: HighCombined, TotalTeamStandings
* new prize categories: overseed, roundrecord, separator, totalteam
* new triggers (hooks): hook_division_update
* bug fixes to: ABSP ratings, Addscore, bye assignment, ChewPair,
   HTTP client, KOTH, LCT ratings, PRiZEs, RANDomscores, SUBMIT, UPSETs,
   UpdatePLAYers
* preparation for self-service scoring system

Thank you all for your continuing feedback that lets me keep improving tsh.

John
--
John Chew (poslfit on MD) * jjchew@... * http://www.poslfit.com

--- In tsh-users@yahoogroups.com, "G. Vincent Castellano" <gvc@...> wrote:
>
> Has anyone used tsh on Vista when uploading results?  Carole Denton got an
error
> when doing the submission from her Vista laptop, and since I refuse to waste
any
> time debugging networking on Vista I just did it from my machine so I don't
have
> any further diagnostics.

Alternately, if anyone has tsh on Vista, and the time to work
with me to see what if anything needs to be done to make rating
submission work from Vista, please contact me privately.

John
--
John Chew (poslfit on MD) * jjchew@... * http://www.poslfit.com

Has anyone used tsh on Vista when uploading results?  Carole Denton got an error
when doing the submission from her Vista laptop, and since I refuse to waste any
time debugging networking on Vista I just did it from my machine so I don't have
any further diagnostics.
--gvc

I've attached config and .t file for division A as of round 8 from the event I
assisted with last weekend.  We did round-robin for 8 rounds (Day 1), then I
tried to do Chew pairings for the next day's morning session.  This resulted in
player 10 (ranked 15 of 16) matched against player 12 (ranked 13 of 16) in both
rounds 10 and 12.

I resolved this by backing out Chew Pairings and specifying "ns 0" (I may have a
followup question on this later).

I have a particular aversion to repeat pairings because they tend to generate a
lot of complaints, and having repeats two games apart is especially problematic.
   Should I not be using "cp" to pair this many rounds at once?  I would prefer
to do "one game ahead" but I needed to minimize the amount of pairings generated
outside of the breaks.

Is there any prospect of Wikifying the tsh documentation?  In particular, I
think a 'gallery' of config files for particular event setups would be useful.

--gvc

division a a.t
division b b.t
division c c.t
division d d.t

config max_rounds = 14
config initial_schedule = 8
config prize_bands{'a'} = [1,2,3]
config prize_bands{'b'} = [1,2,3,4]
config prize_bands{'c'} = [1,2,3]
config prize_bands{'d'} = [1,2,3,4]
config director_name = "Ted Gest"
config event_date = "March 21-22, 2009"
config event_name = "Washington DC/Rockville"
config force_koth = 1
config session_breaks = [12]
config no_boards = 1

autopair a 8 9 cp 8 a
autopair a 8 10 cp 8 a
autopair a 8 11 cp 8 a
autopair a 8 12 cp 8 a
autopair a 12 13 koth 1 12 a
autopair a 13 14 koth 2 13 a

autopair b 8 9 cp 8 b
autopair b 8 10 cp 8 b
autopair b 8 11 cp 8 b
autopair b 8 12 cp 8 b
autopair b 12 13 koth 1 12 b
autopair b 13 14 koth 2 13 b

autopair c 8 9 cp 8 c
autopair c 8 10 cp 8 c
autopair c 8 11 cp 8 c
autopair c 8 12 cp 8 c
autopair c 12 13 koth 1 12 c
autopair c 13 14 koth 2 13 c

autopair d 8 9 cp 8 d
autopair d 8 10 cp 8 d
autopair d 8 11 cp 8 d
autopair d 8 12 cp 8 d
autopair d 12 13 koth 1 12 d
autopair d 13 14 koth 2 13 d

I'll be using an experimental new feature of tsh this weekend to
automatically report on the top players each round at the Dallas
Open via the social network Twitter.  This is in addition of course
the usual coverage that Sherrie and I provide.

If you are a Twitter user and want to receive the DO "tweets",
you should follow the "NASPA" user on Twitter.

If you do not want to join Twitter but are curious what I'm talking
about, you can see the NASPA twitter feed at either

http://www.scrabbleplayers.org

in the box on the right that may be black if you don't have Flash
or at

http://www.twitter.com/NASPA

John
--
John Chew (poslfit on MD) * jjchew@... * http://www.poslfit.com

Hi all

I've been meaning to ask this for a while- an issue that came up at
my Round Robin tourney in November was the ability (or lack thereof)
to enter tournament results 'out of order', as it were, particularly
in RRs.

The scenario on the day was that due to a scheduling conflict, one
player had to leave early, and to accommodate her, I effectively
switched her player number with that of another player's, thinking I
could enter the results anyway- e.g. if player A was due to play
player C in round 2 and player B played player C in round 6, players
B and C would actually play in round 2 but I'd still be able to enter
the result as a round 6 result despite it still only being round 2.
Turns out that wasn't possible.

I have since been asked to use tsh for the local club's annual
leagues, which are notorious for remaining incomplete at the end of
the year. As they're round robins where every fixture will be decided
before the first game has been played, is there a way for me to enter
the results willy-nilly (as it were)?

cheers

BJ"coherent"W

If I understand the distinction correctly, I think statically would
be o.k., as the information could be updated at the end of each
round, like standings and ratings information.

I think it adds a lot to live coverage to have the player information
pages, like the following from nationals:

http://www.scrabble-assoc.com/tourneys/2008/nsc/build/player/

With information on each player like this:

http://www.scrabble-
assoc.com/tourneys/2008/nsc/build/player/1/001.html

If tsh could generate that, it would be great.  If there is a
convenient way to use tsh data to generate these types of pages
manually, I would be interested to know that also.

Dallas Johnson

--- In tsh-users@yahoogroups.com, "John Chew" <jjchew@...> wrote:
>
> --- In tsh-users@yahoogroups.com, "Dallas Johnson" <dallasjohnson@>
> wrote:
> >
> > What command (or process) creates the html pages in tsh that show
each
> > player on his/her own page, with a list of all opponents and the
> > round-by-round results, including photos?  I have looked for this
in the
> > manual, but I am missing it somehow.
>
> There is none yet, mainly as I'm not sure whether it's better
> to generate that volume of content dynamically or statically.
> Suggestions are welcome.
>
> John
> --
> John Chew (poslfit on MD) * jjchew@... *
> http://www.poslfit.com
>

--- In tsh-users@yahoogroups.com, "Andy Saunders"
<andysaunders1983ca@...> wrote:
>
> I'm surprised to be making this post, because I've never seen TSH make
> incorrect ratings calculations, but I'm not sure if the ratings
> calculations presented by TSH at the Mississauga Open this weekend are
> entirely accurate.

It's possible that a recent version of tsh does not compute
acceleration points correctly.  I have been rewriting the
ratings estimation routines for tsh to exactly match the NSA's
calculations (including esoterica such as the feedback bug:
did you know that whether or not you receive feedback points
contribute to acceleration points (and therefore to other
feedback points) depends on the position of your player number
in the roster?), and building in support for three additional
ratings systems (Australian, German and WESPA), and will have
a new version posted and tested in time for this weekend.

John
--
John Chew (poslfit on MD) * jjchew@... *
http://www.poslfit.com

--- In tsh-users@yahoogroups.com, "Dallas Johnson" <dallasjohnson@...>
wrote:
>
> What command (or process) creates the html pages in tsh that show each
> player on his/her own page, with a list of all opponents and the
> round-by-round results, including photos?  I have looked for this in the
> manual, but I am missing it somehow.

There is none yet, mainly as I'm not sure whether it's better
to generate that volume of content dynamically or statically.
Suggestions are welcome.

John
--
John Chew (poslfit on MD) * jjchew@... *
http://www.poslfit.com

I'm surprised to be making this post, because I've never seen TSH make
incorrect ratings calculations, but I'm not sure if the ratings
calculations presented by TSH at the Mississauga Open this weekend are
entirely accurate. I'm posting this to CGP in case others want to
check their calculations from the weekend, and to TSH-users in case
there's a bug that needs to be fixed, or some sort of documentation
needed to ensure that other directors have TSH set up correctly to do
the proper calculation. Of course, if I've made a mistake as well, I
fully expect that someone will correct me. Please let me know.

http://shellstevens.com/moo/A-ratings-016.html is the URL to the final
standings of the MOO.

As an example, I'm going to go through and calculate my rating change
manually. I will be verbose in my calculations for the benefit of
those not entirely familiar with the ratings system.

Initial rating: 1331
Opponent 1: 1717 (ratings difference -386, win expectation 0.086)
2: 1656 (-325, 0.125)
3: 1458 (-127, 0.327)
4: 1890 (-559, 0.024)
5: 1567 (-236, 0.202)
6: 1805 (-474, 0.047)
7: 1217 (+114, 0.657)
8: 1412 (-81, 0.387)
9: 1210 (+121, 0.666)
10: 1596 (-265, 0.174)
11: 1466 (-135, 0.317)
12: 1519 (-188, 0.253)
13: 1251 (+80, 0.611)
14: 1431 (-100, 0.362)
15: 1670 (-339, 0.115)
16: 1805 (-474, 0.047)

Total win expectation: 4.400
Number of wins: 10.0
Over expectation: 5.600
Ratings multiplier: 20
Base increase: 112 points
Acceleration bonus for a gain exceeding 5 points / round: equal to
amount of gain greater than 5 points / round, which is 112 - (16 * 5)
= 32 points
Ratings increase, including bonus: 144 points.
Feedback points: Opponent 9 gained 94 "base" points over the weekend
(garnering 14 bonus points) and Opponent 13 gained 84 "base" points (4
bonus). As each opponent that played the player getting bonus points
gets 0.05 feedback points for each bonus point earned, I should get
((14 + 4) * 0.05) = 0.9 feedback points.
Total ratings increase: 144.9 points, rounded up to 145

New rating: 1476

It appears as if the ratings calculator used at the MOO did not take
the bonus points into account when making the calculations.

--
Andy Saunders
andy@...

Vince writes:

> Doesn't the 'autopair' (config file directive) take the division as
the first
> argument:
>
> Each auto-pairing configuration line consists of the command word
"autopair", a
> division name, the number of the round whose results must all be in
before
> pairings can be calculated, the number of the round whose pairings
will be
> calculated, and then a tsh command that generates the pairings.
>
> So we'd have
>
> autopair a 7 8 ns 0 7

That's correct.

John
--
John Chew (poslfit on MD) * jjchew@... *
http://www.poslfit.com

John Chew wrote:
> Otherwise the usual solution solution in tsh is to pair all of the
> rounds in a session based on the standings at the beginning of the
> session, either Swiss or Chew:
>
> autopair 7 8 ns 0 7
> autopair 7 9 ns 0 7
> autopair 7 10 ns 0 7
> autopair 7 11 ns 0 7
>

Doesn't the 'autopair' (config file directive) take the division as the first
argument:

Each auto-pairing configuration line consists of the command word “autopair”, a
division name, the number of the round whose results must all be in before
pairings can be calculated, the number of the round whose pairings will be
calculated, and then a tsh command that generates the pairings.

So we'd have

autopair a 7 8 ns 0 7

etc. etc.
--gvc

Vince writes:

> Because the DC event is going to be operated by personnel without a lot
> of technical background, or experience with tsh, I am recommending that
> they set up their sessions in advance (except for the last session of
> three) to minimize the pressure to get everything keyed in while the
> event is in progress.
[...]
> config max_rounds = 14
> config initial_schedule = 3

Since you have more than 14 players, you could certainly increase the
length of the initial schedule from three rounds to say seven, to get
the staff through to Saturday lunch.  The downside to this would be
a moderate decrease in pairings accuracy, but since the NSC was paired
this way until 2006 (though granted with twice as many rounds) the
players won't likely mind too much.

Otherwise the usual solution solution in tsh is to pair all of the
rounds in a session based on the standings at the beginning of the
session, either Swiss or Chew:

autopair 7 8 ns 0 7
autopair 7 9 ns 0 7
autopair 7 10 ns 0 7
autopair 7 11 ns 0 7

or

autopair 7 8 cp 7
autopair 7 9 cp 7
autopair 7 10 cp 7
autopair 7 11 cp 7

John
--
John Chew (poslfit on MD) * jjchew@... *
http://www.poslfit.com

I sent the following to JCIII, but then it ocurred to me that the tsh user
community might be able to ffer their suggestions as well.
--gvc

Because the DC event is going to be operated by personnel without a lot
of technical background, or experience with tsh, I am recommending that
they set up their sessions in advance (except for the last session of
three) to minimize the pressure to get everything keyed in while the
event is in progress.

Unfortunately, I am not available to support the event.

I'll send the event files along later, after we're done with today's
training session.

The sessions go:

Fri PM:  4 games
Sat AM:  3 games
Sat PM:  4 games
Sun AM:  3 games

In config.tsh we have:

config max_rounds = 14
config initial_schedule = 3

To simplify initial setup, I wanted to provide 'canned' commands so the
initial pairings could be done rapidly as soon as attendance is
established.

I proposed:

sp 1 a
sp 2 a
sp 3 a
cp 0 a

(All of this is just going to be iterated over other divisions, of course.)

One concern I have right off is that the above method gives us
apparently inequitable pairings based on the "sum of opponents'
rankings", which, valid or not, I have seen used as a "measure of
goodness" by players.

The final 3 rounds will be 12-KotH-NoRepeat 13+14-KotH-AsTheyLie, and I
am anticipating that by then the operator will be enough up to speed on
data entry that she will be able to handle the data entry and printing
for the last two rounds on the spot.

My specific question is:  given a scenario where we are going to have
three sessions which I want pairings based on results of the previous
session (or on initialization), what are the preferred config options
and pairing commands?

Thanks for your help.  If you want to call me to discuss any of this you
can reach me at 703.282.3611.
--gvc

Here's what the release notes say (nothing too major):

* Added "config division_rating_system{'A'} = 'none'" which overrides
"config rating_system" on a per-division basis (Andy Saunders).
* Removed "config max_name_length", which is no longer necessary as
all tables are dynamically formatted.
* Removed "config name_format", which is no longer necessary as all
tables are dynamically formatted.
* "RESETEVERYTHING" also deletes "prizes.doc" (Brad Mills).
* "RoundRATings" command now uses vulgar fractions, correctly handles
empty divisions, correctly handles "config no_boards = 1".
* "ShowPairings" command now correctly handles "config no_boards = 1".
* Commands triggered by hooks do not change which report "BrowseLast"
displays.
* "config avoid_sr_runs = 1" is now the default setting in the ABSP
realm (Paul Thomson).
* Added experimental new command "UpdatePLAYers", which uses the new
configuration variable "config cross_tables_id" to choose and download
a player roster from cross-tables.com.

John
--
John Chew (poslfit on MD) * jjchew@... *
http://www.poslfit.com

Major changes in version 3.250:
- Added support for Big Apple Showdown pairings
- Corrected serious issues with Gibsonization
- Rewrote and redesigned electronic ScoreBoard

Minor changes:
- EditScores partially rewritten, allows you to specify career game total
- Improved Chew pairings for small divisions
- PRiZes now computes class prizes
- TeamRoundRobin will sometimes compute necessary byes
- NSA ratings computed correctly for partially scored rounds

Enjoy,
John
--
John Chew (poslfit on MD) * jjchew@... *
http://www.poslfit.com