The World Warlords II League
Last Revised: 6-Jul-98


Speed Warlords Ideas

The following messages outline ideas for playing "Speed Warlords", similar to "Speed Chess". We outlined the scheme while dreaming up Tournament ideas many months ago, but decided it would have to wait for the WW2L before it would really work. Well, now the WW2L is here, and we just need to see if anyone is motivated enough to seriously try it out! It would be good to playtest it informally and see how it does, before asking Elam to do it.
X-Sender: elam@roo.madness.net
Mime-Version: 1.0
Date: Wed, 28 Jan 1998 11:33:20 -0500
To: "Robert F. Heeter" 
From: "Elam T. Birnbaum" 
Subject: Re:"Speed Warlords"
Cc: jpanagos@cellophane.com, mleung@uclink.berkeley.edu, MTMGsMax@aol.com

Ok, here are my comments...

>I think this is a great idea, but we need to keep in mind that in
>real wars and chess tournaments, the players are spending 100% of
>their time on the game while the game is going on.  In PBEM Warlords,
>we have the problem that players also have real-world lives to
>deal with.

I agree here. Technically, the Remailer right now simulates exactly what
happens in chess. Player 1 makes his move and presses the button of a
dual-clock timer which simultaneously stops his clock and starts the clock
for player 2. Instead of a couple of hours for chess, we have a couple of
days. But, like Bob said, we gave 2 days because people have lives to live
and they do their turn when they have free time. In that sense the Remailer
does take time off their limit even if they don't know its their turn
because they haven't checked their e-mail yet.

>However, we could probably do something like this, which wouldn't
>require making too many tweaks in the Remailer either (?):
>
>(1) Say Player 1 completes Turn 5 and sends gamefile to the Remailer.
>(2) The Remailer sends out the "change-of-turn" notice (as usual),
>    but HOLDS ONTO THE GAMEFILE.  Remailer then holds all
>    archive messages (but sends notice to any player that the
>    archive message was received and held in the queue).
>    (The idea here is that the game is "paused" until the next
>     player is ready to think about his move, and no one gets
>     any of the info needed to plan the move until he's ready.)

This is the only way it would work. You have to deny the player the
gamefile until he is ready. Then it is like saying "ready-set-go" and they
are timed from when they receive the gamefile. And yes, archive messages
would have to be held as well.

This part is not actually that hard to do. I simply create a temporary
archive for the game and any subsequent message gets appended to the
archive. When it is time, the script takes each individual message in the
archive and sends it out before deleting the temp file.

>(3) Player 2 has some fixed amount of time to deal with real life
>    and ignore the game (say the standard 2-day turnaround time).
>(4) When Player 2 is ready to move, he sends a "start the clock"
>    message to the Remailer.  The Remailer handles this by doing
>    3 things:
>       (a) sending the gamefile to all players who should receive it.
>       (b) sending any queued archive messages to all players.
>       (c) starting the timer clock for that player.

This isn't that hard to do either because this action is prompted by an
e-mail, not by a specific time (though there would still have to be an
application running in the backround to enforce the total turn limit).
Adding this ability to the Remailer script wouldn't be too hard, though we
would have to discuss what happens if someone uses up their time.

>(5) Player 2 takes his turn as fast as he likes, and sends the
>       gamefile back to the Remailer.  (Trick: he has to send the
>       gamefile first, wait for the change-of-turn notice from
>       the Remailer, and then send the archive message once he
>       knows the Remailer has clocked the end of his turn and is
>       waiting to start the next turn.)

Unless the player thinks there might be an error with his attachment, he
can send the archive message immediately after the gamefile. The scripts
work pretty fast, so I doubt there would be a problem with the archive
message arriving at the same time the queue is being processed. There are
people who have slow e-mail and they may not want to wait until the end or
turn message is received.

[clip]

>I think this could work really well for 2-player head-to-head games.
>For 3 or 4 player games it does not work so well because the
>players would get the intermediate gamefiles.  (Then you could
>do much of the planning for your own move before asking the
>Remailer for the gamefile.  And then a lot would depend upon
>whether you were at war with the player moving just before you
>or just after you - in the one case you'd be at a disadvantage
>because you wouldn't have any time to think.)  One fix for that
>is to have the Remailer hold onto all the gamefiles for a particular
>player and then send him *all* the other players' gamefiles and
>archive messages *only* when his turn comes up again.  But
>that gets more complicated, and leads to the problem that players
>will not be able to catch rules errors and battle errors right
>away, so that a lot of games will need to be backed up a full turn
>(which leaves those players who just revealed their strategies
>feeling screwed) instead of just 1 or maybe 2 moves (which is
>usually no big deal).

After a bit of thinking I also believe that this method would work pretty
much only for 2 player games. In 4 player tournament games, making this
rule work would put more complexity into the rules and could potentially
make it harder to detect other abuses. To use the tournament games going on
now as an example, the first player (white) would have even more of an
advantage than he already has. Early in the game he is rarely in a war with
blue, who is on the opposite side of the board. He could complete his
entire turn in between moves and start and stop the clock within an hour,
thereby giving him a LOT of extra time later in in the game for the future
when he might be involved with blue. All he would need to do is duplicate
his moves (and reverts) with the new gamefile he received. Then, on turn 18
or so, he could take a week to do his turn due to the time he saved from
early turns. 2 player games would work because your move is completely
determined by the move of the other guy, you can't think ahead until you
see what he does.

[clip]

>That sounds like it might work, but we ought to discuss it
>more first, and bring in some more Moderators for their ideas.
>Also we have to see if Elam is sufficiently enthused about the
>idea to want to program it in as a Remailer gaming option. :)

>The other tricky situation I can imagine right now is if
>someone started the clock for their turn and checked out the
>gamefile from the Remailer, had a rules question, and then
>needed an answer while his clock was running.  Do you suspend
>the game in that situation?  (That would give him more time
>to consider his move.)  Maybe the thing to do there is to
>have the player take his best guess at the correct rules
>interpretation, send in his move, and then suspend the game
>once the clock stopped running.  If the player was right about
>the rule, the game would continue.  If he was wrong, the clock
>would be reset, he'd get a rules error, and he'd take his turn
>over.  If he was questioning another player's move; he'd
>still take his own, and then afterwards we'd back the game
>up and reset the clock and make them both move over again.

I could see another abuse. A player who is pressed for time sends in a
gamefile with an illegal move or action. That action would not be seen
until the next person was ready, at which time it has to go back to the
original player who by now has already had the time to figure out his move.
Even if this wasn't done on purpose, there is wasted time that could
technically be noticed immediately, especially with a major battle. This
types of things is what makes me think that these rules are best used for 2
player games.

>But what if the player starts the clock for his move and then
>gets called away or distracted by some crisis or emergency?
>Maybe that's just a tough break and it's the player's job to
>arrange a good time to do his move, within the limits of the
>turnaround-time, time limit, and vacation restrictions for the game.

Before I talk about other things, I'll say that there will have to be some
changes made to the Remailer. That are not major, but they will need to be
tested and I'd say the modifications would have to take a day or so,
especially since I want to make this a seperate option, so people can
choose either way to play. But I don't think it will work for the
tournament, any time savings would be slight, and I think what we'd find is
that those people who really took advantage of the new rules would be those
who naturally get their turn in quickly anyway. I like the 2 day limit
because it properly handles the time you need for a complicated turn, and
early on in the game, rarely do people use up their full 2 day limits.

>Speaking of vacations, Elam, some of the moderators have
>suggested that we do something like this with the vacation time:
>
>(1) Each player is given a number of vacation days in the game
>    settings, just like timeouts.
>
>(2) Games can only be suspended by Web.  The CGI script that
>    suspends the game gives you a radio button where you select
>    who to charge the suspension time to.  It can either go
>    as player vacation for one of the players, in which case
>    the Remailer keeps track and deducts vacation days,
>    or else it can be a moderator intervention which isn't
>    charged to anyone.
>
>(3) Moderators should be able to manipulate vacation days just
>    like timeouts.

Well, this is also doable, its just a matter of adding vactions the same
way timeouts exist. In fact, it would probably be quicker for me to simply
consider them timeouts with another name. It would be quick to set up, but
everything would have to be tested. We would have to discuss in more detail
things like what heppens if a player's vacation time ran out, do we turn
off e-mail suspension only for games with vacations tracked by the
Remailer, etc.

Overall, all the changes you are proposing here are possible, but do
require more programming and testing. It doesn't seem like any of these
proposals are urgent, so I could see doing them when I had some free time...

Elam

---------------------=#| http://www.pixgen.com/elam |#=---------------------
| Elam T. Birnbaum  | Programming Engineer | "An engineer is someone who   |
| elam@pixgen.com   | Macintosh Evangelist |  spends 3 hours trying to do  |
| New York City, NY | Expert LEGO Builder  |  a 2 hour problem in 1 hour." |
----------------------------------------------------------------------------
One thing I noticed while skimming over it is that it will work well for both 2-player and multiplayer games as long as the gamefile (and accumulated archive messages) only goes to one player each time the turn changes. The archive messages could then simply accumulate, with each player adding new things to the end and passing the result on to the next player, instead of going to all players all the time... - Bob
From: MTMGsMax 
Date: Wed, 28 Jan 1998 12:15:13 EST
To: Bob.Heeter
Cc: elam@pixgen.com, jpanagos@cellophane.com, mleung@uclink.berkeley.edu
Subject: Re:  "Speed Warlords"
Organization: AOL (http://www.aol.com)

Hey Gents:

Elam's objection based on "possible abuses" is well founded.  It doesn't take
much to imagine that happening.  

When I mentioned to Bob the idea of limiting the time (I think Elam and I even
touched on this once), it was so that everyone got the same allotment.  I
don't think it was intended to be "speed chess" as Warlords, but more to give
a basic guideline of time.  I think I thought around 100 hours for a 20 turn
game.  (5 hours a turn, obviously the early turns could build up credit for
the later ones.) 

In the Duel2 playtesting game (which was also 80 cities and a BIG map, and
other things) we had some turns that took forever.  I, personally, lost two
days of my life by spending a good 16 hours debating a turn!  (There was a
hero revert in there, too.)  That was an awful lot to spend on a game......  

To "time" the matches, you almost want any "rules" issues to be non-existent.
And, in a head to head game, you sort of get this.  One of my A-rounds has two
static Sides.  The other two guys are already beating up on each other,
besides the statics.  The game moves VERY quickly (they've done 4 turns in a
week?) and I haven't heard a peep out of them--- they monitor each other so
intently (we presume) that *they'll* find the error anyway!  (Esp. if both are
solid players.)  

The next problem, though, is that if it *really* is "speed Warlords" so that
moves must be done quickly, then you WILL get errors, even by good players.
I'm sure we all have seen enough instances already of "oops... sent the wrong
saved version" to know that this kind of thing just happens.  

And more so, if they feel that "every minute counts."  

My idea, for the Remailer controlled time limit, was to monitor the time.  Or
"allot it."  This way, you don't have those with "no lives" [ :/] spending 12
hours a turn when everyone else spends 3-4.  

It might be worth playtesting to see if it matters.  We would *like* to think
that good warlords are good, given a "reasonable" amount of time to do their
moves.  And bad warlords would still be bad, given all day to move.  I guess
the defining issue is whether some good warlords are playing worse as they
don't have enough "personal" time to combat those who do, or-- the corollary--
fair warlords are playing better because they DO have lots of time to rethink
moves and examine all possibilities.  

I think that, someone with some free time, *should* get a certain advantage.
(It's like a "skill" they have.) I don't think we should make everyone play in
50 hours if some people do have 100 hours..... [we could call the "Overworked
Warlords Tournament".  :-) ] 

But, it would be interesting to see-- when a game was over-- IF someone who
did the best in the game had spent 127hours when You  came in second
spending 77.....  I don't think there would be punishment for such, etc. just
cause for reflection and "what if...."?  However, were it reversed-- the top
finisher ALSO spent the least amount of time, then it would be readily agreed
that this person was the best Warlord in the game......  (if we still choose
to believe that time matters to player performance.)


One last issue, esp. with Head-2-Head games:  it depends on the opponent,
doesn't it?  For instance, I might want to get into a "leisurely" game with
Bob, where we decide "gentlemanly" to try to do the turn every 4 days, etc.
But, there may be others where playing in a SPEED game with them might seem an
advantage.  (GARY, for instance!!  He can't do anything quick....  :) )  

If we choose NOT to try to give this option to the 4-player games, then maybe
we can just do it by "agreement" in Head to Heads?  Say Mike and Jim want to
play each other in a timed game and agree to it?  But, maybe Bob and I don't
think we have the energy or the necessary blocks of time to do it regularly,
and thus don't?  


Given people's "time restrictions" already, and the overabundant use of
timeouts, the "choosing" for when the gamefile is sent-- "okay, I'm ready"--
is an aide to this problem.  Unfortunately for some, it would still come on
the THIRD day..... :-(

Thus, I don't think we should REQUIRE that everyone plays "timed" games unless
all players decide that they are ready and "up" to such feats.  Perhaps by
Round D we'll have enough "die hards" that they'd all love to try it?  That
might be a nice way to weed out a little more "wheat from the chaff" by that
point!  :*)

{I smile-- Like an idiot.  I'll have been on permanent Moderator duty by that
Point!]

m

Date: Wed, 28 Jan 1998 13:04:27 -0800
X-Sender: jpanagos@pacificrim.net
Mime-Version: 1.0
To: MTMGsMax 
From: jpanagos@cellophane.com (Jim  Panagos)
Subject: Re:  "Speed Warlords"
Cc: Bob.Heeter, elam@pixgen.com, mleung@uclink.berkeley.edu

Hi Guys
Re:  Max's comments below, I think you have to bear in mind that for the
ladder to work, you have to have a time limit (hopefully short), because of
the other players wishing to challange these players.  If you are only
required to play one head to head at a time, Mike and I would have ruled
out anyone challanging us for over a year (maybe a decade :-)).  Sooo, the
idea of a leisurely game, should be limited to pleasure warlording only,
not the Head to Head ladder.
later
Jim

At 12:15 PM 1/28/98, MTMGsMax wrote:
>Hey Gents:

>One last issue, esp. with Head-2-Head games:  it depends on the opponent,
>doesn't it?  For instance, I might want to get into a "leisurely" game with
>Bob, where we decide "gentlemanly" to try to do the turn every 4 days, etc.
>But, there may be others where playing in a SPEED game with them might seem an
>advantage.  (GARY, for instance!!  He can't do anything quick....  :) )
>
>If we choose NOT to try to give this option to the 4-player games, then maybe
>we can just do it by "agreement" in Head to Heads?  Say Mike and Jim want to
>play each other in a timed game and agree to it?  But, maybe Bob and I don't
>think we have the energy or the necessary blocks of time to do it regularly,
>and thus don't?
>
>
>Given people's "time restrictions" already, and the overabundant use of
>timeouts, the "choosing" for when the gamefile is sent-- "okay, I'm ready"--
>is an aide to this problem.  Unfortunately for some, it would still come on
>the THIRD day..... :-(
>
>Thus, I don't think we should REQUIRE that everyone plays "timed" games unless
>all players decide that they are ready and "up" to such feats.  Perhaps by
>Round D we'll have enough "die hards" that they'd all love to try it?  That
>might be a nice way to weed out a little more "wheat from the chaff" by that
>point!  :*)


JP
Cellophane Square's home page
http://www.cellophane.com

Cel-lo-phane Square (sel'-a-fain skwar) n.
1. the Northwest's best stores for used and new CDs, tapes and records.  2.
the best selection of independent labels, local artists and neat stuff you
can find.  3. the coolest record store in the western world.  Cel-lo-phile
n. Cel-lo-phaned v.


From: MTMGsMax 
Date: Thu, 29 Jan 1998 09:36:59 EST
To: Bob.Heeter
Subject: Re:  WCMod: Timed Warlords Idea
Organization: AOL (http://www.aol.com)

Bob and Mods:

Couple quick ideas for possibilities on the 4-player timed play.

1) The MODERATOR gets the files and messages before requested.  This way at
least *someone* has checked out the last player's move for errors before it is
sent to the next player.  (Labor intensive.)

2) The ARCHIVE is sent to all, but not the gamefile.  Allows those concerned
to check-- from their OLD gamefiles-- the reported battles to make sure they
go as stated.  This gives some hints to all-- equally-- about what someone has
done, but without being able to move their own armies and knowing where the
other players armies have ended, it's not that much of an aide.

3) Errors.  For timed play, what if there is no such thing as an error?  Well,
not like we have now.  Depending on the game, if it is a real "time crunch"
game where it seems possible (like speed chess) that you WILL run out of time,
then what is *most* crucial is the time itself.  Thus, I think that some form
of "time" should be the penalty in these games.  That, say, any delay caused
by sending a wrong file counts (half?) against YOUR time!  Thus, YOU would be
the *last* person to do such a thing intentionally, if one assumes you could
get the "right" file out in just a few hours!  

4) Or, maybe instaid of timeouts, we have a "reverse" conversion.  Everyone is
given 4 errors to start the game.  (I'm assuming a "crunch" game, like Bob's
60-hour example.)  Throughout the "rushing" in the game, everyone is allowed
to have transferred the gamefile wrong, or reported wrong, or whatever four
times.  (Something that causes a "suspension or delay" to the moving of the
game, either needing correction or not.)  Then, if you do NOT use your allowed
errors or don't think you'll need them, you can convert them for an extra 2
hours each (or something.....)  The risk being that, without "errors to give"
(like timeouts), the next error made by (attributed to) You would cause your
turn to be skipped.


Hopefully this will tweak the brain of someone's who works better than mine!
:*)  


[Elam?]

M

X-Sender: plang@mailhost.vwc.edu
Date: Thu, 29 Jan 1998 09:38:41 -0500
To: "Robert F. Heeter" 
From: Pat Lang 
Subject: Re: WCMod: Timed Warlords Idea
Mime-Version: 1.0


>
>What do people think?
>
>-- Bob
>
>I think the internet is too quirky for this to be very reliable. We could
try it, but I fear many problems with the internet in general, with
particular local connections, with AOL, etc.
This is all assuming the Remailer is working fine.

Pat


X-Sender: rheeter@mailhost
Mime-Version: 1.0
Date: Thu, 29 Jan 1998 10:40:29 -0500
To: rheeter
From: "Robert F. Heeter" 
Subject: "Speed Warlords"

Hi Elam -

Jim Panagos has an interesting idea for a new timing scheme for
the Head-to-Head ladder.  The basic idea is to have an overall
time limit for all of a player's moves, rather than (or in
addition to) individual time limits for each move.   I'm not sure
how it might work, but it's worth thinking about.

Max Fulton made a similar suggestion for the tournament as well,
but I'm not sure it would work so well in the tournament
context (see below), unless we could be a bit creative about
it and playtest it first, and there might not be time.

I've cc'ed Jim and Mike Leung from the h1-duel game, and Max.

Here's what Jim writes:
>
>I was wondering if rather than the current system we could go with a clock
>like approach, where each player is allotted so much time for the game and
>the remailer would track the hours used.  Then a player would be rewarded
>for quick turnaround by conserving his time and if he ran out of time he
>would lose.  This may not be doable, just an idea!  I do not see this as an
>option in Mike and my "unique" battle, although I do have some element of
>doubt as to whether this game will ever be resolved!
>
>I do think that part of making Warlords realistic is the fact that a
>"general" in the field does not have the option of waiting a month or more
>to move his armies and consider every conceivable option, he must react
>quickly.  In that respect H2H is very unrealistic.  That is why they clock
>chess games!  I know Mike likes to think  several moves down the line, and
>in order to be a worthy opponent, I must attempt to do the same!  It seems
>like each turn takes twice as long as the previous, sooo to extrapalate

I think this is a great idea, but we need to keep in mind that in
real wars and chess tournaments, the players are spending 100% of
their time on the game while the game is going on.  In PBEM Warlords,
we have the problem that players also have real-world lives to
deal with.

However, we could probably do something like this, which wouldn't
require making too many tweaks in the Remailer either (?):

(1) Say Player 1 completes Turn 5 and sends gamefile to the Remailer.
(2) The Remailer sends out the "change-of-turn" notice (as usual),
    but HOLDS ONTO THE GAMEFILE.  Remailer then holds all
    archive messages (but sends notice to any player that the
    archive message was received and held in the queue).
    (The idea here is that the game is "paused" until the next
     player is ready to think about his move, and no one gets
     any of the info needed to plan the move until he's ready.)
(3) Player 2 has some fixed amount of time to deal with real life
    and ignore the game (say the standard 2-day turnaround time).
(4) When Player 2 is ready to move, he sends a "start the clock"
    message to the Remailer.  The Remailer handles this by doing
    3 things:
       (a) sending the gamefile to all players who should receive it.
       (b) sending any queued archive messages to all players.
       (c) starting the timer clock for that player.
(5) Player 2 takes his turn as fast as he likes, and sends the
       gamefile back to the Remailer.  (Trick: he has to send the
       gamefile first, wait for the change-of-turn notice from
       the Remailer, and then send the archive message once he
       knows the Remailer has clocked the end of his turn and is
       waiting to start the next turn.)
(6) When the Remailer receives the gamefile, it stops that player's
       clock, and starts up the 2-day turnaround time for the
       next player.

I think this could work really well for 2-player head-to-head games.
For 3 or 4 player games it does not work so well because the
players would get the intermediate gamefiles.  (Then you could
do much of the planning for your own move before asking the
Remailer for the gamefile.  And then a lot would depend upon
whether you were at war with the player moving just before you
or just after you - in the one case you'd be at a disadvantage
because you wouldn't have any time to think.)  One fix for that
is to have the Remailer hold onto all the gamefiles for a particular
player and then send him *all* the other players' gamefiles and
archive messages *only* when his turn comes up again.  But
that gets more complicated, and leads to the problem that players
will not be able to catch rules errors and battle errors right
away, so that a lot of games will need to be backed up a full turn
(which leaves those players who just revealed their strategies
feeling screwed) instead of just 1 or maybe 2 moves (which is
usually no big deal).

In the event of a rules dispute or vacation, the normal suspend/revive
routine would work just fine.  And the current set of timeouts
could still work for the regular "pick up your move" turnaround time.

The overall clock time available per player would have to be
included in the turn-change notice, and it would have to be
a game setting that could be tweaked if desired.  Once a player's
clock ran out, they would not be allowed to move, and the Remailer
would let the other player continue moving until his clock ran out
too.  If neither player concedes, whoever holds more cities when
both clocks expire is the winner.

That sounds like it might work, but we ought to discuss it
more first, and bring in some more Moderators for their ideas.
Also we have to see if Elam is sufficiently enthused about the
idea to want to program it in as a Remailer gaming option. :)

The other tricky situation I can imagine right now is if
someone started the clock for their turn and checked out the
gamefile from the Remailer, had a rules question, and then
needed an answer while his clock was running.  Do you suspend
the game in that situation?  (That would give him more time
to consider his move.)  Maybe the thing to do there is to
have the player take his best guess at the correct rules
interpretation, send in his move, and then suspend the game
once the clock stopped running.  If the player was right about
the rule, the game would continue.  If he was wrong, the clock
would be reset, he'd get a rules error, and he'd take his turn
over.  If he was questioning another player's move; he'd
still take his own, and then afterwards we'd back the game
up and reset the clock and make them both move over again.

But what if the player starts the clock for his move and then
gets called away or distracted by some crisis or emergency?
Maybe that's just a tough break and it's the player's job to
arrange a good time to do his move, within the limits of the
turnaround-time, time limit, and vacation restrictions for the game.

Speaking of vacations, Elam, some of the moderators have
suggested that we do something like this with the vacation time:

(1) Each player is given a number of vacation days in the game
    settings, just like timeouts.

(2) Games can only be suspended by Web.  The CGI script that
    suspends the game gives you a radio button where you select
    who to charge the suspension time to.  It can either go
    as player vacation for one of the players, in which case
    the Remailer keeps track and deducts vacation days,
    or else it can be a moderator intervention which isn't
    charged to anyone.

(3) Moderators should be able to manipulate vacation days just
    like timeouts.

I don't know if that's do-able either, but it's another idea...

-- Bob

X-Sender: rheeter@mailhost
Mime-Version: 1.0
Date: Thu, 29 Jan 1998 11:47:41 -0500
To: mas-mgo@pop.lu.se, Bob.Heeter, elam@madness.net, geez@opendoor.com,
        Bob.Heeter, birwin@gte.net, bill.irwin@gsc.gte.com, geez@opendoor.com,
        mleung@uclink.berkeley.edu, jpanagos@pacificrim.net,
        kchang@u.washington.edu, Bob.Heeter, gsb@compupick.com,
        plang@mailhost.vwc.edu, Bob.Heeter, alex.vickers@utoronto.ca,
        andrey@ksu.edu, aubjf@mail.vtx.ch, bbrook@rna.bio.mq.edu.au,
        dembe739@student.liu.se, gsb@compupick.com, hartman@ucla.edu,
        ijb@nbnet.nb.ca, MrZakk@aol.com, MTMGsMax@aol.com,
        qbradley@gulf.csc.UVic.CA
From: "Robert F. Heeter" 
Subject: WCMod: Timed Warlords Idea

Hi Mods -

A few of us have been discussing a new rules idea
that was inspired by Max Fulton and Jim Panagos.
At the moment we think it will probably work only
for the Head-to-Head ladder where the games are 1-on-1,
but we'd like your opinions and ideas.

[ Elam, Max, and Jim, I've replied to your comments (sometimes
  directly and sometimes by adding new thoughts) in what follows. ]

The basic idea is to run the game somewhat like tournament
style chess games, where there is a timer and an overall
limit for the amount of time each player can spend making
his/her moves (in addition to the turnaround-time limit
for each move).  The allowed overall time limit can be
changed to allow games with a more relaxed pace as well
as "speed warlords" games where all moves must be made in haste.
The main value is that it puts busy players on an even footing
with those who have lots of free time.

Here's how it might work:

(1) Say Player 1 completes Turn 5 and sends gamefile to the Remailer.
(2) The Remailer sends out the "change-of-turn" notice (as usual),
    but HOLDS ONTO THE GAMEFILE.  Remailer then holds all
    archive messages (but sends notice to any player that the
    archive message was received and held in the queue).
    (The idea here is that the game is "paused" until the next
     player is ready to think about his move, and no one gets
     any of the info needed to plan the move until he's ready.)

Elam comments:
>This part is not actually that hard to do. I simply create a temporary
>archive for the game and any subsequent message gets appended to the
>archive. When it is time, the script takes each individual message in the
>archive and sends it out before deleting the temp file.

(3) Player 2 has some fixed amount of time to deal with real life
    and ignore the game (say the standard 2-day turnaround time).
(4) When Player 2 is ready to move, he sends a "start the clock"
    message to the Remailer.  The Remailer handles this by doing
    3 things:
       (a) sending the gamefile to all players who should receive it.
       (b) sending any queued archive messages to all players.
       (c) starting the timer clock for that player.

Elam comments:
>This isn't that hard to do either because this action is prompted by an
>e-mail, not by a specific time (though there would still have to be an
>application running in the backround to enforce the total turn limit).
>Adding this ability to the Remailer script wouldn't be too hard, though we
>would have to discuss what happens if someone uses up their time.

(5) Player 2 takes his turn as fast as he likes, and sends the
       gamefile back to the Remailer.
(6) The Remailer stops Player 2's timer, holds the gamefile and
    archive messages, and sends out the change-of-turn notice to
    the next player.  The next 2-day turnaround time then starts up.

If a player runs out their clock, then either they forfeit, or
perhaps the Remailer skips that turn and all their remaining turns
until their opponent(s) run out of time.  If no one surrenders,
then whoever holds the most cities when all the time runs out is
the winner.

A serious 20-turn game might allow 400 hours of contemplation.
(say 20 hours per turn, giving players a chance to really think
about their moves, sleep, eat, go to work in the meantime, etc.)
Players could take extra time on some turns without being slapped
with timeouts, while speeding through other moves and conserving
their clock time for when they really need it.

Meanwhile, busy players might go for speed, knowing that when
their turn comes up they'll have at most (on average) 3 hours
each night to make their move.  Then the game would have a 60-hour
limit.  Those who want real speed could go for a 20-hour limit
(1 hour per move!).  Going below 20 hours could be a problem
because of network sluggishness.  (If the Remailer doesn't get
your gamefile for 2 hours because you have AOL, that's hard to
live with.)

Another possible problem is that this system rewards players
with faster computers and more efficient operating systems.
Someone on an old 486 running DOS will have a bit more trouble
generating WarBOT reports and turn reports and running the game
swiftly than someone on a G3 PowerMac with 64 megs of RAM who
can run multiple WarBOTs at once and revert in 5 seconds or whatever.
(Wish I *had* a G3 PowerMac with 64 megs of RAM, but nevermind. :) )

Still, the idea is that the overall limit could be negotiated to
a level where both players were comfortable, and all players would
live or die by their choices...  For the Head-to-Head Ladder games
we would simply set guidelines on what was allowed.  Say the
upper limit would be a 7-day turnaround time, 21 days of
vacation/timeout time, and 600 hours of contemplation time.
This also keeps players with really important head-to-head
matches from taking forever to finish their moves.

I think this could work really well for 2-player head-to-head games.
For 3 or 4 player games it does not work so well because the
players would get the intermediate gamefiles.  (You could
do much of the planning for your own move before asking the
Remailer for the gamefile.  And then a lot would depend upon
whether you were at war with the player moving just before you
or just after you - in the one case you'd be at a disadvantage
because you wouldn't have any time to think.)  One fix for that
is to have the Remailer hold onto all the gamefiles for a particular
player and then send him *all* the other players' gamefiles and
archive messages *only* when his turn comes up again.  But
that gets more complicated, and leads to the problem that players
will not be able to catch rules errors and battle errors right
away, so that a lot of games will need to be backed up a full turn
(which leaves those players who just revealed their strategies
feeling screwed) instead of just 1 or maybe 2 moves (which is
usually no big deal).

Elam captures the idea better than I did:
>After a bit of thinking I also believe that this method would work pretty
>much only for 2 player games. In 4 player tournament games, making this
>rule work would put more complexity into the rules and could potentially
>make it harder to detect other abuses. To use the tournament games going on
>now as an example, the first player (white) would have even more of an
>advantage than he already has. Early in the game he is rarely in a war with
>blue, who is on the opposite side of the board. He could complete his
>entire turn in between moves and start and stop the clock within an hour,
>thereby giving him a LOT of extra time later in in the game for the future
>when he might be involved with blue. All he would need to do is duplicate
>his moves (and reverts) with the new gamefile he received. Then, on turn 18
>or so, he could take a week to do his turn due to the time he saved from
>early turns. 2 player games would work because your move is completely
>determined by the move of the other guy, you can't think ahead until you
>see what he does.

[ Point 1:  The Head-to-Head Ladder will be starting up fairly soon
   anyway (I'm aiming for an official launch during the middle of
   Round B in the Tournament), so it's great if it'll work there. ]
[ Point 2:  Can anyone think of a way we can get around these problems
   and get this working in a 4-player tournament game?  Maybe my
   objection about holding all the messages for a given player
   goes away when we get to the late rounds and players have lots
   of experience? ]

In the event of a rules dispute or vacation, the normal suspend/revive
routine would work just fine.  And the current set of timeouts
could still work for the regular "pick up your move" turnaround time.

The overall clock time available per player would have to be
included in the turn-change notice, and it would have to be
a game setting that could be tweaked if desired.  Once a player's
clock ran out, they would not be allowed to move, and the Remailer
would let the other player continue moving until his clock ran out
too.  If neither player concedes, whoever holds more cities when
both clocks expire is the winner.

The other tricky situation I can imagine right now is if
someone started the clock for their turn and checked out the
gamefile from the Remailer, had a rules question, and then
needed an answer while his clock was running.  Do you suspend
the game in that situation?  (That would give him more time
to consider his move.)  Maybe the thing to do there is to
have the player take his best guess at the correct rules
interpretation, send in his move, and then suspend the game
once the clock stopped running.  If the player was right about
the rule, the game would continue.  If he was wrong, the clock
would be reset, he'd get a rules error, and he'd take his turn
over.  If he was questioning another player's move; he'd
still take his own, and then afterwards we'd back the game
up and reset the clock and make them both move over again.

Elam adds:
>I could see another abuse. A player who is pressed for time sends in a
>gamefile with an illegal move or action. That action would not be seen
>until the next person was ready, at which time it has to go back to the
>original player who by now has already had the time to figure out his move.
>Even if this wasn't done on purpose, there is wasted time that could
>technically be noticed immediately, especially with a major battle. This
>types of things is what makes me think that these rules are best used for 2
>player games.

[ My answer to this is that the same problem already exists, and this
  is where the rules-violation points come in.  But it would still
  work best with experienced players in a Head-to-Head game. ]

But what if the player starts the clock for his move and then
gets called away or distracted by some crisis or emergency?
Maybe that's just a tough break and it's the player's job to
arrange a good time to do his move, within the limits of the
turnaround-time, time limit, and vacation restrictions for the game.
Or we could allow each player one "emergency" move per game,
where they charge themselves a rules error and lose a timeout
but they can take an extra dau to do their turn.

At this point it seems like there are lots of neat possibilities,
and the trick now is to brainstorm those possibilities until
we have all the avenues explored, and then to identify the
simplest possible set of principles by which these rules could
work, so that we can implement something that's reasonably easy
for Elam to program, reasonably easy for players to learn,
and which doesn't have too many bizarre side-effects and
ambiguous situations that can pop up and make life hell for
the moderators and rules gurus. :)

What do people think?

-- Bob

X-Sender: rheeter@mailhost
Mime-Version: 1.0
Date: Thu, 29 Jan 1998 14:55:50 -0500
To: mas-mgo@pop.lu.se, Bob.Heeter, elam@madness.net, geez@opendoor.com,
        Bob.Heeter, birwin@gte.net, bill.irwin@gsc.gte.com, geez@opendoor.com,
        mleung@uclink.berkeley.edu, jpanagos@pacificrim.net,
        kchang@u.washington.edu, Bob.Heeter, gsb@compupick.com,
        plang@mailhost.vwc.edu, Bob.Heeter, alex.vickers@utoronto.ca,
        andrey@ksu.edu, aubjf@mail.vtx.ch, bbrook@rna.bio.mq.edu.au,
        dembe739@student.liu.se, gsb@compupick.com, hartman@ucla.edu,
        ijb@nbnet.nb.ca, MrZakk@aol.com, MTMGsMax@aol.com,
        qbradley@gulf.csc.UVic.CA
From: "Robert F. Heeter" 
Subject: WCMod: Early Feedback on Timed Warlords

Date: Thu, 29 Jan 1998 09:38:41 -0500
From: Pat Lang 

I think the internet is too quirky for this to be very reliable. We could
try it, but I fear many problems with the internet in general, with
particular local connections, with AOL, etc.
This is all assuming the Remailer is working fine.

Pat


******************************************

From: MTMGsMax 
Date: Thu, 29 Jan 1998 09:36:59 EST

Bob and Mods:

Couple quick ideas for possibilities on the 4-player timed play.

1) The MODERATOR gets the files and messages before requested.  This way at
least *someone* has checked out the last player's move for errors before it is
sent to the next player.  (Labor intensive.)

2) The ARCHIVE is sent to all, but not the gamefile.  Allows those concerned
to check-- from their OLD gamefiles-- the reported battles to make sure they
go as stated.  This gives some hints to all-- equally-- about what someone has
done, but without being able to move their own armies and knowing where the
other players armies have ended, it's not that much of an aide.

3) Errors.  For timed play, what if there is no such thing as an error?  Well,
not like we have now.  Depending on the game, if it is a real "time crunch"
game where it seems possible (like speed chess) that you WILL run out of time,
then what is *most* crucial is the time itself.  Thus, I think that some form
of "time" should be the penalty in these games.  That, say, any delay caused
by sending a wrong file counts (half?) against YOUR time!  Thus, YOU would be
the *last* person to do such a thing intentionally, if one assumes you could
get the "right" file out in just a few hours!

4) Or, maybe instaid of timeouts, we have a "reverse" conversion.  Everyone is
given 4 errors to start the game.  (I'm assuming a "crunch" game, like Bob's
60-hour example.)  Throughout the "rushing" in the game, everyone is allowed
to have transferred the gamefile wrong, or reported wrong, or whatever four
times.  (Something that causes a "suspension or delay" to the moving of the
game, either needing correction or not.)  Then, if you do NOT use your allowed
errors or don't think you'll need them, you can convert them for an extra 2
hours each (or something.....)  The risk being that, without "errors to give"
(like timeouts), the next error made by (attributed to) You would cause your
turn to be skipped.


Hopefully this will tweak the brain of someone's who works better than mine!
:*)


[Elam?]

M


Date: Thu, 29 Jan 1998 16:37:26 -0800
X-Sender: jpanagos@pacificrim.net
Mime-Version: 1.0
To: "Elam T. Birnbaum" 
From: jpanagos@cellophane.com (Jim  Panagos)
Subject: Re:WCMod: Early Feedback on Timed Warlords
Cc: "Robert F. Heeter" , mas-mgo@hermes.lu.se,
        Elam Birnbaum , geez@opendoor.com, birwin@gte.net,
        bill.irwin@gsc.gte.com, mleung@uclink.berkeley.edu,
        jpanagos@pacificrim.net, kchang@u.washington.edu, gsb@compupick.com,
        plang@mailhost.vwc.edu, alex.vickers@utoronto.ca, andrey@ksu.edu,
        aubjf@mail.vtx.ch, bbrook@rna.bio.mq.edu.au, dembe739@student.liu.se,
        hartman@ucla.edu, ijb@nbnet.nb.ca, MrZakk@aol.com, MTMGsMax@aol.com,
        qbradley@gulf.csc.UVic.CA

I still think it is important to think about how the ladder will work
before going to far with game play options ie. form follows function.

This is why I think it is important to have relatively quick games.  If a
player only has to play one or two challange games at a time, we need to
consider the protocol of accepting and offering challanges.  People will
lose interest if they are denied the option of challanging a player for
several months (or years).
In order to have a flow, we need to accept that some players may go down in
rank by virtue of "being too busy", but by similar token when time allows,
may go up in rank!
One solution I think is very simple computes as follows.

20 move game
each player gets 48 hrs per move  = 40 days or 960hrs
Multiply this by the two players  = 80 days or 1920hrs

Thus you have a game that goes nearly 3mos.  Forget about timeouts and
vacations (if someone is gone for 10 days that means they must complete
their 20 turns in 30 days), send the file as soon as the remailer gets it
and deduct the time from their initial allotment of 960.

What I like about this, is that a player knows when a game will finish (80
days ,usually sooner), and when he is entitled to challange.  The scenarios
presented so far, seem to allow for games that will go to long and make the
ladder stagnant!

My 2 cents
later
Jim


X-Sender: rheeter@mailhost.jet.uk (Unverified)
Mime-Version: 1.0
Date: Mon, 16 Feb 1998 09:13:28 +0000
To: rheeter
From: "Robert F. Heeter" 
Subject: Timed Warlords only okay for H2H.

The timed warlords idea falls down when you realize that different
players have computers with different speeds, and different abilities
to write turn reports. Two native English speakers with fast computers
and big screesn plus the ability to simultaneously run email,
multiple copies of Warlords, and multiple WarBOTs, will be able to
do much more thinking in a given amount of time than someone in a
foreign country with DOS who has to do half the turn report by hand
while frantically reloading Warlords and WarBOT to switch from one
to the other.

The only way the timed idea can be fair is if both players compare
system setups and agree that it will be reasonable to use a timing
system during head-to-head play.  In that case I wonder if it would
be worth it for Elam to do the coding in the Remailer...

He could set the Remailer to send out a notice whenever the
"Resend" command is used, and then allow the gamefile not to be
sent anywhere when it is sent in (except to the Moderator).
Then a player sends the gamefile to the Remailer, the Remailer
holds onto it, and it only goes to the player when the resend
command is used.  The players themselves can keep track of
the time used by manually subtracting the time elapsed between
the "resend" from the Remailer and the sending of the next turn
by the player...

-- Bob


The World Warlords II League (WW2L):