Let's start a ladder tournament. As with a tennis
ladder, we start with a list of names in an arbitrary
order. You can challenge the person immediately above
you in the list. If you win, you exchange position
with them in the ranking. <br><br>I suppose we could
start a ladder for each variant of Octi, but perhaps
for starters we should create a ladder for the basic
3-square game. Games can, of course, be played on-line, by
email, in person, or whatever.<br><br>Let me know if
you'd like to participate (by replying to this
posting), and when we have a few names, I'll post the
ladder on the Yahoo! Octi club site, updating it as
needed.
The best prices I've seen lately for OCTI
(www.octi.net) are<br> <br>www.kbtoys.com
$14.95<br>BoulderGames.com $18.80<br>www.funagain.com $23.95<br> <br>Grab
the kbtoys special before it disappears.
Perhaps the computer should always check to see
whether it is possible to lose the game in, say, 5 or six
moves. If so, it needs to adopt a set of 'yikes'
heuristics -- either it's capable of winning faster than the
opponent, or it better shift gears and
defend.<br><EOM>
Ok, here is the deal... the puzzle image is at
<a href=http://host.bip.net/trax/octi/strat/octi0011.gif
target=new>http://host.bip.net/trax/octi/strat/octi0011.gif</a><br><br>And the
goal...<br>i) Red to play and win - assuming
octi for kids rules (easy)<br>ii) Blue to play and win
- assuming octi for kids rules (moderate)<br>iii)
Red to play and win - assuming capture of opponent's
pieces, no capture of own pieces (hard)<br><br>Telling
the number of moves to the goal would be making it a
little too easy (and to hard for myself), so give it
your best shot.<br><br>/cjr
I like your octi page(s) a lot. The only thing I
don't quite understand are the lines that emanate from
some of the pieces -- are these ways of showing how
the pieces could be jumped?<br><br>I still don't
quite get the puzzle. I generally like blue's chances,
but how many moves do you think remain until victory
is achieved?
For the puzzle, any player to play and win (i
think). I also worked a little on a new Octi page, to be
in my series of abstract games pages only including
Trax so far but intended to include all my favorite
games sometime when i can afford a domain.
<a href=http://host.bip.net/trax/octi
target=new>http://host.bip.net/trax/octi</a><br><br>/cjr<br><br>PS: Either cjr
or Carl (my real name) please (i don't
like Cjr since their my initials :) )
Great. We will have an on-line tourney as soon as
Eoin produces the new version of the the octi on-line
code -- I hope that's in the next two weeks!<br><br>In
the meantime, it might be nice to create some sort of
'challenge board' like the ones used in tennis ladders. Do
you want to organize it?<br><br>--Octimon
Nice graphics, Cjr. I need a bit more guidance as
to the octi-for-kids puzzle. Whose turn is it, and
how many turns until victory?<br><br>I think that the
comparison of the two square formations is very interesting,
thinking back to the proposed heuristic about maximizing
possible moves. These squares are good for very different
things, depending on how many enemy pieces they face and
in which direction. Can you give us some more
detailed examples? <br><br>-octimon
I thought i would instead collect my thoughts on
Octi on a web page, just as I did with trax... (hope
you don't mind, Don) starting with a small
illustrated comment on heuristics (i hardly begun yet, hope i
will have some time for it, it may improve in just 1
or 2 days in best case). So, tune in to
<a href=http://host.bip.net/trax/strategy.htm
target=new>http://host.bip.net/trax/strategy.htm</a><br><br>And do try my Octi
puzzle, at bottom of that page
Red/blue to play and win... (didnt check it all that
thouroughly, please report if you think is
mistake).<br><br>/cjr
11. Try to cause your opponent to run short of
prongs and pods. <br><br>12. If you have an advantage in
materiel, play for time, so that you can build.<br><br>13.
If your opponent is on your side of the board, try
to keep a pod in reserve, so that it can emerge,
disruptively, onto a home base.<br><br>14. It's an advantage to
defend behind one's home squares.<br><br>15. If you have
3 or fewer pods remaining and your opponent has 4
or more, you urgently need to secure an enemy
base.<br><br>16. An attack formation consisting of two pieces will
usually be completely disrupted after it attacks. An
attack force with three pieces is often much stronger.
4. Maximize the number of possible moves for
friendly pieces.<br><br>5. Minimize the number of possible
moves for enemy pieces.<br><br>6. Minimize the number
of enemy pieces. (: When choosing between enemy
pieces, choose the one with the greatest attacking
strength - attacking strength being defined (as a first
stab) as number of possible moves. This will
effectively mean, usually, the one with the most prongs, but
not always - this is a good way to get the computer
to decide that getting rid of one mid-pronged
well-placed piece is better than getting rid of a spider.
This seems like a tough call, though.<br><br>7. Build
your prongs 180 degrees from the places where your
opponent has no prongs, in directions from which you could
possibly approach.<br><br>8. If a friendly pod is in a
potentially dangerous situation, give it an escape prong.
(potentially dangerous is hard to define, though)<br><br>9.
Guard or capture enemy bases. Often (unless you need
captured pods back) it is better to have one pod guarding
2 enemy bases than to put a pod on top of each base
or one pod on top of one base.<br><br>10. Defend
your pods by keeping two in a row along lines of
probable enemy attack.
Post the heuristics that you think an AI version
of OCTI ought to consider.<br><br>For starters, here
are some of mine.<br><br>1. Try to keep friendly
pieces together in some kind of formation or mobile
stack.<br><br>2. Try to build or move so as to have prongs
pointing at your opponent, while few enemy prongs point at
you.<br><br>3. If possible, force an opponent to move by
building rather than moving.
Now there is another zillions-of-games version of
octi. Unlike the one on www.octi.net, this one plays an
offensive game. It's not bad (and more fun to play), but so
far it has yet to win against the zrf version on the
website, at 3 minutes per move.
Join the Yale OCTI club at noon at the Donaldson
Commons (corner of Prospect and Sachem) at noon on
Thursdays. <br><br>Also, come play OCTI on the Old
Campus/New Haven Green during Communiversity day, 4/15.
Well, my club, has grown 36-fold in two weeks
time, which means it may encompass many of those
ex-terristrials we don't know about yet with the same growth rate
next six months. (which would sure be interesting, how
unlikely it may sound)<br><br>And I would guess my part
would play the 2 player game.
Consider the rate of growth. Our OCTI club
membership has expanded by a factor of seven over two weeks'
time. If we grow by seven-fold every two weeks for six
months, we will encompass all of humanity. <br><br>But
will they play the 2 or 4 player version?<br>--octimon
When does the law of large numbers kick in?<br>10
doesn't seem very large -- huge standard errors.<br><br>I
might even find some unstandard errors to make (like
thinking we're playing a one square game).
Are you experiencing bugs or freeze-ups when logging on to octi on-line? If so,
let us know the computer and type of browser you're using. <br>--octimon
URL www.octi.net has zillions code and graphics for the kids version of OCTI.
Great for on-line play. It will soon be posted also at
www.zillions-of-games.com