Losing Streaks
EdmondT
I'd like to pose a question for people to discuss: What do you do to adjust
your games when you are in the middle of a losing streak? I think we all have
them, and when your in them it can be very frustrating. It often FEELS like
the dice are just against you.
My solution: I tighten WAY up on the cube. Try to get it, keep it until I can
close out matches with it, unless a VERY high liklihood of gammon.
Torben Østergreen-Johansen
In my humble opinion you shouldn't change the way you play at all just
because you're on a losing streak. Except, of course, if the losing streak
is a fair sign of your level of play. In that case you should study your
games and mistakes very carefully, to which effect I recommend the PC
programme JellyFish Analyzer 3.0 or even better Snowie pro version 1.0 (I
have both programmes, so I should know what I am talking about).
If we assume that your losing streak is caused by sheer bad luck, there is
nothing you can do about it, except taking a break and stop steaming. If you
try to play more conservatively (checker play or cube actions or both) your
game will only deteriorate even further since you are bound to lose equity
if you don't make the objectively correct play. Therefore, you will lose
equity both through bad luck AND bad play. You can't really prepare against
bad luck - if you "underplay" your position you will not get full value from
it, and who knows...next roll may be your breakout from the bad luck. And
wouldn't you be sorry you didn't double just because you feared your bad
luck, now only winning one point (by double and pass) instead of the four
points (by gammon on a 2-cube) you could have won if you had doubled
correctly?
I have a good friend, who happened to be on a very long losing streak
(several months). He was a fine player, though not an expert at that time,
but his results were indeed becoming increasingly poor. Frustrated, he tried
to change his play, but his results only got even worse. He was now
suffering from both bad luck and bad play. He took a break from backgammon
(a few weeks) and then started again with a fresh mind, returning to his
play as it was before Lady Luck got angry with him. In a few months he was
one of the very higest rated players in Denmark! I should mention, however,
that during his break from active backgammon he studied his own matches plus
expert matches very carefully with the help of JellyFish and reached the
conclusion, that his former play was not to blame for the poor results.
However, in the process he was able to catch a few misconceptions in his
games, thereby sharpening his play.
If you buy or already own Snowie, you can have your matches analyzed move by
move, and see if you lose because of bad play, bad luck or a combination of
both. The programme will tell you exactly how unlucky you were compared to
your opponent, and it also tell you how severe mistakes you and your
opponent made. If you want to read about Snowie, go to www.oasya.com. For
JellyFish, go to www.effect.no.
One last advice: Don't get angry with the dice and do NEVER scold you
opponent for your bad luck or his good luck. There is nothing more pathetic
than an opponent whining about his bad luck (often combined with bad play
or...dare I say it...caused by bad play). Good play results in more good
rolls on the next shake (disguised as "good luck") and bad play results in
fewer good rolls and more bad rolls (disguised as "bad luck"). Having said
that, I do of course acknowledge that even a perfect player (I know of none
such) will end up on a losing streak from time to time. But the streak will
be relatively short and he will be able to cut his losses to a minimum.
Best regards and good luck!
"torjoh" on FIBS
By the way, please excuse my english as it is not my native language. Danish
is, but you wouldn't like me to write in Danish, would you ;-)
EdmondT skrev i meddelelsen <19980920082845...@ng40.aol.com>...
John Goodwin
>
>I'd like to pose a question for people to discuss: What do you do to adjust
>your games when you are in the middle of a losing streak? I think we all have
>them, and when your in them it can be very frustrating. It often FEELS like
>the dice are just against you.
>
>My solution: I tighten WAY up on the cube. Try to get it, keep it until I can
>close out matches with it, unless a VERY high liklihood of gammon.
You should not make any modification to your playing style.
A losing streak is just like a streak of heads when tossing a coin
(unless you beleive that you are playing badly.)
Just because you've lost n games is no reason to suppose that you're
either more or less likely to win the next game (unless, of course,
you've become "spooked" by ypur streak).
JG
EdmondT
Two people have replied to my message about what people do to end losing
streaks. Both gave very sound advice about not changing the way you play just
because you are getting bad rolls.
I certainly agree with that. But I find that when I get on a losing streak,
its not that I play the same as when I was winning. Moreover, its very hard
for a non-professional to play the same way when they are getting bad dice for
a long period.
[I also have a pet peeve about BG players, which is that they generally ignore
their own good dice, and complain about luck when they get bad rolls. So I try
to avoid blaming luck when I'm losing.]
My experience with my own play is that I tend to get on losing streaks when I
get overly aggressive. Overblitzing, or leaving too many blots when I'm in the
middle of a blitz. Or doubling too early, or accepting too many doubles.
Thats why tightening up with the cube helps me when I'm on a streak.
Phill Skelton
>
> Two people have replied to my message about what people do to end
> losing streaks. Both gave very sound advice about not changing the
> way you play just because you are getting bad rolls.
>
> I certainly agree with that. But I find that when I get on a losing
> streak, its not that I play the same as when I was winning. Moreover,
> its very hard for a non-professional to play the same way when they
> are getting bad dice for a long period.
>
> streaks when I get overly aggressive. Overblitzing, or leaving too
> many blots when I'm in the middle of a blitz. Or doubling too early,
> or accepting too many doubles.
>
I think there are two kinds of losing streak. One happens when the
dice just don't go your way for a long time. Nothing you can do about
this except study your games to convince yourself that you are
following the correct strategy, to the best of yuor knowledge, as
previous posters have said.
The other sort, which I think is what you are on about, is when you
change the style of your game (consciously or not). I recently read
Robertie's 'Backgammon for serious players', and picked up on some of
the tips in there, particularly about slotting. I lost 19 of my next 25
games - not a huge losing streak, but significant. It takes a while to
get an intuitive feel for balancing up the risks and rewards for a
slightly different strategy, and during that time you are likely to
lose more than you win.
Phill
Murat Kalinyaprak
>I'd like to pose a question for people to discuss: What do
>you do to adjust your games when you are in the middle of a
>losing streak? I think we all have them, and when your in
>them it can be very frustrating. It often FEELS like the
>dice are just against you.
This is one thing I noticed no later than within
my first 4-5 days on the FIBS and wondered about.
When I was winning, it was going like 2 wins, 1
losses, 3 wins, 1 loss, 2 wins, 2 losses, etc.
But almost periodicly I would loose what seem to
be a considerable number of matches all in a row,
which I though was a little odd. Then I heard so
many other players mention their "loosing streaks"
while chatting, that I kind of thought maybe they
were intentional in order to keep a ceiling on the
ratings system or to avoid clustering of players
at the top and bottom with few in between. With my
"reputation":) in this newsgroup, I didn't want to
be the one who initiated a thread about it but I
was impatiently waiting for someone else to write
about it so I could get it out of my chest too...
Beside what to do about them, does anyone have any
ideas as to why do they even happen...? Since it
seems like nobody is immune to it, is it a common
human psychological phenomenon (i.e. selective
remembering) or just having off-days, etc...?
MK
Tapio Palmroth
>
> I'd like to pose a question for people to discuss: What do you do to adjust
> your games when you are in the middle of a losing streak? I think we all have
> them, and when your in them it can be very frustrating. It often FEELS like
> the dice are just against you.
>
> close out matches with it, unless a VERY high liklihood of gammon.
>
If i remember right , my losing streak has already been nearly
twenty years old :)
But what the heck , i am still playing :))
But.....maybe i have some winning streaks too , but one has some kind
of difficulties to remember those :)
That is one thing sure happens in real life too :)
But just keep on your game ( or whatever ) and the only thing
you have to tighten is your brains :))
It is just how life goes on :)
tapio
Us
won't have long to wait before you are winning again!
EdmondT wrote in message <19980920082845...@ng40.aol.com>...
>
>I'd like to pose a question for people to discuss: What do you do to
adjust
Albert Steg
> In <19980920082845...@ng40.aol.com> EdmondT wrote:
>
> >I'd like to pose a question for people to discuss: What do
> >losing streak?
Losing streaks are inevitable if you play enough, but the less expert you
are, the more frequent they will be. When I have a bad stretch, I hit the
books and try to reinforce my knowledge of the game, confirming to myself
that I'm playing as best as I know how.
If your game knowledge is solid, at least relative to the field you are
playing in, your aim should be to alter your playing *not at all*. Think
about it: you are playing the best way you know how, moving checkers and
wielding the cube in order to maximize your equity. Now that you've been
losing, though, you are thinking of playing differently? By making
sub-standard decisions in any way, you only increase your likelihood of
extending your losing streak.---well, at least you won't have the dice to
blame anymore!
Albert
Andrew Bokelman
I've been going through some bad luck. I almost laugh every time I hit my
opponent on my five point because for a week they ALWAYS hit me back when
they reenter. Also, for two days straight I lost EVERY match I played --
including a 7 and a 5 point match to people with ratings under 1400 (my
rating is over 1650).
So I start to doubt whether I'm really a 1650 plus player or I just got up
there by being lucky before. And I'd like to study more.
The only book I read (actually I studied it) is Magriel, and this was at
least 15 years ago. I'm rereading it now and I am going to order two books:
Advanced Backgammon Vol. 1 by Robertie, and How to Play Tournament
Backgammon by Woolsey -- unless someone here has strong feelings that these
are not good books to read next.
I'm also interested in either Jellyfish or Snowie. Can both of them replay
and advise on an entire match that I've played on FIBS.
..Andrew
No Spam Thanks
Andrew Bokelman <73457...@CompuServe.COM> wrote:
>Advanced Backgammon Vol. 1 by Robertie, and How to Play Tournament
>Backgammon by Woolsey -- unless someone here has strong feelings that these
>are not good books to read next.
_Tournament_ is a (relatively) short monograph that shows how
cube handling changes in match play vs. money play. It is
_the_ source for the match equity table. It also goes into
detail on how sometimes gammons matter; sometimes not.
Highly statistical, highly technical, lots of math, although
the math is the sort that uses _numbers_, so it's not all that bad. :-)
On the other hand, I think you should also consider _New Ideas..._
by Woolsey and Heinrich. It is a collection of 100-odd problems
with analysis by Woolsey and rollouts by JellyFish. All checker
play, no cube actions. I found the commentary useful not only
in and of itself, but on a meta-level because it showed so many
different ways to analyze positions, and it helped in developing
a technical vocabulary for thinking about a position.
(Although sometimes I wonder if the arguments were contrived
to support the rollouts, but the intro to the priming section
makes me think they're not...at least not always.)
As a 1650-plus player, your needs might be different from mine
(I've just clawed my way back to 1500 yay!), but I think a better
value than _Tournament_ would be _New Ideas_ by Woolsey and Heinrich.
It's twice the price, but I think more than twice the value.
See also the reviews on
http://www.gamesdomain.com/backgammon/books/
--
Craig E. Groeschel >Craig at Metro Link dot com< Not speaking for my employer.
"Do not play this piece fast. It is never right to play Ragtime fast." Joplin
h...@agis.net in...@agis.net sa...@agis.net docu...@agis.net
Chuck Bower
Reminds me of trying to choose between sugar and honey (and no, I
don't mean the decision between your wife and your mistress ;)
Seriously, although both "...Tournament BG" and "New Ideas..." are
excellent books, my feeling is that having not read (and absorbed)
Robertie's "Advanced BG" indicates that a student would be ahead
of him/herself to work through "New Ideas". I think Andrew's original
plan (buying "Advanced BG, volume I" and "Tournament BG") is a better
one. "New Ideas" could really throw a person into a tailspin because
it is really about "where experts screw up" or "exceptions to rules-
of-thumb". If you were to build your game around these problems
WITHOUT a solid basis in the fundamentals (as provided by Magriel and
Robertie), you might never recover. Sorta like picking up a calculus
book before you learn algebra....
"Tournament BG" is good for anyone who plays matches. However,
a solid understanding of money doubling strategy may be a prerequisite.
I would opt for buying BOTH of Robertie's volumes and ABSORBING them
first. (They cover a LOT of money cube decisions.) Then get "Tournament
BG" and finally, after getting about a good year's worth of experience
playing (and using the ideas in theses books), ask for "New Ideas" for
Christmas.
(Sorry, Kit. I may have just killed a couple royalty payments....)
Chuck
bo...@bigbang.astro.indiana.edu
c_ray on FIBS
EdmondT
I think Kit's "New Ideas" is more advanced than the two Robertie books: I'd
read them first, and once absorbed, go on to Kit's.
Torben Østergreen-Johansen
I will not comment on your losing streak, as I think my thoughts about this
is covered in my previous post.
However, I can inform you, that only Snowie is able to analyze your matches,
and it's very easy to import matches from FIBS and GamesGrid for these
analysis. In addition Snowie has a lot of other great features which are not
incorporated in JellyFish (though JellyFish is a very strong programme -
about equal to Snowie in playing strength).
Snowie does cost a few bucks more than JellyFish, but in my humble opinion
you do get extremely good value for these extra bucks. And don't wait too
long before deciding...Snowie's salesprice of 300 $ is an introductory
price, so expect to pay 400 $ or so if you wait much longer before ordering!
Best regards
Torben
"torjoh" on FIBS
Andrew Bokelman <73457...@CompuServe.COM> skrev i meddelelsen ...
>Torben,
>
>I've been going through some bad luck. I almost laugh every time I hit my
>opponent on my five point because for a week they ALWAYS hit me back when
>they reenter. Also, for two days straight I lost EVERY match I played --
>including a 7 and a 5 point match to people with ratings under 1400 (my
>rating is over 1650).
>
>So I start to doubt whether I'm really a 1650 plus player or I just got up
>there by being lucky before. And I'd like to study more.
>
>The only book I read (actually I studied it) is Magriel, and this was at
>least 15 years ago. I'm rereading it now and I am going to order two
books:
>Advanced Backgammon Vol. 1 by Robertie, and How to Play Tournament
>Backgammon by Woolsey -- unless someone here has strong feelings that these
>are not good books to read next.
>
Torben Østergreen-Johansen
Torben Østergreen-Johansen
Torben Østergreen-Johansen
Tom Vavasour
>you do get extremely good value for these extra bucks. And don't wait too
>long before deciding...Snowie's salesprice of 300 $ is an introductory
>price, so expect to pay 400 $ or so if you wait much longer before ordering!
Well I've read 4 copies of your post now, and I'm STILL not going to
buy Snowie....sorry. :)
...Tom (aka diGriz )
Money implies poverty
- Iain M Banks, The State of the Art
thehub
adjust
have
>them, and when your in them it can be very frustrating. It often FEELS
like
>the dice are just against you.
Losing streaks are natural and innevitable. Study and analysis will improve
your game if you are making inferior moves or cube decisions. However I
think it is just as important to be physically and mentaly sharp when you
play. That way you won't make as many bad decisions.
thehub
Murat Kalinyaprak
>Losing streaks are natural and innevitable. Study and
>analysis will improve your game if you are making
>inferior moves or cube decisions. However I think it
>is just as important to be physically and mentaly
>sharp when you play. That way you won't make as many
>bad decisions.
While I was keeping an eye on this thread, I had
one more of those loosing streaks and now really
starting to wonder why do they happen...?
Since I had last talked about it a week ago, my
rating had slowly gone up another 30 points or
so. At about that stage, I started thinking "a
loosing streak should be due any moment now"...
Sure enough, I lost 8-10 games in a row in less
than 2 days with 2 wins and 3 more losses in an
interleaved order following them.
After the first 6-7, I found myself playing the
next several games almost in a feeling of total
resignation that probably made things even worse.
But after busting your brains at no avail for a
number of matches, it becomes easy to fall into
a state of mind which submits to the idea that
one will loose the next "N" number of matches in
an almost predetermined manner no matter what
anyway. (on FIBS, that is:)
I have seen other readers comment that loosing
streaks are "normal", "expected", etc. and now
you are also saying "natural" and "inevitable".
But, why so...?
Mine perhaps wasn't even a "major" one compared
to what I have heard from other players. Do we
have "off-days" that last several days and weeks
when we are not as "sharp" (or all our opponents
get more sharp all of a sudden) and make worse
moves/cube decisions than "normal" (or all our
opponents make better moves/cube decisions all
of a sudden)? Assuming that there is some kind
of evenness/consistency out there in terms of
ratings+experience and that one doesn't change
his habits of choosing opponents based on them,
why such "noticeable" cases of unevennesses do
happen...?
Do people also have "winning streaks" that they
are not telling about? Personally, after I first
started paying attention to this "loosing streak"
phenomenon, I also paid attention to see if I get
"winning streaks" and can't say that I do. I'm
not sure whether I should base anything on my own
experience because I may have mismatching checker
play and cube play skills (which I go against its
wisdom sometimes) but then again I'm at the same
time probably quite consistent in doing what I'm
doing also. So why don't I win 8-9 matches in a
row every now and then, as well as loosing 8-9 in
a row...?
What about people who play bg on a regular basis
in real life...? Do they experience such "loosing
streaks" also...?
Does anybody have any "scientific knowledge":) or
"wild speculations":) regarding "loosing streaks",
beyond comments like "Oh, yeah, they are normal.
They happen to all of us"...?
MK
EdmondT
I think losing streaks can start in a couple of ways. First, you can just run
into bad dice for awhile. Second, after winning a lot, you can inadvertently
change your style of play, being too aggressive for example, or accepting too
many doubles on the theory that you will win somehow.
Once you are in a losing streak, it can continue for several reasons. First,
the dice can remain bad. More likely though, is that the player is affected by
the psychology of losing, and changes his style of play -- not necessarily
knowing that he is doing it.
For example, when I am losing I start feeling that any blot I leave anywhere
will be hit, and I often try not to leave any. This is obviously a mistake,
and can prolong the streak.
Murat Kalinyaprak
Before I wrote any more on this subject, I wanted to
and did take some more time to observe myself. I'll
make a few follow up comments which hopefully may be
useful to others.
>I think losing streaks can start in a couple of ways. First,
>you can just run into bad dice for awhile. Second, after
>winning a lot, you can inadvertently change your style of
>play, being too aggressive for example, or accepting too
>many doubles on the theory that you will win somehow.
I agree with both of the above. I still find it odd
that one can get enough "bad dice" to loose not just
so many games but so many matches in a row, but it
seems to happen. I suppose the reverse can happen
also. Initially, I had linked my loosing the first
5-6 matches to the dice. Since then, I observed that
I won 3 matches in a row twice. I guess one could
win 5-6 matches in a row just as easily as well.
>Once you are in a losing streak, it can continue for
>several reasons. First, the dice can remain bad. More
>likely though, is that the player is affected by the
>psychology of losing, and changes his style of play --
>not necessarily knowing that he is doing it.
>For example, when I am losing I start feeling that any
>blot I leave anywhere will be hit, and I often try not
>to leave any. This is obviously a mistake, and can
>prolong the streak.
I'm not sure if I modify my checker strategy during
the games but I agree with your comment that one can
and perhaps prolong the streak in such similar ways.
For example, I noticed/remembered that after the
above mentioned first 5-6 losses, I had resigned too
early (i.e. when technically I still had a however
small chance to win) in at least two occasions which
cost me the entire match. In one instance, I remember
apologizing to the other player for my emotional act,
thinking that it may come across as throwing "it" at
his face.
After the 8-9 matches which I had called a "streak",
wins and losses started to alternate but I was still
trying to do this. In one instance the other player
refused my trying to resign a regular game with cube
at 2. Sure enough, I rolled 3 doubles during bear off
and ended up winning the game. That's a whopping 4
points difference, 2 that I didn't loose plus 2 that I
won. Shortly after that, I found myself having clicked
on the resign button again in a similar situation. But
that time I noticed it soon enough to ask "could it
happen again?" and to click on the cancel button. And
sure enough, I won another game (and the match if I
remember correctly) that I had considered already
lost.
To make it short, at least in my case and at least
for the prolonged part of it, what I perceived as a
"loosing streak" was self-inflicted with observable
reasons like breaking some basic rules such as never
giving up until the very last chance. (BTW: my above
related behavior is clearly contradictory to my own
prior arguments that "it's not over until the lady
sings")...
One question of curiosity that still lingers in my
mind though, is whether robots experience "loosing
streaks"? Assuming some robot owners may have the
means of doing, do any robot owners keep statistics
which may provide a clue on this?
MK
Gary Wong
> I agree with both of the above. I still find it odd
> that one can get enough "bad dice" to loose not just
> so many games but so many matches in a row, but it
> seems to happen. I suppose the reverse can happen
> also. Initially, I had linked my loosing the first
> 5-6 matches to the dice. Since then, I observed that
> I won 3 matches in a row twice. I guess one could
> win 5-6 matches in a row just as easily as well.
Winning or losing 5 or 6 matches in a row should not be considered at
all uncommon. If you are playing with somebody of approximately your
own skill, then you can regard the outcome of a single point match as
a random event with two outcomes of probability roughly 50% (the same
model as a coin toss, for instance). Try tossing a coin 30 times. I
did it just now and got the following results (I hope nobody's going
to accuse my coin of cheating :-) --
THHHHHTTTHTHTHHTHTTTTTHTTTHHHH
There are two "winning" or "losing" streaks of 5 tosses ("matches") in
a row in that sample alone. I regard this sample as utterly
unremarkable; streaks of consecutive results are to be expected in any
sample of independent random variables. In fact, if streaks were NOT
found in a sufficiently large sample, that may be evidence of some
large scale structure (ie. non-randomness) in the data.
Separating the "signal" from the "noise" can sometimes be very
difficult to do. Preconceptions about randomness are very often
misleading. In fact, entire industries survive on the effects of
people expecting order in inherently disordered systems, and wanting
to predict random events from measurements of independent variables:
astrology, "how to beat the casino" scams, and, when I'm feeling
cynical, economics. Sports commentary probably falls in this
category, too -- I don't know much about baseball, but I've heard
commentators describe batters as being in a "slump" merely because
they haven't hit a ball in some short period of time. They are almost
certainly just measuring random noise. Surely nobody would suggest
that my coin was in a "tails slump" when I tossed 5 heads in a row,
earlier (implying that some property of the coin had changed) -- so
why do people suddenly attach importance to the existence of "batting
slumps" and "losing streaks" (and worse, assume that this will have
some effect on subsequent trials) when the same effect occurs in other
contexts?
(I don't mean to imply that every losing series of games can be
directly attributed to random noise. But a little probability theory
goes a long way in describing the behaviour of independent random
variables. It pays to understand how random variables behave WITHOUT
the existence of external influences, before invoking these external
influences as an explanation of what you see.)
> One question of curiosity that still lingers in my
> mind though, is whether robots experience "loosing
> streaks"?
If a losing streak is defined as a sequence of consecutive losses, then yes,
of course they do.
> Assuming some robot owners may have the
> means of doing, do any robot owners keep statistics
> which may provide a clue on this?
I keep records of each game Abbott plays (I originally wrote the thing to
carry out the experiment suggested by Don Banks at the end of last year,
but Peter Fankhauser beat me to it with yabs and yabe :-). Here is a
record of the last 1000 games it played ("w" = Abbott won, "l" = Abbott
lost):
lllllwwlllllwwwlwlwllllllllwllwlwlllwlwwwwlwllwwwl
llwwllllwlwlwllwwwwllwwlwwllwwwllllwwlwwwwwwllwllw
wlllwwlwwwwwlllllwwwwwlwwwlllwllllwlwwllwllllwllll
wwlwwlllwwlwlwlllwwwwlllllwlwlllwwllllwwwwllwlwlll
lwlllwlwwllwlwllwwwwllwllwllllwwlllwwwllwwwllwlwll
llwlwwlwwlwwllwwwlwlwlwwlwwwlwllwllwlwllwlwlwwwwwl
lwlwlllwllllwllwwllwwwlllwlwwwwwllwlwlwlwwlwlllwll
wwwwlllwlwwwwwlllwlllwlwlwlllwlwlllwwlwlwllwwlwlll
lwwwwwlllllllwwwwwwlwllllllwwwwwlwllllwwwwwlwwllll
llwwwwlwllwlwwwlllwlllwwlwllllllllwwwlwwlwllwlwwll
wllwwlwlwwwwllwwwlllwlwlwwwwwllwlwwlllwlllwlllllll
wwlwllwwwlwwllwwwlllwlllwlwlwwwllllllwwlwllllllwll
wwlwlllwllwwlllwwllwlwwwwllllwlwwwllwlwwwwlwwllllw
wwlwllwwlwwllwllwwlwlwwlwwlwwlllwllwllwwlwllwwwwwl
lllllllllwllwlllllllllwwwwllwwwlwwwlwwlwwwlwlllllw
llwwwwlllwlwwwwllwlllwlllwllwllllwwwwllwwwllwwwlll
wlwwwwwwwwwllwllllwlwllwllwwlwllwwllwwlwwlwwwlwwww
wwlllwwwlllwwlwwllwwlwllwllwlwwwwlllllwwwlllwllwwl
llwllwwwwlwwlwllwlwlwlllwlwllwwwlllwlllwllwwwlllww
lwlwwwllwwlllwlwlwwwllwwlllllllwllwwllwlwlwwwllllw
The ratings and experience of its opponents are a bit long to include here,
but I saved a summary of the same sample of 1000 games (23KB) at:
ftp://ftp.cs.arizona.edu/people/gary/abbott.log
Each line is in this format:
789 1234.56 w fred
indicating that Abbott won ("w") one game against fred, whose rating after
the game was 1234.56 and experience was 789.
Cheers,
Gary.
--
Gary Wong, Department of Computer Science, University of Arizona
ga...@cs.arizona.edu http://www.cs.arizona.edu/~gary/
charles platter
>I'd like to pose a question for people to discuss: What do you do to adjust
>your games when you are in the middle of a losing streak? I think we all have
>them, and when your in them it can be very frustrating. It often FEELS like
>the dice are just against you.
>My solution: I tighten WAY up on the cube. Try to get it, keep it until I can
>close out matches with it, unless a VERY high liklihood of gammon.
While not simply a function of probability, streaks are definitely an
attribute of randomness, which is consistent with the fact that the
random arrangement of the stars looks like clumping instead of evenly
spaced individual objects. Steven Jay Gould has an amusing essay on
probabliity in "Bully for Brontosaurus" His subject is Joe Dimagio's
hitting streak (cutting to the chase, 56 games is so unlikely it's not
worth calculating).
Murat Kalinyaprak
> mu...@cyberport.net Murat Kalinyaprak writes:
>> One question of curiosity that still lingers in my
>> mind though, is whether robots experience "loosing
>> streaks"?
>If a losing streak is defined as a sequence of
>consecutive losses, then yes, of course they do.
>> Assuming some robot owners may have the
>> means of doing, do any robot owners keep statistics
>> which may provide a clue on this?
>I keep records of each game Abbott plays (I originally
>wrote the thing to carry out the experiment suggested
>by Don Banks at the end of last year, but Peter
>Fankhauser beat me to it with yabs and yabe :-). Here
>is a record of the last 1000 games it played ("w" =
>Abbott won, "l" = Abbott lost):
I'm happily surprised that you had exactly the data
which could provide some answers to an unusual(?)
question. Since "Abbot" played against opponents of
ratings varying from very high to very low, what can
be derived from the below stats may be questionable;
but let's take a quick look at it anyway.
>lllllwwlllllwwwlwlwllllllllwllwlwlllwlwwwwlwllwwwl
>llwwllllwlwlwllwwwwllwwlwwllwwwllllwwlwwwwwwllwllw
>wlllwwlwwwwwlllllwwwwwlwwwlllwllllwlwwllwllllwllll
>wwlwwlllwwlwlwlllwwwwlllllwlwlllwwllllwwwwllwlwlll
>lwlllwlwwllwlwllwwwwllwllwllllwwlllwwwllwwwllwlwll
>llwlwwlwwlwwllwwwlwlwlwwlwwwlwllwllwlwllwlwlwwwwwl
>lwlwlllwllllwllwwllwwwlllwlwwwwwllwlwlwlwwlwlllwll
>wwwwlllwlwwwwwlllwlllwlwlwlllwlwlllwwlwlwllwwlwlll
>lwwwwwlllllllwwwwwwlwllllllwwwwwlwllllwwwwwlwwllll
>llwwwwlwllwlwwwlllwlllwwlwllllllllwwwlwwlwllwlwwll
>wllwwlwlwwwwllwwwlllwlwlwwwwwllwlwwlllwlllwlllllll
>wwlwllwwwlwwllwwwlllwlllwlwlwwwllllllwwlwllllllwll
>wwlwlllwllwwlllwwllwlwwwwllllwlwwwllwlwwwwlwwllllw
>wwlwllwwlwwllwllwwlwlwwlwwlwwlllwllwllwwlwllwwwwwl
>lllllllllwllwlllllllllwwwwllwwwlwwwlwwlwwwlwlllllw
>llwwwwlllwlwwwwllwlllwlllwllwllllwwwwllwwwllwwwlll
>wlwwwwwwwwwllwllllwlwllwllwwlwllwwllwwlwwlwwwlwwww
>wwlllwwwlllwwlwwllwwlwllwllwlwwwwlllllwwwlllwllwwl
>llwllwwwwlwwlwllwlwlwlllwlwllwwwlllwlllwllwwwlllww
>lwlwwwllwwlllwlwlwwwllwwlllllllwllwwllwlwlwwwllllw
>The ratings and experience of its opponents are a bit
>long to include here, but I saved a summary of the
>same sample of 1000 games (23KB) at:
> ftp://ftp.cs.arizona.edu/people/gary/abbott.log
Based on the above and what's at your web page, I
really can't say I see "loosing streaks". In my own
case, I was talking about at least 8-9 matches but
wasn't willing to blame any more than 5-6 of them
on the dice. I have the impression that many other
people mean larger numbers of consecutive losses
when they talk about "loosing streaks". Even just
5-6 matches of 5 points can easily equal 30-40
games. Of course one doesn't loose all matches in
a "loosing streak" always "N-0" (in terms of games)
but at least would most likely loose a big majority
of them.
In the above sequences, a few sections come close
to equalling a few matches lost in a row, but they
are against players rated at 300-400 points above
"Abbott", in which situations "Abbott" would be
expected to loose anyway. So, personally, not only
I don't see that "Abbott" had "loosing streaks" but
to the contrary it seemes to have done better than
expected against even much higher rated opponents.
The reason I had wondered about these things was
because I don't quite understand how can a rating
system be maintained with a game based on dice and
when players are free to choose their opponents
(i.e. they are not restricted to opponents in their
rating brackets).
Let's consider my case. Not only I'm supposed to be
quite weak with the cube, but I also play in a way
as though I'm conducting an experiment to see to
what degree I can neuralize the cube with checker
play, etc. To that end, (and at the expense of maybe
offending some lower rated players by not accepting
their invitations), for the past many weeks I almost
exclusively played against opponents rated 1650-1900+.
And my rating still remains at around high 1600's,
occasionally even going over 1700.
If I can be allowed to deduct from this that my
checker play is perhaps way stronger than what my
rating reflects, let me continue.
When I first joined FIBS, I had stuck to 1 pointers
almost exclusively for at least the first 100 or so
games. I was also somewhat cautious to play against
similar rated players. For the sake of simplicity,
let's just say that I could win 2 out of 3 games
against players in the 1500's. By an oversimplified
ratings calculation, that would give me about 350
points in those 100 games. Yet, at the end of those
first 100 games, my rating hadn't gone up to 1850.
It hadn't gone up to 1750 or even 1700 either. Why?
My own answer to that was that I was having periodic
"loosing streaks" that would pull me down 30-40-50
points at a time.
At 100-200, 200-300, 300-400 experience brackets, I
could still earn 4x, 3x and 2x points but I didn't
make it to up there during those stages either. It
almost felt like your climb up was somehow ensured
to be a very slow crawl (although I later read about
how some players go around this by certain tricks).
I asked myself the same questions about robots like
Jellyfish, Snowie, MonteCarlo, etc... Supposedly
only a few people on this planet can beat them on
a regular basis. I'm sure those few people aren't
on-line 24-hours a day playing against them and
keeping their ratings down. So, what prevents them
or some top human players from achieving ratings
like 3700, 4200, 5995, etc...?
Frankly, I had thought of a possibility of some
sort of a "bonzai trimming" in order to "fake" a
stable looking rating system with some resonable
range like 1200's-2100's... I may attract more
criticism for making speculations again but I
just can't keep from wondering... :) Let me also
emphasize the "wondering" part just to prevent
possible requests for "strong evidence"...
MK