probabilty, MLB "subway series", and coin toss
Tomoyuki Tanaka
today (Sunday) around noon on TV i heard George Will (?) say
that in the past the winner of the 1st game of the 7-game
World Series won the Series 59% of the time.
(did he say "69%" ?)
i thought ... is there such a thing as "momentum"?
or, in other words, is this (59%) "better than chance"?
so i took a few minutes to solve the following problem:
i'm tossing a coin 7 times. the first toss yielded
Heads. what is the probability that i'm going to get
more Heads than Tails?
and, is that better than 59%?
i'll post the answer in a few days if nobody else does.
>--------------------------------------------------------------------
> D.R.Hofstadter: "alien" "inscrutable" "Oriental mind"
>--------------------------------------------------------------------
>> in "Metamagical Themas" (1985) Hofstadter self-righteously
>> preached nonsexist language (word choices, etc) with
>> hypersensitivity.
>>
>> in "Le Ton beau de Marot" (1997) Hofstadter casually makes
>> fun of Asians with the phrases such as
>> "inscrutable" "the Oriental mind" and other outdated
>> (and inherently racist) stereotypes.
>>
>--------------------------------------------------------------------
> from Douglas Hofstadter's book "Le Ton beau de Marot" (1997)
>
> "Could it be that the very idea of transculturation
> itself is a Western one, and strikes the Oriental mind
> as alien?" (Page 148)
>
> "By virtue of being overly Oriental, it would be
> extraordinarily disorienting!" (Page 149)
>
> have you read another book that's published in the last 20
> years or so that uses the words like "inscrutable" and "the
> Oriental mind" (or other racist stereotypes) to make fun of
> Asians?
>
> if so, could you let me know?
> i'm esp. interested in books by non-comedians.
>
>>--------------------------------------------------------------------
>> NYT's review article of Douglas Hofstadter's book "Le
>> Ton beau de Marot"
>> http://www.nytimes.com/books/97/07/20/reviews/970720.20altert.html
>> (Prof. Alter, using lenient language, points out
>> Hofstadter's superficial understanding of literature
>> and translation.)
>>--------------------------------------------------------------------
Tomoyuki Tanaka
Dr J D Collins <jcol...@yahoo.com.au> wrote:
>
>1-Prob(a team wins 0, 1 or 2 games in 6 games)
>
>1-(6C0+6C1+6C2)/64
>
>1-(1+6+15)/64
>
>1-22/64
>
>0.656 or 65.6%
>
>so he probably said 69% since the better team is slightly
> more likely to win first.
>
>Regards
>John Collins
>
posting the following; email bounced.
--------------------------------------------------------------------
(personal email. not a copy of a Usenet post)
hi. thanks. that was quick.
actually, it took me more than a few minutes.
i used Pascal's triagle.
>
> so he probably said 69% since the better team is slightly
> more likely to win first.
>
> Regards
> John Collins
>
i agree, since one team is usually better than the other,
but...
i thought about it some more and i'm pretty sure he said 59%.
i'm hoping a baseball buff can confirm this.
tomoyuki
Dr J D Collins
1-(6C0+6C1+6C2)/64
1-(1+6+15)/64
1-22/64
0.656 or 65.6%
so he probably said 69% since the better team is slightly more likely to
win first.
Regards
John Collins
> today (Sunday) around noon on TV i heard George Will (?) say
phunt
tan...@web1.calweb.com (Tomoyuki Tanaka) wrote:
> ....
> i'll post the answer in a few days if nobody else does.
What is the probability of obtaining 3 or more successes
in six trials without replacement?
- - - - - - - - - -
>
> today (Sunday) around noon on TV i heard George Will (?) say
> that in the past the winner of the 1st game of the 7-game
> World Series won the Series 59% of the time.
>
> (did he say "69%" ?)
>
> i thought ... is there such a thing as "momentum"?
> or, in other words, is this (59%) "better than chance"?
>
> so i took a few minutes to solve the following problem:
> i'm tossing a coin 7 times. the first toss yielded
> Heads. what is the probability that i'm going to get
> more Heads than Tails?
>
> and, is that better than 59%?
>
> i'll post the answer in a few days if nobody else does.
Sent via Deja.com http://www.deja.com/
Before you buy.
Dale Hicks
> > that in the past the winner of the 1st game of the 7-game
> > World Series won the Series 59% of the time.
> > (did he say "69%" ?)
> > i thought ... is there such a thing as "momentum"?
> > or, in other words, is this (59%) "better than chance"?
> >
> > so i took a few minutes to solve the following problem:
> > i'm tossing a coin 7 times. the first toss yielded
> > Heads. what is the probability that i'm going to get
> > more Heads than Tails?
Dr J D Collins <jcol...@yahoo.com.au> wrote in article <G6QI5.3281$KY1....@news1.rivrw1.nsw.optushome.com.au>...
>
> 0.656 or 65.6%
gahh... a sci.math guy beat all the resident statheads (to my server, at least).
(or puzzles guy?)
> so he probably said 69% since the better team is slightly more likely to
> win first.
it was 50/87 counting the series from '03 to '90 that didn't have a tie to start
the series, so perhaps 59% is right. I can't find a web source for the
breakdowns.
finishing it up manually, we have
91 WWLLLWW
92 LWWWLW
93 WLWWLW
95 WWLWLW
96 LLWWWW
97 WLWLWLW
98 WWWW
99 WWWW
56/95 = 59%
--
Cranial Crusader dgh...@bellsouth.net
Octavio G. Pulsringwilson
>Date: 10/23/00 2:11 AM Eastern Daylight Time
>Message-id: <01c03cb8$b158a6c0$7ff84cd8@celeron>
The major problem with this calculation, is that it makes the assumption that
both teams are exactly even strength all the time. Which is not only false,
but can never, ever be true even in one case.
--------
"I gotta say something about that guy up there, and I can sum it all up in just
one word: courage, dedication, daring, pride, pluck, spirit, grit, mettle, and
G-U-T-S, guts!"
Dale Hicks
> [coin toss analogy to WS]
> The major problem with this calculation, is that it makes the assumption that
> both teams are exactly even strength all the time. Which is not only false,
> but can never, ever be true even in one case.
Yeah, but if it failed, it theoretically would fail the
other way, with the stronger team more likely to win game 1.
Of course, true sports fans know that it can depend on
pitching matchups, etc., etc. so it's not something you
can reduce to a coin toss.
--
Cranial Crusader dgh...@bellsouth.net
(who was surprised that the Braves have lead 1-0 in 3/5 series)
Tom MacIntyre
Tanaka) wrote:
>
> today (Sunday) around noon on TV i heard George Will (?) say
> that in the past the winner of the 1st game of the 7-game
> World Series won the Series 59% of the time.
>
> (did he say "69%" ?)
>
> i thought ... is there such a thing as "momentum"?
> or, in other words, is this (59%) "better than chance"?
>
> so i took a few minutes to solve the following problem:
> i'm tossing a coin 7 times. the first toss yielded
> Heads. what is the probability that i'm going to get
> more Heads than Tails?
Each toss is an individual event, and the probability each time is
50%.
Tom
>
> and, is that better than 59%?
>
> i'll post the answer in a few days if nobody else does.
>
>
>
>
Matthew T. Russotto
Tomoyuki Tanaka <tan...@web1.calweb.com> wrote:
}
} today (Sunday) around noon on TV i heard George Will (?) say
} that in the past the winner of the 1st game of the 7-game
} World Series won the Series 59% of the time.
}
} (did he say "69%" ?)
}
} i thought ... is there such a thing as "momentum"?
} or, in other words, is this (59%) "better than chance"?
No need to invoke "momentum" even if it is better than chance.
Remember that the teams are not necessarily evenly matched. It may be
that independently of all other games, one term has a >50% chance of
winning any particular game. If that's the case, then the winner of
the first game will be correlated with the winner of the Series.
--
Matthew T. Russotto russ...@pond.com
"Extremism in defense of liberty is no vice, and moderation in pursuit
of justice is no virtue."
David A Karr
>Remember that the teams are not necessarily evenly matched. It may be
>that independently of all other games, one term has a >50% chance of
>winning any particular game. If that's the case, then the winner of
>the first game will be correlated with the winner of the Series.
But it seems the actual observed percentage is 59% (unless the source
of that figure miscalculated), while the prediction assuming even
chances on each game is about 66%. Does that then imply an effect
that is the _opposite_ of "momentum"?
--
David A. Karr "Groups of guitars are on the way out, Mr. Epstein."
ka...@shore.net --Decca executive Dick Rowe, 1962
Christoph Gaard
lta...@my-deja.com
ka...@shore.net (David A Karr) wrote:
> >No need to invoke "momentum" even if it is better than chance.
> >Remember that the teams are not necessarily evenly matched. It may
be
> >that independently of all other games, one term has a >50% chance of
> >winning any particular game. If that's the case, then the winner of
> >the first game will be correlated with the winner of the Series.
>
> But it seems the actual observed percentage is 59% (unless the source
> of that figure miscalculated), while the prediction assuming even
> chances on each game is about 66%. Does that then imply an effect
> that is the _opposite_ of "momentum"?
>
> --
> David A. Karr
It's possible that the format 2 home - 3 away - 2 home has some effect
on the odds. This might be expected to dampen slightly the advantage
the home team gets when it wins the first game. Maybe some stat hounds
could run a few simulations with various reasonable parameters? (For
example making the slightly better team a 55-45 favorite at home, a
46-54 underdog away).
Of course it's also possible that the 66-59 discrepancy is not that
significant statistically, given the size of the sample.
Larry Tapper
Voros
I'd bet on the latter though I haven't checked.
--
Voros McCracken
vo...@daruma.co.jp
http://www.baseballstuff.com/mccracken/
Danil
> But it seems the actual observed percentage is 59% (unless the source
> of that figure miscalculated), while the prediction assuming even
> chances on each game is about 66%. Does that then imply an effect
> that is the _opposite_ of "momentum"?
Guess I had better throw this in before somebody else discovers it.
I spent Saturday dinner looking at game by game breakdowns of the world
series 1909 - 1999. Specifically, I was looking at how often the leader
in the series wins the next game.
1-0: 42 of 87
2-0: 19 of 42
2-1: 32 of 68
3-1: 18 of 35
Those all look like coin flips to me, but the net result is 10 games in
favor of "anti-momentum".
3-0: 16 of 19
That looks like something real, rather than a coin flip.
3-2: 20 of 53
A real effect in the opposite direction?
Please note that the numbers above haven't been reviewed, aside from a
consistency check.
Danil
Tomoyuki Tanaka
i'm glad some people here seem to know what they are talking
about (rare on Usenet).
i wish i knew more statistics so that i can evaluate the
significance of the sample size (56/95) and the discrepancy
("hypothetical" 66% vs. actual 59%).
maybe this would make a good article for Scientific American.
three possible reasons for the "anti-momentum" effect.
1. the winner of the 1st game becomes overconfident, careless.
2. the loser of the 1st game becomes energized
3. a fixer would design it so that the Series loser wins the
1st game. makes it more unexpected ... better for
his profits.
the famous fixed World Series of 1919 doesn't fit this pattern:
the White Sox threw the Series, but it also lost the 1st game.
were there other suspected fixed World Series games?
bkmaj
game series? Anybody know the percentage for winning the series after
winning the first game of a five game series? I'd guess significantly
higher.
maj
Christoph Gaard wrote in message
<20001023120620...@ng-fu1.aol.com>...
Gregory Bunimovich
: i'm glad some people here seem to know what they are talking
: about (rare on Usenet).
: i wish i knew more statistics so that i can evaluate the
: significance of the sample size (56/95) and the discrepancy
: ("hypothetical" 66% vs. actual 59%).
: maybe this would make a good article for Scientific American.
: three possible reasons for the "anti-momentum" effect.
: 1. the winner of the 1st game becomes overconfident, careless.
: 2. the loser of the 1st game becomes energized
: 3. a fixer would design it so that the Series loser wins the
: 1st game. makes it more unexpected ... better for
: his profits.
: the famous fixed World Series of 1919 doesn't fit this pattern:
: the White Sox threw the Series, but it also lost the 1st game.
: were there other suspected fixed World Series games?
Often the easiest is to simulate the distribution. I find that the probability
of having 56 or fewer teams out of 95 win the series after winning the 1st
game, assuming that the probability of winning each game is .5, is about
10.5%. Significant? That all depends on your definition (I'm a Bayesian,
anyway), but assuming standard measures of significance, no.
I don't particularly go for any of the explanations above, and I can't come
up with a good explanation of my own as to why the probability would be less
than 42/64 = .65625.
We could construct a scenario where the probabality would be less. Consider,
for example, a case of unequal teams. Team 1 is generally worse, and has only
a 40% chance of winning games 2-7. However, for some reason, it has a 60%
chance of winning game 1. Then the probability that the team that wins game
1 wins the series falls to .601728 (exact calculated probability, not
simulated).
We could have something of a similar situation due to home field advantage if
the worse team has home field advantage. It has home field in game 1, so it
has a >50% chance of winning, but (on average) in games 2-7 the home field is
equal, so in those games it has a <50% chance of winning. (on average -- this
isn't quite the right way to think about it, but I believe it gives the right
intuition).
In any case, I don't find such a scenario particularly realistic, just wanted
to show that you could play around with the numbers to make the probability
closer to 59%. I really don't believe that's necessary, though. I'm fairly
confident it's just chance.
Greg
WhoWins
[regarding Game 1 World Series winners winning the Series]:
>it was 50/87 counting the series from '03 to '90 that didn't have a tie to
start
>the series, so perhaps 59% is right. I can't find a web source for the
>breakdowns.
>
>56/95 = 59%
A number of points, most of which are trivial, and all of which are open to
correction:
1) Take Dale Hicks' presentation as correct for all World Series from 1903
through 1999
(inclusive), including the 1907 Series in which Game 1 ended in a 12-inning
tie. But
four of those 95 Series were of a best-of-nine format: 1903, 1919-1921. In
those four,
the Game 1 winner had a 2-2 Series record. Factor out those four, and the
best-of-7
Series record for the Game 1 victor becomes 54 Series wins in 91 Series ...
still 59%.
Thus, it's very likely that the Game 1 winner wins the Series 59% (not 69%)
of the time,
historically.
2) Dr. J.D. Collins' calculation of a 21/32 = .65625 best-of-7 series
victory probability
for the Game 1 winner assumes a per-game victory probability of .5, which is
a well-
founded assumption, given the coin-flip scenario presented him. But as
Octavio G.
Pulsringwilson, Dale Hicks, and others have noted, the assumption does not
extend
to games of sport that well -- it would be far better to use the individual
game probabilities
submitted regularly to r.s.bb by Douglas T. Massey, for example. If you
consider the
Pascal's triangle-based approach alluded to by Tomoyuki Tanaka, you get a
cumbersome
equation that states series victory probability as a function of individual
game probabilities.
For the Game 1 winner in a best-of-7 series, that equation features (Lord
help us) twenty
terms with coefficient +1, fifteen with coefficient -3, six with +6, and one
with -10. Make
the facile assumption that the Game 1 winner has an equal probability of
winning each of
the remaining games, and this forty-two-term monstrosity mercifully
collapses to
P(Game 1 winner wins series) = 20*p^3 - 45*p^4 +36*p^5 - 10*p^6, where p =
probability
of the Game 1 winner winning each of the remaining games. (Note how p = .5
leads to
P = .65625.) Solve that sixth-order polynomial for P = 54/91, and you get p
~ .467 - .468.
That value of p is so close to .5 that I would reject the notion that the
.65625 theoretical
series victory probability differs significantly from the 54/91 historical
series victory
probability.
3) Larry Tapper raised the issue of the 2-3-2 World Series format, and of
home-field
advantage. If you forgive us our considering the 1907 Series a four-game
sweep for the
Cubs (ignoring that first-game tie), you can find a comparison of best-of-7
series results
for Game 1 winners [for not only MLB but also the NBA and NHL (all rounds)]
at
http://www.whowins.com/tables/up10.html. It is often noted in r.s.bb that
home-field
advantage is far less important in baseball than in other sports. Indeed: In
best-of-7 World
Series play, when the home team wins Game 1, its Series record is 34-20
(63% -- not
much greater than 59-60%); by contrast, when the home team wins Game 1 in
the best-of-7
NBA Quarterfinals, its Quarterfinals series record is a lopsided 89-11
(.890). It amazes
me how extreme the home-court advantage is in the NBA.
4) Despite the overall 56-39 World Series record of Game 1 winners, the last
time that
the Game 1 winner lost the Series in even two consecutive years was
1985-1986. Since
then, Game 1 winners have been on a 10-2 Series run. Atlanta, in 1992 and
1996, is
responsible for the two Series losses in that time. The longest drought in
Series wins among
Game 1 winners in best-of-7 play was 1955-1959, inclusive.
Heath K.-W. Atomstramm
www.whowins.com
Octavio G. Pulsringwilson
>David A Karr says...
>> But it seems the actual observed percentage is 59% (unless the source
>> of that figure miscalculated), while the prediction assuming even
>> chances on each game is about 66%. Does that then imply an effect
>> that is the _opposite_ of "momentum"?
>
>Guess I had better throw this in before somebody else discovers it.
>
>I spent Saturday dinner looking at game by game breakdowns of the world
>series 1909 - 1999. Specifically, I was looking at how often the leader
>in the series wins the next game.
>
>1-0: 42 of 87
>2-0: 19 of 42
>2-1: 32 of 68
>3-1: 18 of 35
>
>Those all look like coin flips to me, but the net result is 10 games in
>favor of "anti-momentum".
>
>3-0: 16 of 19
>
>That looks like something real, rather than a coin flip.
It may be something real, but 16 of 19 doesn't sound like a very significant
sample size to me. Why do teams down 3-1 overcome desperation to win 50% of
the time, but teams down 3-0 don't? If there really is a "give-up" effect, it
should apply to both, to some degree, not apply entirely to one and completely
bypass the other, reducing it to a coin flip.
Paul G. Wenthold
I think I am objecting to everyone calling 59% the "probability"
that a team will win. After the Yanks won Saturday, does anyone
think there is really a 59% chance that they will win the series?
59% is how often the team that won the first game has won the
series, but I don't think it is the probability that the team
who wins game one _will_ win the series.
paul
Voros
> Danil wrote:
>>David A Karr says...
>>> But it seems the actual observed percentage is 59% (unless the source
>>> of that figure miscalculated), while the prediction assuming even
>>> chances on each game is about 66%. Does that then imply an effect
>>> that is the _opposite_ of "momentum"?
>>
>>Guess I had better throw this in before somebody else discovers it.
>>
>>I spent Saturday dinner looking at game by game breakdowns of the world
>>series 1909 - 1999. Specifically, I was looking at how often the leader
>>in the series wins the next game.
>>
>>1-0: 42 of 87
>>2-0: 19 of 42
>>2-1: 32 of 68
>>3-1: 18 of 35
>>
>>Those all look like coin flips to me, but the net result is 10 games in
>>favor of "anti-momentum".
>>
>>3-0: 16 of 19
>>
>>That looks like something real, rather than a coin flip.
> It may be something real, but 16 of 19 doesn't sound like a very significant
> sample size to me. Why do teams down 3-1 overcome desperation to win 50% of
> the time, but teams down 3-0 don't? If there really is a "give-up" effect, it
> should apply to both, to some degree, not apply entirely to one and completely
> bypass the other, reducing it to a coin flip.
I think the biggest difference in the 3-0 scenarios is that a team down
3-0 is generally a weaker team, though definitely not always.
John DiFool
Octavio G. Pulsringwilson wrote:
>
> >3-0: 16 of 19
> >
> >That looks like something real, rather than a coin flip.
>
> It may be something real, but 16 of 19 doesn't sound like a very significant
> sample size to me. Why do teams down 3-1 overcome desperation to win 50% of
> the time, but teams down 3-0 don't? If there really is a "give-up" effect, it
> should apply to both, to some degree, not apply entirely to one and completely
> bypass the other, reducing it to a coin flip.
I believe you will find a similar pattern in NHL & NBA best-of-seven
series (look at all the sweeps in both sports' finals this past decade...).
John DiFool
Douglas T. Massey
fonzi...@aol.comquat (Octavio G. Pulsringwilson) writes:
>>
>>I spent Saturday dinner looking at game by game breakdowns of the world
>>series 1909 - 1999. Specifically, I was looking at how often the leader
>>in the series wins the next game.
>>
>>1-0: 42 of 87
>>2-0: 19 of 42
>>2-1: 32 of 68
>>3-1: 18 of 35
>>
>>Those all look like coin flips to me, but the net result is 10 games in
>>favor of "anti-momentum".
>>
>>
>>That looks like something real, rather than a coin flip.
>
> It may be something real, but 16 of 19 doesn't sound like a very significant
> sample size to me. Why do teams down 3-1 overcome desperation to win 50% of
> the time, but teams down 3-0 don't? If there really is a "give-up" effect, it
> should apply to both, to some degree, not apply entirely to one and completely
> bypass the other, reducing it to a coin flip.
Well, I can think of two reasons; one statistical and one emotional.
The statistical: a 3-0 lead is more likely to indicate that the
leading team is actually much better than the trailing team, and thus
much more likely to win game four (or any game, for that matter).
If two unevenly matched teams reach the series and one has, let's
say, a 55% chance of beating the other, then it has a 16.6% chance
of going up 3-0, but only a 9.1% chance of being *down* 0-3. If
the better team wins 60% of the time, it's even more extreme: 21.6%
chance of being up 3-0, 6.4% chance of being down 0-3. So of all
the 3-0 leads that have been seen, you might expect two-thirds
of those leads to be taken by the better team, and thus the team
with the lead is likely to have the advantage in game four. But
that hardly explains a 16-3 record in game four; it might explain
a 10-9 or 11-8 record.
The emotional: no baseball team has ever recovered from an 0-3
deficit. This is mentioned ad nauseum every time there's an 0-3
deficit. It may seem an insurmountable obstacle to the trailing
team. 1-3 deficits have been overcome on numerous occassions, though,
so the "give up" mentality shouldn't be nearly as strong.
The records of the winning teams following a 3-1 lead (18-17) and
a 3-0 lead (16-3) might not be much different, when you consider
the difference in the "give up" factor, the fact that the leading
team is usually the better team (and a 3-0 lead indicates a bigger
difference than a 3-1 lead), and the magnitude of luck in a 19-
or 35-game sample.
If we guess that a 3-0 leading team is naturally going to win 52%
of game fours, but that is boosted to, say, 60% by the "give up"
factor, and that a 3-1 leading team is naturally going to win 51%
of game fives, but that's boosted to, say, 55% by the "give up"
factor, then a 16-3 (or better) record in 19 game-fours-with-a-3-0-
lead would occur 2.3% of the time, and an 18-17 (or worse) record
in 35 game-fives-with-a-3-1-lead would occur 39% of the time.
So to see both numbers seems like a 1-in-100 kind of thing to me:
strange, but not absurdly so. Maybe there are other factors that
really boost a 3-0 leading team's chances in game 4; I can't think
of any significant ones, though . . .
Maybe the trailing team gets desparate and puts in their game one
start on three days' rest, only to see him get rocked. Maybe the
trailing team has destroyed their bullpen in the first three games
and that hurts their chances. Anything else?
Doug
--
-----------------------------------------------------------------------
___, IBM Microelectronics Division, Burlington, Vermont
\o ASICs Product Development Engineering |>
| Phone: (802)769-7095 t/l: 446-7095 fax: x6752 |
/ \ E-mail: mas...@btv.ibm.com |
. Doug's Homepage: http://members.tripod.com/~masseyd (|)
Gregory Bunimovich
: I think I am objecting to everyone calling 59% the "probability"
: that a team will win. After the Yanks won Saturday, does anyone
: think there is really a 59% chance that they will win the series?
: 59% is how often the team that won the first game has won the
: series, but I don't think it is the probability that the team
: who wins game one _will_ win the series.
It's a empirical probability, based on historical data. It may differ from
the theoretical probability due to small (finite) sample size. It may differ
from 42/64 because that model is overly simplistic. But that's the question,
which I and others have attempted to answer. You simply state the probability
isn't 59%. That may be, but unlike others you've made no argument and shown no
proof.
Greg
Danil
> Danil wrote:
> >3-0: 16 of 19
> >
> >That looks like something real, rather than a coin flip.
>
> It may be something real, but 16 of 19 doesn't sound like a very significant
> sample size to me.
Perhaps - however, we are also talking about a very large deviation from
the average expectancy: 2.98 sigma. You'd get a distribution like this
maybe 3 or 4 times in 1000 trials of 19 coin flips each.
WhoWins
>Octavio G. Pulsringwilson wrote:
>> >3-0: 16 of 19
>> >That looks like something real, rather than a coin flip.
>> It may be something real, but 16 of 19 doesn't sound like a very
significant
>> sample size to me. Why do teams down 3-1 overcome desperation to win 50%
of
>> the time, but teams down 3-0 don't? If there really is a "give-up"
effect, it
>> should apply to both, to some degree, not apply entirely to one and
completely
>> bypass the other, reducing it to a coin flip.
>I believe you will find a similar pattern in NHL & NBA best-of-seven
>series (look at all the sweeps in both sports' finals this past decade...).
Disregarding the tie in the '07 World Series and treating '07 as a four-game
sweep, here
are the Game 4 records of the team leading 3-0 in MLB, NBA, and NHL
best-of-7 playoff
series:
World Series: 17-3 (.850)
World Series and League Championship Series: 20-5 (.800)
World Series, NBA Finals, and NHL Finals: 43-12 (.782)
all rounds -- MLB, NBA, and NHL: 142-77 (.648)
Further breakouts for teams leading 3-0 (also considering home-field
advantage) are listed at
http://www.whowins.com/tables/up30.html. Especially eye-catching is the Game
4 record of the
team leading 3-0 by best-of-7 playoff round:
Finals (MLB, NBA, NHL): 43-12 (.782)
Semis (MLB, NBA, NHL): 36-25 (.590)
Qtrs (NBA, NHL): 48-29 (.623)
Prelim (NHL): 15-11 (.577)
A bit of a "one more win and we've _finally_ won it all!" effect seems to be
characterizing
the teams leading 3-0 in the World Series, NBA Finals, and NHL Finals.
Heath K.-W. Atomstramm
www.whowins.com
Dale Hicks
> Dale Hicks wrote in message <01c03cb8$b158a6c0$7ff84cd8@celeron>...
> [regarding Game 1 World Series winners winning the Series]:
> start
> >the series, so perhaps 59% is right. I can't find a web source for the
> >breakdowns.
> >
> >56/95 = 59%
>
> 1) Take Dale Hicks' presentation as correct for all World Series from 1903
> through 1999
> (inclusive), including the 1907 Series in which Game 1 ended in a 12-inning
> tie.
You're right, I tried to omit '07, but I had the effect of
omitting '04, which I missed as a no-series year in the list.
Luckily (again), I was counting first game losses, and the
'07 series was counted (correctly) as a win.
--
Cranial Crusader dgh...@bellsouth.net
Ray Vickson
"Matthew T. Russotto" wrote:
> In article <39f3c497$1...@news3.calweb.com>,
> Tomoyuki Tanaka <tan...@web1.calweb.com> wrote:
> }
> } today (Sunday) around noon on TV i heard George Will (?) say
> } that in the past the winner of the 1st game of the 7-game
> } World Series won the Series 59% of the time.
> }
> } (did he say "69%" ?)
> }
> } i thought ... is there such a thing as "momentum"?
> } or, in other words, is this (59%) "better than chance"?
>
> No need to invoke "momentum" even if it is better than chance.
> Remember that the teams are not necessarily evenly matched. It may be
> that independently of all other games, one term has a >50% chance of
> winning any particular game.
Even if that's not the case, the first winner has an advantage: they don't
need to win as many of the remaining games as the first loser. Before the
first game is played, each team is equally likely to get that first-game
advantage (assume 50% winn probability), but after the first game the odds
change.
> If that's the case, then the winner of
> the first game will be correlated with the winner of the Series.
> --
> Matthew T. Russotto russ...@pond.com
> "Extremism in defense of liberty is no vice, and moderation in pursuit
> of justice is no virtue."
--
R. G. Vickson
Department of Management Sciences
University of Waterloo
Waterloo, Ontario, CANADA