SPSSX Discussion

Tennis / likelyhood of winning a second time

Classic

List

Threaded

12 messages Options

drfg2008

Tennis / likelyhood of winning a second time

This post was updated on .

The question is whether a tennis player who won the first set, is more likely to win the second set.

For this I have two variables: Set 1 and Set 2 for the results of the sets

(SET1_win / SET2_win 1 / 1= Player 1 wins, 2 = Player 2 wins).

In addition, I have the game results in set 1 and set 2

ScoreSet1T1 for the number of points of player 1 in the first set
ScoreSet1T2 for the number of points of Player 2 in the first set
ScoreSet2T1 for the number of points of player 1 in the second set
ScoreSet2T2 for the number of points of player 2 in the second set

The record looks like this:

ScoreSet1T1
ScoreSet1T2
ScoreSet2T1
ScoreSet2T2
Set1_win
Set2_win

I could also provide some data.

Now I'm not quite clear which method is best. Any help appreciated.

Thanks

Dr. Frank Gaeth

lori.andersen

Re: Tennis

I would do a chi-square. You have four possible outcomes for a 2 set match: win no sets, win set 1 only, win both sets, win set 2 only. A chi-square will answer the question if these four outcomes occur with equal likelihood or not. You would be most interested in the "win both sets" outcome.

On Mon, May 14, 2012 at 6:03 AM, drfg2008 [via SPSSX Discussion] <[hidden email]> wrote:

The question is whether a tennis player who won the first set, is more likely to win the second set.

For this I have two variables: Set 1 and Set 2 for the results of the sets

(SET1_win / SET2_win 1 / 1= Player 1 wins, 2 = Player 2 wins).

In addition, I have the game results in set 1 and set 2

ScoreSet1T1 for the number of points of player 1 in the first set
ScoreSet1T2 for the number of points of Player 2 in the first set
ScoreSet2T1 for the number of points of player 1 in the second set
ScoreSet2T2 for the number of points of player 2 in the second set

The record looks like this:

ScoreSet1T1
ScoreSet1T2
ScoreSet2T1
ScoreSet2T2
Set1_win
Set2_win

I could also provide some data.

Now I'm not quite clear which method is best. Any help appreciated.

Thanks

Dr. Frank Gaeth
FU-Berlin

If you reply to this email, your message will be added to the discussion below:
http://spssx-discussion.1045642.n5.nabble.com/Tennis-tp5708820.html

To start a new topic under SPSSX Discussion, email [hidden email]
To unsubscribe from SPSSX Discussion, click here.
NAML

--
Lori Andersen
Ph.D. student, Educational Policy, Planning & Leadership
College of William & Mary
Williamsburg, VA

drfg2008

Re: Tennis

This post was updated on .

As far as I understand a Chi-square test is testing independence between the two sets, which is definitely not the case since the winner of set 1 is stronger and will be more likely to win set 2, too. And this is not my question. The question is if the likelyhood increases to win again when the player has won before.

I thougth about the McNemar test first since this tests if a change from winning to loosing is equal likely as the change from loosing to winning. If the hypothesis (see above) is true (namely that a winner is more likely to win again) there should be more changes from one side to the other.

But this can not be the solution, since the winner is the non-looser and the looser is the non-winner in equal proportions.

Dr. Frank Gaeth

Marta Garcia-Granero

Re: Tennis

Hi:

I think you have paired data, since in the first match, one member of
the pair wins, and, therefore, the other one loses. You have to take
into account the pairing and treat the data as a matched cohort study.
McNemar test will do that, but you get only a p-value, not an estimate
of the increase of likelihood of victory. You need a relative risk that
takes into account the matching. Look for this paper: The Stata Journal
(2004) 4, Number 3, pp. 274–281, Analysis of matched cohort data, by
Peter Cummings Barbara McKnight.

I have also read that log-binomial or Poisson regression can be used for
that.

I hope this helps,
Marta GG

El 14/05/2012 14:06, drfg2008 escribió:

> As far as I understand a Chi-square test is testing independence between the
> two sets, which is definitely not the case since the winner of set 1 is
> stronger and will be more likely to win set 2, too. And this is not my
> question. The question is if the likelyhood increases to win again when the
> player has won before.
>
> I thougth about the McNemar test first since this tests if a change from
> winning to loosing is equal likely as the change from loosing to winning. If
> the hypothesis (see above) is true (namely that a winner is more likely to
> win again) there should be more changes from one side to the other.
>
>
>
> -----
> Dr. Frank Gaeth
> FU-Berlin
>
> --
> View this message in context: http://spssx-discussion.1045642.n5.nabble.com/Tennis-tp5708820p5708910.html
> Sent from the SPSSX Discussion mailing list archive at Nabble.com.
>
> =====================
> To manage your subscription to SPSSX-L, send a message to
> [hidden email] (not to SPSSX-L), with no body text except the
> command. To leave the list, send the command
> SIGNOFF SPSSX-L
> For a list of commands to manage subscriptions, send the command
> INFO REFCARD
>

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD

David Marso

Re: Tennis / likelyhood of winning a second time

Administrator

In reply to this post by drfg2008

As is typical, your question is as clear as MUD!
Please carefully reformulate this with the addition of whatever assumptions you are making about the data and try again. First step back and reread your question to a hopefully forgiving coworker and see if you get anything other than a blank stare. You don't do yourself any favors by repeatedly posting nebulous carelessly formulated questions to this list which require 4-5 iterations to gel. You seem to think this group is here as a free consulting service and that you don't need to use your own thinking about your research! Last week it was Futball, this week Tennis, a month or so ago some nebulous 5000 variable GLM and Poisson regression prior to that and Python in the mix as well. This wouldn't be so much a problem if it weren't for the fact that you *NEVER* bother to pitch in and help anyone else.
Look up Reciprocity in your German/English dictionary.
---

drfg2008 wrote

The question is whether a tennis player who won the first set, is more likely to win the second set.

For this I have two variables: Set 1 and Set 2 for the results of the sets

(SET1_win / SET2_win 1 / 1= Player 1 wins, 2 = Player 2 wins).

In addition, I have the game results in set 1 and set 2

ScoreSet1T1 for the number of points of player 1 in the first set
ScoreSet1T2 for the number of points of Player 2 in the first set
ScoreSet2T1 for the number of points of player 1 in the second set
ScoreSet2T2 for the number of points of player 2 in the second set

The record looks like this:

ScoreSet1T1
ScoreSet1T2
ScoreSet2T1
ScoreSet2T2
Set1_win
Set2_win

I could also provide some data.

Now I'm not quite clear which method is best. Any help appreciated.

Thanks

Please reply to the list and not to my personal email.
Those desiring my consulting or training services please feel free to email me.
---
"Nolite dare sanctum canibus neque mittatis margaritas vestras ante porcos ne forte conculcent eas pedibus suis."
Cum es damnatorum possederunt porcos iens ut salire off sanguinum cliff in abyssum?"

drfg2008

Re: Tennis / likelyhood of winning a second time

This post was updated on .

@David

This wouldn't be so much a problem if it weren't for the fact that you *NEVER* bother to pitch in and help anyone else. Look up Reciprocity in your German/English dictionary.

Look at that: http://www.statistik-tutorial.de/forum/index.php

I posted over 1382 answers here for people asking questions.

Look at that: youtube. Here I posted more than 60 films / introductions into SPSS (more than 100 thousand downloads)

All for free, David.

Dr. Frank Gaeth

David Marso

Re: Tennis / likelyhood of winning a second time

Administrator

Congratulations on your contributions to German intellectual culture.
I was referring to the SPSS-X list (here, this place...).
You still need to clarify your question.

drfg2008 wrote

@David

This wouldn't be so much a problem if it weren't for the fact that you *NEVER* bother to pitch in and help anyone else. Look up Reciprocity in your German/English dictionary.

Look at that: http://www.statistik-tutorial.de/forum/index.php

I posted over 1382 answers here for people asking questions.

Look at that: youtube. Here I posted more than 60 films / introductions into SPSS (more than 100 thousand downloads)

All for free, David.

drfg2008

Re: Tennis / likelyhood of winning a second time

In reply to this post by drfg2008

Thought about using a binary log Model, but couldn’t find a proper solution.

The question is part of an ongoing scientific discussion. Examples [1] [2]

[1] http://www.plosone.org/article/info:doi%2F10.1371%2Fjournal.pone.0024532

[2] http://statracket.net/A.%20James%20O%27Malley.pdf

[3] @ D. M.:

146.063 Downloads 25.03.2008 – 14.05.2012

Top Countries
Germany
Austria
Switzerland
Netherlands
United States
United Kingdom
Spain
Italy
India
Sweden

Dr. Frank Gaeth

Marta Garcia-Granero

Re: Tennis / likelyhood of winning a second time

Frank:

Did you read my reply to your first message?

See below a simple dataset, adapted from the Stata Journal paper I
mentioned then. I have changed the variable names to Set1 & Set2.

This is the syntax I used for the conditional log-binomial regression model:

* Ecuaciones de estimación generalizadas.
GENLIN Set2 (REFERENCE=FIRST)
BY Set1 (ORDER=DESCENDING)
/MODEL Set1
INTERCEPT=YES
DISTRIBUTION=BINOMIAL
LINK=LOG
/CRITERIA METHOD=FISHER(1) SCALE=1
MAXITERATIONS=100 MAXSTEPHALVING=5
PCONVERGE=1E-006(ABSOLUTE)
SINGULAR=1E-012
ANALYSISTYPE=3 CILEVEL=95
/REPEATED
SUBJECT=Match
WITHINSUBJECT=Set1
SORT=YES CORRTYPE=INDEPENDENT
ADJUSTCORR=YES COVB=ROBUST
MAXITERATIONS=100
PCONVERGE=1e-006(ABSOLUTE) UPDATECORR=1
/MISSING CLASSMISSING=EXCLUDE
/PRINT CPS DESCRIPTIVES MODELINFO FIT SUMMARY SOLUTION(EXPONENTIATED).

HTH,
Marta

Dataset:

* Use numerical values for Set1&Set2 (0=Loser; 1=Winner).
Match Set1 Set2
1 Winner Winner
1 Loser Winner
2 Loser Winner
2 Winner Winner
3 Winner Winner
3 Loser Winner
4 Winner Winner
4 Loser Winner
5 Winner Winner
5 Loser Winner
6 Loser Winner
6 Winner Winner
7 Loser Winner
7 Winner Winner
8 Winner Winner
8 Loser Winner
9 Loser Winner
9 Winner Winner
10 Loser Winner
10 Winner Winner
11 Winner Winner
11 Loser Winner
12 Loser Winner
12 Winner Winner
13 Winner Winner
13 Loser Winner
14 Winner Winner
14 Loser Winner
15 Winner Winner
15 Loser Winner
16 Loser Winner
16 Winner Winner
17 Loser Winner
17 Winner Winner
18 Winner Winner
18 Loser Winner
19 Loser Winner
19 Winner Winner
20 Winner Winner
20 Loser Winner
21 Winner Winner
21 Loser Winner
22 Winner Winner
22 Loser Winner
23 Winner Winner
23 Loser Winner
24 Winner Winner
24 Loser Winner
25 Loser Winner
25 Winner Winner
26 Winner Winner
26 Loser Winner
27 Winner Winner
27 Loser Winner
28 Loser Winner
28 Winner Winner
29 Winner Winner
29 Loser Winner
30 Loser Winner
30 Winner Winner
31 Winner Winner
31 Loser Winner
32 Winner Winner
32 Loser Winner
33 Loser Winner
33 Winner Winner
34 Loser Winner
34 Winner Winner
35 Loser Winner
35 Winner Winner
36 Loser Winner
36 Winner Winner
37 Winner Winner
37 Loser Winner
38 Winner Winner
38 Loser Winner
39 Loser Winner
39 Winner Winner
40 Winner Winner
40 Loser Winner
41 Winner Winner
41 Loser Winner
42 Loser Winner
42 Winner Winner
43 Winner Winner
43 Loser Winner
44 Loser Winner
44 Winner Winner
45 Loser Winner
45 Winner Winner
46 Loser Winner
46 Winner Winner
47 Winner Winner
47 Loser Winner
48 Winner Winner
48 Loser Winner
49 Winner Winner
49 Loser Winner
50 Winner Winner
50 Loser Winner
51 Winner Winner
51 Loser Winner
52 Winner Winner
52 Loser Winner
53 Winner Winner
53 Loser Winner
54 Loser Winner
54 Winner Winner
55 Loser Winner
55 Winner Winner
56 Loser Winner
56 Winner Winner
57 Loser Winner
57 Winner Winner
58 Winner Winner
58 Loser Winner
59 Winner Winner
59 Loser Winner
60 Winner Winner
60 Loser Winner
61 Winner Winner
61 Loser Winner
62 Winner Winner
62 Loser Loser
63 Loser Loser
63 Winner Winner
64 Loser Loser
64 Winner Winner
65 Winner Winner
65 Loser Loser
66 Winner Winner
66 Loser Loser
67 Winner Winner
67 Loser Loser
68 Winner Winner
68 Loser Loser
69 Loser Loser
69 Winner Winner
70 Loser Loser
70 Winner Winner
71 Winner Winner
71 Loser Loser
72 Winner Winner
72 Loser Loser
73 Loser Loser
73 Winner Winner
74 Winner Winner
74 Loser Loser
75 Loser Loser
75 Winner Winner
76 Loser Loser
76 Winner Winner
77 Winner Winner
77 Loser Loser
78 Loser Loser
78 Winner Winner
79 Winner Winner
79 Loser Loser
80 Winner Winner
80 Loser Loser
81 Winner Winner
81 Loser Loser
82 Loser Loser
82 Winner Winner
83 Winner Winner
83 Loser Loser
84 Loser Loser
84 Winner Winner
85 Winner Winner
85 Loser Loser
86 Winner Winner
86 Loser Loser
87 Loser Loser
87 Winner Winner
88 Winner Winner
88 Loser Loser
89 Loser Loser
89 Winner Winner
90 Loser Loser
90 Winner Winner
91 Winner Winner
91 Loser Loser
92 Winner Winner
92 Loser Loser
93 Loser Loser
93 Winner Winner
94 Loser Loser
94 Winner Winner
95 Winner Winner
95 Loser Loser
96 Winner Winner
96 Loser Loser
97 Loser Loser
97 Winner Winner
98 Loser Loser
98 Winner Winner
99 Loser Loser
99 Winner Winner
100 Winner Winner
100 Loser Loser
101 Loser Loser
101 Winner Winner
102 Winner Winner
102 Loser Loser
103 Loser Loser
103 Winner Winner
104 Winner Winner
104 Loser Loser
105 Winner Winner
105 Loser Loser
106 Loser Loser
106 Winner Winner
107 Loser Loser
107 Winner Winner
108 Loser Loser
108 Winner Winner
109 Winner Winner
109 Loser Loser
110 Winner Winner
110 Loser Loser
111 Winner Winner
111 Loser Loser
112 Winner Winner
112 Loser Loser
113 Loser Loser
113 Winner Winner
114 Winner Winner
114 Loser Loser
115 Loser Loser
115 Winner Winner
116 Winner Winner
116 Loser Loser
117 Winner Winner
117 Loser Loser
118 Loser Loser
118 Winner Winner
119 Loser Loser
119 Winner Winner
120 Loser Loser
120 Winner Winner
121 Winner Winner
121 Loser Loser
122 Winner Winner
122 Loser Loser
123 Loser Loser
123 Winner Winner
124 Loser Loser
124 Winner Winner
125 Winner Winner
125 Loser Loser
126 Winner Winner
126 Loser Loser
127 Loser Loser
127 Winner Winner
128 Loser Loser
128 Winner Winner
129 Loser Loser
129 Winner Winner
130 Loser Loser
130 Winner Winner
131 Winner Winner
131 Loser Loser
132 Loser Loser
132 Winner Winner
133 Winner Winner
133 Loser Loser
134 Loser Loser
134 Winner Winner
135 Winner Winner
135 Loser Loser
136 Loser Loser
136 Winner Winner
137 Loser Loser
137 Winner Winner
138 Winner Winner
138 Loser Loser
139 Winner Winner
139 Loser Loser
140 Winner Winner
140 Loser Loser
141 Winner Winner
141 Loser Loser
142 Winner Winner
142 Loser Loser
143 Loser Loser
143 Winner Winner
144 Winner Winner
144 Loser Loser
145 Loser Loser
145 Winner Winner
146 Loser Loser
146 Winner Winner
147 Winner Winner
147 Loser Loser
148 Loser Loser
148 Winner Winner
149 Winner Winner
149 Loser Loser
150 Loser Winner
150 Winner Loser
151 Winner Loser
151 Loser Winner
152 Winner Loser
152 Loser Winner
153 Loser Winner
153 Winner Loser
154 Loser Winner
154 Winner Loser
155 Loser Winner
155 Winner Loser
156 Winner Loser
156 Loser Winner
157 Winner Loser
157 Loser Winner
158 Loser Winner
158 Winner Loser
159 Loser Winner
159 Winner Loser
160 Winner Loser
160 Loser Winner
161 Loser Winner
161 Winner Loser
162 Winner Loser
162 Loser Winner
163 Loser Winner
163 Winner Loser
164 Winner Loser
164 Loser Winner
165 Loser Winner
165 Winner Loser
166 Winner Loser
166 Loser Winner
167 Winner Loser
167 Loser Winner
168 Loser Winner
168 Winner Loser
169 Winner Loser
169 Loser Winner
170 Loser Winner
170 Winner Loser
171 Winner Loser
171 Loser Loser
172 Winner Loser
172 Loser Loser
173 Winner Loser
173 Loser Loser
174 Loser Loser
174 Winner Loser
175 Winner Loser
175 Loser Loser
176 Winner Loser
176 Loser Loser
177 Winner Loser
177 Loser Loser
178 Winner Loser
178 Loser Loser
179 Winner Loser
179 Loser Loser
180 Winner Loser
180 Loser Loser
181 Loser Loser
181 Winner Loser
182 Loser Loser
182 Winner Loser
183 Loser Loser
183 Winner Loser
184 Winner Loser
184 Loser Loser
185 Loser Loser
185 Winner Loser
186 Winner Loser
186 Loser Loser
187 Loser Loser
187 Winner Loser
188 Winner Loser
188 Loser Loser
189 Loser Loser
189 Winner Loser
190 Loser Loser
190 Winner Loser
191 Winner Loser
191 Loser Loser
192 Winner Loser
192 Loser Loser
193 Loser Loser
193 Winner Loser
194 Loser Loser
194 Winner Loser
195 Winner Loser
195 Loser Loser
196 Winner Loser
196 Loser Loser
197 Loser Loser
197 Winner Loser
198 Winner Loser
198 Loser Loser
199 Winner Loser
199 Loser Loser
200 Winner Loser
200 Loser Loser
201 Loser Loser
201 Winner Loser
202 Winner Loser
202 Loser Loser
203 Winner Loser
203 Loser Loser
204 Loser Loser
204 Winner Loser
205 Loser Loser
205 Winner Loser
206 Loser Loser
206 Winner Loser
207 Loser Loser
207 Winner Loser
208 Loser Loser
208 Winner Loser
209 Loser Loser
209 Winner Loser
210 Loser Loser
210 Winner Loser
211 Loser Loser
211 Winner Loser
212 Winner Loser
212 Loser Loser
213 Winner Loser
213 Loser Loser
214 Loser Loser
214 Winner Loser
215 Loser Loser
215 Winner Loser
216 Winner Loser
216 Loser Loser
217 Loser Loser
217 Winner Loser
218 Winner Loser
218 Loser Loser
219 Winner Loser
219 Loser Loser
220 Winner Loser
220 Loser Loser
221 Winner Loser
221 Loser Loser
222 Winner Loser
222 Loser Loser
223 Winner Loser
223 Loser Loser
224 Loser Loser
224 Winner Loser
225 Winner Loser
225 Loser Loser
226 Winner Loser
226 Loser Loser
227 Loser Loser
227 Winner Loser
228 Winner Loser
228 Loser Loser
229 Winner Loser
229 Loser Loser
230 Winner Loser
230 Loser Loser
231 Loser Loser
231 Winner Loser
232 Loser Loser
232 Winner Loser
233 Loser Loser
233 Winner Loser
234 Winner Loser
234 Loser Loser
235 Loser Loser
235 Winner Loser
236 Loser Loser
236 Winner Loser
237 Winner Loser
237 Loser Loser
238 Winner Loser
238 Loser Loser
239 Loser Loser
239 Winner Loser
240 Winner Loser
240 Loser Loser
241 Loser Loser
241 Winner Loser
242 Loser Loser
242 Winner Loser
243 Winner Loser
243 Loser Loser
244 Winner Loser
244 Loser Loser
245 Loser Loser
245 Winner Loser
246 Winner Loser
246 Loser Loser
247 Winner Loser
247 Loser Loser
248 Winner Loser
248 Loser Loser
249 Winner Loser
249 Loser Loser
250 Winner Loser
250 Loser Loser
251 Loser Loser
251 Winner Loser
252 Loser Loser
252 Winner Loser
253 Winner Loser
253 Loser Loser
254 Loser Loser
254 Winner Loser
255 Loser Loser
255 Winner Loser
256 Winner Loser
256 Loser Loser
257 Winner Loser
257 Loser Loser
258 Winner Loser
258 Loser Loser
259 Loser Loser
259 Winner Loser
260 Winner Loser
260 Loser Loser
261 Loser Loser
261 Winner Loser
262 Loser Loser
262 Winner Loser
263 Winner Loser
263 Loser Loser
264 Loser Loser
264 Winner Loser
265 Loser Loser
265 Winner Loser
266 Loser Loser
266 Winner Loser
267 Winner Loser
267 Loser Loser
268 Loser Loser
268 Winner Loser
269 Winner Loser
269 Loser Loser
270 Loser Loser
270 Winner Loser
271 Loser Loser
271 Winner Loser
272 Loser Loser
272 Winner Loser
273 Loser Loser
273 Winner Loser
274 Winner Loser
274 Loser Loser
275 Loser Loser
275 Winner Loser
276 Winner Loser
276 Loser Loser
277 Winner Loser
277 Loser Loser
278 Winner Loser
278 Loser Loser
279 Winner Loser
279 Loser Loser
280 Winner Loser
280 Loser Loser
281 Loser Loser
281 Winner Loser
282 Winner Loser
282 Loser Loser
283 Winner Loser
283 Loser Loser
284 Loser Loser
284 Winner Loser
285 Loser Loser
285 Winner Loser
286 Loser Loser
286 Winner Loser
287 Winner Loser
287 Loser Loser
288 Winner Loser
288 Loser Loser
289 Loser Loser
289 Winner Loser
290 Loser Loser
290 Winner Loser
291 Loser Loser
291 Winner Loser
292 Winner Loser
292 Loser Loser
293 Loser Loser
293 Winner Loser
294 Winner Loser
294 Loser Loser
295 Loser Loser
295 Winner Loser
296 Loser Loser
296 Winner Loser
297 Loser Loser
297 Winner Loser
298 Winner Loser
298 Loser Loser
299 Winner Loser
299 Loser Loser
300 Winner Loser
300 Loser Loser
301 Winner Loser
301 Loser Loser
302 Loser Loser
302 Winner Loser
303 Winner Loser
303 Loser Loser
304 Loser Loser
304 Winner Loser
305 Winner Loser
305 Loser Loser
306 Loser Loser
306 Winner Loser
307 Loser Loser
307 Winner Loser
308 Loser Loser
308 Winner Loser
309 Winner Loser
309 Loser Loser
310 Loser Loser
310 Winner Loser
311 Loser Loser
311 Winner Loser

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD

drfg2008

Re: Tennis / likelyhood of winning a second time

Thanks Marta!

This is exactly what I was looking for. I tested the Syntax (probably Match should be unique?) with your data.

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

Parameter Estimates
Parameter B Std. Error 95% Wald Confidence Interval Hypothesis Test Exp(B) 95% Wald Confidence Interval for Exp(B)
Lower Upper Wald Chi-Square df Sig. Lower Upper
(Intercept) -.652 .0544 -.759 -.546 143.829 1 .000 .521 .468 .579
[Set1=2] .346 .0641 .220 .472 29.158 1 .000 1.414 1.247 1.603
[Set1=1] 0a . . . . . . 1 . .
(Scale) 1
Dependent Variable: Set2
Model: (Intercept), Set1
a Set to zero because this parameter is redundant.

Dr. Frank Gaeth

Marta Garcia-Granero

Re: Tennis / likelyhood of winning a second time

El 15/05/2012 21:30, drfg2008 escribió:
> Thanks Marta!
>
> This is exactly what I was looking for. I tested the Syntax (probably Match
> should be unique?) with your data.

Yes, Match values should be, as in the example, 1 1 2 2 3 3.... It
identifies the two players (winner&loser of the first set) that played
one against the other. You can't have more than a pair of "1", or "2"...

BTW, I did a bit of math with the help of old fashioned scrap paper &
pencil and realized that your design can be seen as either a matched
cohort design (you classify players for exposition - 1t set outcome- and
watch the 2nd set to classify them as loser/winner), or matched
case-control design (you classify players for the outcome of the 2nd
set, and look retrospectively for the outcome of the first). This makes
the output of a conditional log-binomial model and conditional logistic
model exactly the same. Also, you could use contingency tables
stratified by Mach, and ask for the Mantel-Haenszel OR, and get the same
result.

Regards,
Marta GG

>
> *************************
>
> Parameter Estimates
> Parameter B Std. Error 95% Wald Confidence Interval Hypothesis Test
> Exp(B) 95% Wald Confidence Interval for Exp(B)
> Lower Upper Wald Chi-Square df Sig. Lower Upper
> (Intercept) -.652 .0544 -.759 -.546 143.829 1 .000 .521 .468 .579
> [Set1=2] .346 .0641 .220 .472 29.158 1 .000 1.414 1.247 1.603
> [Set1=1] 0a . . . . . . 1 . .
> (Scale) 1
> Dependent Variable: Set2
> Model: (Intercept), Set1
> a Set to zero because this parameter is redundant.
>
>
>
> -----
> Dr. Frank Gaeth
> FU-Berlin
>
> --
> View this message in context: http://spssx-discussion.1045642.n5.nabble.com/Tennis-likelyhood-of-winning-a-second-time-tp5708820p5710325.html
> Sent from the SPSSX Discussion mailing list archive at Nabble.com.
>
> =====================
> To manage your subscription to SPSSX-L, send a message to
> [hidden email] (not to SPSSX-L), with no body text except the
> command. To leave the list, send the command
> SIGNOFF SPSSX-L
> For a list of commands to manage subscriptions, send the command
> INFO REFCARD
>

drfg2008

Re: Tennis / likelyhood of winning a second time

In reply to this post by Marta Garcia-Granero

@Marta García-Granero-2

Thank you for the very interesting ideas you provided!

Just a question to your data: As far as I understand now, each game is not unique, each game occurs two times and has two times the same matchnumber: 1-1, 2-2, etc.. But if you consider a Best of 3 game for example you only have the following alternatives:

win - win (2:0)
win - lose - win (2:1)
lose - win - lose (1:2)
lose - lose (0:2)

So if you test the hypothesis of a higher probability after winning a set, you only have the (2:0) solution in best of three (best of five is more complicated). How can you have two lines for one game as in your example.

Frank

Dr. Frank Gaeth