Jump to content

Madison Tecmo X: By The Numbers


hoigaard

Recommended Posts

I’ve always been infatuated with statistics/demographics, etc and this tourney is no different. I wanted to post some stuff about the tourney but didn’t want to clog up the main Madison thread, so I thought I’d start this one as a sort of repository for any posts/discussions about “Madison, By the Numbers.” Things like team usage, player stats, locations, many of the things that Josh tweets from Madison Tecmo Twitter acct, etc. If you find any stats/numbers you think are interested, add them to the thread. To get us started, I’m going to re-post a couple things that I posted last year:


 


To preemptively talk myself off the ledge after a disastrous showing on Saturday (especially after the first two “major” previews *Flo & CVogt* predicted me to advance), here are some numbers to put in perspective the difficulty of leaving this tourney in good standing:


 


25% of the field will go 0-2.


25% of the field will go 1-2.


12.5 to 25% of the field will go 2-2.


 


Half of the field will leave with a losing record, and only 25 to 37.5% of all entrants will depart with any semblance of a "winning" (above .500) record.


 


Next, I wanted to see what the general rate of team selection has been in the Madison Tourney, as well as team winning percentages. Tony O started designing a page for that but I don’t believe it’s quite ready, so I did some pen/paper/Excel work. Here's what I got:


 


Data collection information- The current method of team selection started in Tecmo VI, so for this study, I only looked at Tecmo VI, VII, and VIII (*Note, this study was done prior to Tecmo IX, and I haven’t updated. I think the general theme/trends would hold, even if I did). I chose not to look at tournaments from other cities because I wanted to limit this strictly to Madison (you can be certain to a strong degree that most of these numbers would hold across multiple tourneys). From my shitty math, there were 796 games played in Tecmo VI, VII, and VIII. To get a reasonable sample size w/ 90% accuracy, I needed about 200 data points (roughly 25%). I chose, at random, various guys' results from those tourneys (making sure not to duplicate. Made-up example: Say Joe Blow beats Jim Balls in Tecmo VII. I would not count that same result if I was tallying Jim Balls results).


 


Things this study did not factor in- Strength of player, Who called the matchup, Early-round games vs late-tourney games (many instances teams weren't listed in the results, so I couldn't use them), Controller 1/2, etc. This was strictly "What were the two teams selected, and who won that particular game?"


 


Results (let me repeat: This was FAR from an exact science. It's possible a number or team was screwed up here or there, but all in all, it gives you the general idea of what we're looking for)-


 


Total # of Uses:


 


Total Uses (Usage %)...


SD 24 (11.62%)


Min 23 (11.14%)


Cin 22 (10.66%)


Den 22 (10.66%)


Dal 22 (10.66%)


Wash 21 (10.16%)


Det 20 (9.68%)


Rams 20 (9.68%)


Jets 18 (8.72%)


Phx 18 (8.72%)


Atl 17 (8.24%)


Mia 16 (7.74%)


TB 16 (7.74%)


Cle 15 (7.26%)


KC 14 (6.78%)


Chi 14 (6.78%)


NO 14 (6.78%)


Pit 12 (5.82%)


GB 12 (5.82%)


Buf 11 (5.32%)


NYG 10 (4.84%)


Phi 10 (4.84%)


NE 9 (4.36%)


Oak 8 (3.88%)


Sea 8 (3.88%)


SF 7 (3.38%)


Ind 5 (2.42%)


Hou 5 (2.42%)


 


 


Winning % of Teams:


Wins (Win %)


SF 6 (.857)


TB 13 (813)


Phi 8 (.800)


Pit 8 (.667)


Buf 7 (.636)


Rams 12 (.600)


Hou 3 (.600)


Cin 13 (.591)


SD 14 (.583)


GB 7 (.583)


NO 8 (.571)


Min 13 (.565)


NE 5 (.556)


Cle 8 (.533)


Wash 11 (.524)


Phx 9 (.500)


KC 7 (.500)


Chi 7 (.500)


Dal 10 (.455)


Det 8 (.400)


Ind 2 (.400)


Mia 6 (.375)


NYG 3 (.300)


Atl 5 (.294)


Sea 2 (.250)


Den 5 (.227)


Jets 4 (.222)


Oak 1 (.125)


Link to comment
Share on other sites

I always like reading these.  


I was surprised, both in your results and Tony O's grid, that TB had a winning record against virtually every team they played.  TB and DET are teams that really increase in value in skilled man vs. man Tecmo, and the numbers prove it.


Has really adjusted my "top 3" matchups for this year.  Usually I like calling TB v. PHX or TB v. ATL, but now I think I can almost step up TB and have TB v. SD and TB v. DEN and not really be too worried.


Link to comment
Share on other sites

Last night I wondered “Which region has the most Madison experience? The most wins? The most losses? Etc, etc, etc” So, let’s take a look! (A spreadsheet with the raw data is attached to follow along):


 


The first thing that sticks out is the lack of Madison (and overall tourney) experience in Group 3. Only 1 of the 8 entrants has any Madison experience (William K, 3 years worth). Only two other groups have 3 players with Madison experience, and no others have fewer than 4.


 


Three groups (22, 25, and 32) have 6 of the 8 entrants with prior Madison experience. Oddly, though, in Group 22 those 6 players have only combined to go 17-23.


 


If you look specifically at collective group years with Madison experience, Group 6 is interesting. They have 5 guys who have played in Madison before for a COMBINED 22 years of experience, so each of them averages more than 4 years experience on their own. The next closest groups are 14 and 17 with 19 years of Madison experience each.


 


The groups with the most Madison wins is 29. Their collective 63 wins are led by Jay H and Jimmy B. The next group is Group 9 (Chet's) with 58 wins, followed by Group 24's 57 wins.


 


The fewest Madison wins belongs to Group 16 and 13. Despite 4 players with prior experience in each, they enter Tecmo X at 9-10 and 10-9, respectively. The next fewest is Group 3, whose 11 wins belong to one guy.


 


Group 18 has 13 years of Madison experience among 4 guys, but has the worst all-time winning %, mustering only a 19-28 record and .404. When other tournaments are factored in, that % drops to .373. In fact, 5 groups have an all-time sub-.500 record in recorded tournament play (not just Madison, but others registered through the site): 18 (.373), 16 (.409), 15 (.440), 13 (.476), and 19 (.477). 


 


Here is the grid of groups, sorted by total years of collective Madison experience, then winning % (these may not copy/paste cleanly, my apologies):


 



 

Mad Win

Mad Loss

Tot. Win

Tot. Loss

Mad. Yrs

 

# w/ Mad Ex

Mad Win %

Tot. Win %

Group 6

49

39

64

47

22

 

5

.557

.577

Group 14

49

33

96

61

19

 

5

.598

.611

Group 17

53

36

131

79

19

 

5

.596

.624

Group 29

63

26

63

28

17

 

5

.708

.692

Group 26

54

26

153

55

16

 

4

.675

.736

Group 32

47

23

92

47

16

 

6

.671

.662

Group 25

36

27

69

40

16

 

6

.571

.633

Group 28

32

25

64

54

16

 

5

.561

.542

Group 19

31

34

31

34

16

 

5

.477

.477

Group 24

57

23

79

39

15

 

5

.713

.669

Group 9

58

24

157

38

15

 

5

.707

.805

Group 11

36

25

45

37

15

 

4

.590

.549

Group 5

36

24

76

49

14

 

5

.600

.608

Group 8

35

21

53

32

13

 

4

.625

.624

Group 18

19

28

19

32

13

 

4

.404

.373

Group 22

17

23

32

28

12

 

6

.425

.533

Group 12

34

17

90

41

11

 

5

.667

.687

Group 20

29

20

61

35

11

 

5

.592

.635

Group 27

31

17

39

22

10

 

5

.646

.639

Group 21

28

17

81

41

10

 

4

.622

.664

Group 7

22

21

37

36

10

 

4

.512

.507

Group 15

18

19

22

28

10

 

4

.486

.440

Group 1

32

17

62

27

9

 

4

.653

.697

Group 4

15

18

62

27

9

 

5

.455

.697

Group 2

24

13

48

27

8

 

4

.649

.640

Group 10

20

13

74

37

8

 

4

.606

.667

Group 31

19

16

31

22

8

 

5

.543

.585

Group 23

15

14

15

14

7

 

3

.517

.517

Group 30

13

10

17

14

6

 

3

.565

.548

Group 13

10

9

20

22

5

 

4

.526

.476

Group 16

9

10

9

13

5

 

4

.474

.409

Group 3

11

3

25

9

3

 

1

.786

.735

 


So, that brings us back to the original question: What is the experience level by region? Here's how they break down:


 



REG. TOTAL

Mad Win

Mad Loss

Tot. Win

Tot. Loss

Mad. Yrs

 

# w/ Mad. Ex.

Mad Win %

Tot. Win %

Fawcett

295

170

528

282

105

 

39

.634

.652

Martin

249

195

449

302

103

 

37

.561

.598

DeLuise

234

150

513

277

88

 

35

.609

.649

Reynolds

224

156

427

254

88

 

32

.589

.627

 


Historical strength seems to favor the Farrah Fawcett Region. They have the most entrants with Madison experience, the most Madison wins, the most overall tourney wins, the highest Madison winning % and the highest overall tournament winning %. The Burt Reynolds Region would be defined as the most "raw," as they have the fewest Madison and overall tournament wins, as well as the fewest guys with previous Madison history. It's interesting that the Dean Martin Region, despite having the 2nd-most Madison experience, actually such an enormous discrepancy in winning percentages (just .561 in Madison and the only group under .600 overall in all tournament play).


 


These numbers are not meant to define "strength" or how difficult anyone's road is to a deep run. With so many unknowns in the tourney each year and new guys from different parts of the country who have proven themselves online or elsewhere but have little live tournament play, you cannot definitively say that any one path is "easier" than another. As we can all assume, trying to seed this tournament and spread talent evenly is the perfect nightmare. But the numbers are fun to look at!


Tecmo 2014 Entrants.xlsx

Link to comment
Share on other sites

Honestly the player strength account for so much in team matchups. This is even more true in the opening rounds of madison where many games are complete blowouts.It would have probably been more useful to either included all the games to get a bigger sample or restrict the sample to the field of 64 games where "in theory" the player skill differential isn't as large. 


 


 


I remember once trying to do something similar with a subset of 30,000 TPC games. It was a nightmare even trying to control for all the variables or the fact that one great player might represent a large portion of the games played with a specific team.


 


 


So these numbers are just FUN information to me. 

Link to comment
Share on other sites

So far in this thread we've looked at % chance at "success" in Tecmo X, team usage/win rate, and previous Madison/tourney experience. Now let's look at the geography of the 256 entrants. (A couple entrants had vague or incomplete location info on their profile page, and the site I used wouldn't let me map Costa Rica, so one or two may be unaccounted for. Data available as an attached spreadsheet).


 


Obviously the tournament is dominated by Wisconsin residents. 99 of the 256 entrants hail from the greatest state in the union. That being said, there are a few groups where WI residents are scarce. Of the 32 groups, only 1 (Group 27) has no players from the state of WI, while 4 groups have just one WI entrant each (groups 4, 10, 16, 30). The most WI-heavy are groups 18 and 19, with 6 entrants.


 


As far as overall state variety in group play, many groups have 6 different states represented, but group 23 leads the charge with a whopping 7 different states. A small handful of groups have only 3 states represented (groups 13, 18, and 19).


 


When broken down by region, Dean Martin sticks out as the least-diverse geographically. 32 of their 64 competitors are from WI, and only 16 total states are represented. This compared to Reynolds (18 states, 25 WI), Deluise (19 states, 19 WI), and Fawcett (19 states, 23 WI).


 


Here are visuals of the 4 regions and overall field:


-Burt Reynolds


-Dom Deluise


-Dean Martin


-Farrah Fawcett


-All Entries


 


The Fawcett region sticks out here as the most "Eastern" of the regions. 17 of the entrants hail from Michigan and eastward. In the Burt Reynolds region, Tanvir H has to carry the western US all by himself, as he is from CA and no other competitor is as far west as Dallas. The Deluise region, visually, is the most Upper Midwest-centric of the 4, as all but 10 entrants hail from the belly of the Tecmo beast (15 from IL, especially).


 


So what does all of this tell us? Well, nothing, really. It's just fun to look at maps and numbers. But it does illustrate the point that the regions are drawn up quite well as far as geographic distribution is concerned. With 256 entrants, the seeding committee managed to make each region's state distribution no fewer than 16 and no more than 19. That's a pretty even swath, all things considered. And it means there will be plenty of opportunities to meet new friends from all over the country, even if only within your group or region.


Tecmo 2014 Geography.xlsx

Edited by hoigaard
Link to comment
Share on other sites

Thanks. Pins on the map weren't my first choice. I really wanted to make a heat/intensity map since the pin maps only shows one pin even if, say, 20 guys are from the same city (whereas a heat map would show a brighter/more intense color based on density of where people are from). But after an unreasonable amount of time searching, I couldn't figure out how/where to do it.


Link to comment
Share on other sites

Also 9 to 10 groups have 2 players with the same name. What's it means? Nothing... probably just easier to track guys that way....say like a potential battle of the matt's between toolie and the godfather.

Edited by red98sethuthut
Link to comment
Share on other sites

Thanks. Pins on the map weren't my first choice. I really wanted to make a heat/intensity map since the pin maps only shows one pin even if, say, 20 guys are from the same city (whereas a heat map would show a brighter/more intense color based on density of where people are from). But after an unreasonable amount of time searching, I couldn't figure out how/where to do it.

The heat/intensity up in the PNW would be pretty high as we are pretty dense up here, just like our fog

Link to comment
Share on other sites

Courtesy of my Map Guy, (a friend who works in city zoning), here is a kinda-sorta "heat map" I wanted to make of the entrants (works best in Chrome, and give it a little bit to load):


http://www.arcgis.com/apps/OnePane/basicviewer/index.html?appid=9ccf92de92f74865a16ab1c13c1932d5


Edited by hoigaard
Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

×
×
  • Create New...