Remove this ad

Lead

Jun 26 17 2:27 PM

Tags : :

i am in the process of finishing up basketball, i'm on 1993 right now (started with 1926). any suggestions of what i should start adding next? for now i am sticking with men's team sports (minimum of 5 players a side), though i'd like to eventually add sepak takraw and tennis. also, i am trying to avoid, or at least delay, adding lesser versions of bigger sports (like rugby sevens and 3x3 basketball). i sort of want to start working on water polo or roller hockey. i dread cricket because of the odd scoring and the fact that the first international was played in 1844, so i'm sort of avoiding it... but still open to doing it. i should mention that i'm in the process of casually working on rugby union, ice hockey, baseball, and hurling... so those are all on the table as well.

COMPLETED SPORTS:
American Football
Aussie Rules
Circle Kabaddi
International Rules Football
Soccer Football
Standard Kabaddi

i got a domain, but not sure if i'm going to keep using it: www.sportsworldrankings.com

Last Edited By: abramjones Jun 26 17 2:34 PM. Edited 1 time

Quote    Reply   
Remove this ad
Remove this ad
avatar

TheRoonBa

Posts: 5,526 Site Admin

#1 [url]

Jun 27 17 1:05 AM

Field Hockey, Handball?

Hurling - there haven't been any international matches in hurling. Only a hybrid shinty/hurling match between Scotland and Ireland, which is neither shinty nor hurling. Shinty, on the other hand, has had 2 international matches (both England v USA).

The cricket scoring system is not that difficult, really. The point of the game is always to score more runs than the other team. Both teams bat until either A) all 10 wickets are lost or B) they reach the target score (in runs).
If there is bad weather, the target score is adjusted (for example, if Team A played 50 overs, but Team B only gets to play 20 because of bad weather, obviously their target would be less - i.e. what Team A would have scored after 20 overs).
For example, in the most recent match - India 310-5 (43.0) West Indies 205-6 (43.0). Match reduced to 43 overs because of bad weather. The final score can be read as 310-205 to India. The margin of victory can be scaled up from 43 to 50 overs if required (334-221 would be the final score).
If we look at a match where the second team won:
South Africa 142-3 (20.0) England 143-1 (14.3) - England reached their target (143) with 5.3 overs still left, and lost only 1 wicket. Therefore, to calculate the score for this match, one could simply design a formula to estimate what England would have scored if they kept batting until the end of 20 overs. Then both scores can be "scaled up" to what they would have been if it had been a 50-over match. For my rankings purposes, this match result for me would be South Africa 251-335 England (the estimated scores the teams would have got had they both batted 50 overs).  If England has lost 9 wickets (instead of 1), there score would have been much reduced, as they would only have been 1 "out" away from their innings ending.  The scaled up scores in this case would have been 251-271.




EDIT: OK, I used 12 lines to explain that cricket scoring is simple.smiley: roll

Last Edited By: TheRoonBa Jun 27 17 1:12 AM. Edited 2 times.

Quote    Reply   

#2 [url]

Jun 27 17 4:27 AM

TheRoonBa wrote:
Field Hockey, Handball?

Hurling - there haven't been any international matches in hurling. Only a hybrid shinty/hurling match between Scotland and Ireland, which is neither shinty nor hurling. Shinty, on the other hand, has had 2 international matches (both England v USA).

The cricket scoring system is not that difficult, really. The point of the game is always to score more runs than the other team. Both teams bat until either A) all 10 wickets are lost or B) they reach the target score (in runs).
If there is bad weather, the target score is adjusted (for example, if Team A played 50 overs, but Team B only gets to play 20 because of bad weather, obviously their target would be less - i.e. what Team A would have scored after 20 overs).
For example, in the most recent match - India 310-5 (43.0) West Indies 205-6 (43.0). Match reduced to 43 overs because of bad weather. The final score can be read as 310-205 to India. The margin of victory can be scaled up from 43 to 50 overs if required (334-221 would be the final score).
If we look at a match where the second team won:
South Africa 142-3 (20.0) England 143-1 (14.3) - England reached their target (143) with 5.3 overs still left, and lost only 1 wicket. Therefore, to calculate the score for this match, one could simply design a formula to estimate what England would have scored if they kept batting until the end of 20 overs. Then both scores can be "scaled up" to what they would have been if it had been a 50-over match. For my rankings purposes, this match result for me would be South Africa 251-335 England (the estimated scores the teams would have got had they both batted 50 overs).  If England has lost 9 wickets (instead of 1), there score would have been much reduced, as they would only have been 1 "out" away from their innings ending.  The scaled up scores in this case would have been 251-271.



EDIT: OK, I used 12 lines to explain that cricket scoring is simple.smiley: roll

field hockey and handball are definitely sports i need to do.

in Ireland, as i'm sure you're aware, hurling is often played between counties. these are in essence national teams, as their roster requirements are much stricter than matches between regions in Australia for Australian Rules Football. speaking of which, Australia has also played a few matches against Irish counties in Gaelic Football before the International Rules Series.

thanks for the information on cricket, i'm going to have to read this over a few times...  i had a very simple formulae to merge all 3 forms of cricket into 1 scoring system, but it was more closely related to net run rate, though it factored in wickets. i'm a little fuzzy on remember the details so i will have to re-investigate.  what about full test matches where overs are not limited? how can this be done with your system?

Last Edited By: abramjones Jun 27 17 5:25 AM. Edited 2 times.

Quote    Reply   
avatar

TheRoonBa

Posts: 5,526 Site Admin

#3 [url]

Jun 27 17 12:42 PM

For cricket, a table is required that calculates percentage of resources remaining (a combination of overs and wickets remaining).

For Test cricket, it's a little difficult. Scoring rate is much lower - what I do is that I re-calibrate the innings as if it were a 50-over match, so for example a 120-over innings becomes a 50-over innings, and the score is adjusted.

You can see an example of a ball-by-ball resource percentage table for cricket here:
http://www.ccua.ca/Duckworth-Lewis_Tables._Table_of_ball_by_ball_.pdf

As you will see, the point in number of overs (e.g. 48.4) does not represent a decimal - but is 48 overs and 4 balls (there are 6 balls in each over) - so it's actually 48.666666 overs.

Quote    Reply   

#5 [url]

Jun 27 17 11:41 PM

TheRoonBa wrote:
For cricket, a table is required that calculates percentage of resources remaining (a combination of overs and wickets remaining).

For Test cricket, it's a little difficult. Scoring rate is much lower - what I do is that I re-calibrate the innings as if it were a 50-over match, so for example a 120-over innings becomes a 50-over innings, and the score is adjusted.

You can see an example of a ball-by-ball resource percentage table for cricket here:
http://www.ccua.ca/Duckworth-Lewis_Tables._Table_of_ball_by_ball_.pdf

As you will see, the point in number of overs (e.g. 48.4) does not represent a decimal - but is 48 overs and 4 balls (there are 6 balls in each over) - so it's actually 48.666666 overs.

this is quite complicated. what does the term "wickets down" mean?

Quote    Reply   

#6 [url]

Jun 27 17 11:45 PM

nfm24 wrote:
Try not to pick sports where there are a lot of "unknown" international matches, historically.

yeah, i'm always afraid of this. most sports are now sufficiently documented as far as main tournaments go. however, friendlies are a completely different matter. even in basketball friendlies are not very well documented. football, rugby, and cricket are the only sports i've seen that document friendlies well. i just started handball yesterday and the only thing i'm having trouble finding are early results for the African championship. in most sports tournaments that are not considered main tournaments can also present a problem for research. without www.theroonba.com , www.todor66.com and www.wikipedia.org i would be screwed. my list of essential sources is getting pretty extensive: https://docs.zoho.com/writer/open/1vl7i972fbd149897431685118db97a9a5f40

Last Edited By: abramjones Jun 28 17 12:07 AM. Edited 2 times.

Quote    Reply   

#7 [url]

Jun 28 17 12:37 PM

Why not try and fill in the gaps in the data yourself, by researching the "missing" tournaments or working with those who do so. 

E.g. the Far East Games results I posted, this sort of thing isn't hard.

Quote    Reply   
avatar

TheRoonBa

Posts: 5,526 Site Admin

#8 [url]

Jun 29 17 12:53 AM

abramjones wrote:
TheRoonBa wrote:
For cricket, a table is required that calculates percentage of resources remaining (a combination of overs and wickets remaining).

For Test cricket, it's a little difficult. Scoring rate is much lower - what I do is that I re-calibrate the innings as if it were a 50-over match, so for example a 120-over innings becomes a 50-over innings, and the score is adjusted.

You can see an example of a ball-by-ball resource percentage table for cricket here:
http://www.ccua.ca/Duckworth-Lewis_Tables._Table_of_ball_by_ball_.pdf

As you will see, the point in number of overs (e.g. 48.4) does not represent a decimal - but is 48 overs and 4 balls (there are 6 balls in each over) - so it's actually 48.666666 overs.

this is quite complicated. what does the term "wickets down" mean?

"Wickets down" is how many wickets a team has lost (i.e. how many players are out).  There are 11 players in a team, and they play in pairs, one at each end of the wicket (a run is scored when both members of the pair run simultaneously (one of whom is the batsman) and reach the other end of the wicket, thus swapping places).  When 10 players are out, that team's innings are over (as the 11th player no longer has a partner with which to score runs). "0 wickets down" is thus the state of play at the start of an innings.


Quote    Reply   
avatar

TheRoonBa

Posts: 5,526 Site Admin

#9 [url]

Jun 29 17 1:00 AM

abramjones wrote:
nfm24 wrote:
Try not to pick sports where there are a lot of "unknown" international matches, historically.

yeah, i'm always afraid of this. most sports are now sufficiently documented as far as main tournaments go. however, friendlies are a completely different matter. even in basketball friendlies are not very well documented. football, rugby, and cricket are the only sports i've seen that document friendlies well. i just started handball yesterday and the only thing i'm having trouble finding are early results for the African championship. in most sports tournaments that are not considered main tournaments can also present a problem for research. without www.theroonba.com , www.todor66.com and www.wikipedia.org i would be screwed. my list of essential sources is getting pretty extensive: https://docs.zoho.com/writer/open/1vl7i972fbd149897431685118db97a9a5f40

Friendlies are not really necessary - and in most cases can complicate things, as teams use squads of variable strength.  In international tournaments, teams will more often than not use their best available players.

I include friendly matches in sports where there are lots (football, futsal and beach soccer, ice hockey, rugby union and cricket) and where they are relatively important.  There are also a fair number in basketball, handball, field hockey, volleyball, water polo - but results of these are poorly documented.  And where they are documented (archives on live score sites), their inclusion is arbitrary - probably fewer than 50% of all friendly matches make it onto these sites.  For these sports, I concentrated only on world tournaments, continental tournaments, regional tournaments and tournaments that are part of multi-sport events (again world, continental and regional).  I also went back only to the year 2000 for most of these sports. I find this is sufficient to produce meaningful comparisons (and also, I am fairly confident I have collected between 95-100% of all the results from these tournaments, whereas if I'd tried to go back beyond 2000, that number would have decreased significantly).  The reason I've only gone back to 2010 on the website is due to lack of time to piss about and format all the results and put in all the venues, dates, etc.  It took me long enough just to go back to 2010...

My spreadsheets of results were started just to have an ordered list of results for ranking purposes, and therefore I didn't include dates, as I didn't need them.








Last Edited By: TheRoonBa Jun 29 17 1:03 AM. Edited 1 time.

Quote    Reply   
Remove this ad

#10 [url]

Jun 29 17 4:49 PM

TheRoonBa wrote:
My spreadsheets of results were started just to have an ordered list of results for ranking purposes, and therefore I didn't include dates, as I didn't need them.
This is fine if competition frequency is fairly regular (and it will be if you are only considering major competitions) but for teams which play infrequently, a few results scattered over a decade say, the same assumptions wouldn't apply and there should be a bias towards most recent results.

Quote    Reply   
avatar

TheRoonBa

Posts: 5,526 Site Admin

#11 [url]

Jun 29 17 6:06 PM

Yes, there is a time bias in all the spreadsheets - it's just ordinal rather than by date. In practice, this doesn't matter too much,

For example, in a list 1-1000 (1000 being most recent), there might be bigger gaps (datewise) between each match, but because I don't have dates, there is a gap of exactly 1 between each game instead of 3 days, 4 weeks, etc... The time bias is exponential also rather than linear. However, I'm not doing too well at the moment getting to work exactly how I want it (although in the most part it works well, for certain cases, it works absolutely terribly). In one instance, one of the program runs resulted in Tristan da Cunha being Number 1, just ahead of Brazil.

Quote    Reply   

#12 [url]

Jun 29 17 9:46 PM

TheRoonBa wrote:
In one instance, one of the program runs resulted in Tristan da Cunha being Number 1, just ahead of Brazil.
This explains why in another older thread, your simulations had Tristan da Cunha having a surprisingly large chance of winning the World Cup smiley: laugh

Quote    Reply   

#13 [url]

Jul 3 17 8:16 AM

nfm24 wrote:
Why not try and fill in the gaps in the data yourself, by researching the "missing" tournaments or working with those who do so. 

E.g. the Far East Games results I posted, this sort of thing isn't hard.

i'm doing what i can, data entry alone is extremely time consuming. it took me about 5 months of constant work just to complete soccer football.

Quote    Reply   

#14 [url]

Jul 3 17 8:30 AM

TheRoonBa wrote:
abramjones wrote:
nfm24 wrote:
Try not to pick sports where there are a lot of "unknown" international matches, historically.

yeah, i'm always afraid of this. most sports are now sufficiently documented as far as main tournaments go. however, friendlies are a completely different matter. even in basketball friendlies are not very well documented. football, rugby, and cricket are the only sports i've seen that document friendlies well. i just started handball yesterday and the only thing i'm having trouble finding are early results for the African championship. in most sports tournaments that are not considered main tournaments can also present a problem for research. without www.theroonba.com , www.todor66.com and www.wikipedia.org i would be screwed. my list of essential sources is getting pretty extensive: https://docs.zoho.com/writer/open/1vl7i972fbd149897431685118db97a9a5f40

Friendlies are not really necessary - and in most cases can complicate things, as teams use squads of variable strength.  In international tournaments, teams will more often than not use their best available players.

I include friendly matches in sports where there are lots (football, futsal and beach soccer, ice hockey, rugby union and cricket) and where they are relatively important.  There are also a fair number in basketball, handball, field hockey, volleyball, water polo - but results of these are poorly documented.  And where they are documented (archives on live score sites), their inclusion is arbitrary - probably fewer than 50% of all friendly matches make it onto these sites.  For these sports, I concentrated only on world tournaments, continental tournaments, regional tournaments and tournaments that are part of multi-sport events (again world, continental and regional).  I also went back only to the year 2000 for most of these sports. I find this is sufficient to produce meaningful comparisons (and also, I am fairly confident I have collected between 95-100% of all the results from these tournaments, whereas if I'd tried to go back beyond 2000, that number would have decreased significantly).  The reason I've only gone back to 2010 on the website is due to lack of time to piss about and format all the results and put in all the venues, dates, etc.  It took me long enough just to go back to 2010...

My spreadsheets of results were started just to have an ordered list of results for ranking purposes, and therefore I didn't include dates, as I didn't need them.








i agree that friendlies can complicate things in several situations. not only do they sometimes use lesser quality rosters as you mention (this can happen in tournaments often as well), but friendlies are often played by countries in the same region. then regional ratings for certain teams may start drifting incorrectly too much because of a lack of variety of games between different regions. i do not include friendlies that are played against countries in the same region. having said that, i still feel using certain friendlies will still provide a more accurate rating. and i feel you on the length of time it takes to document all these matches, that is one reason i am so impressed with this website.

Quote    Reply   

#15 [url]

Jul 3 17 4:53 PM

abramjones wrote:
nfm24 wrote:
Why not try and fill in the gaps in the data yourself, by researching the "missing" tournaments or working with those who do so. 
i'm doing what i can, data entry alone is extremely time consuming. it took me about 5 months of constant work just to complete soccer football.

Well my point more that you haven't "completed" it as such, because there are valid historical matches you don't have in your list (and nor does anyone else except perhaps a couple of researchers such as those on this forum).  Granted, in football the percentage of unknowns is smaller than in other sports, but this is precisely why it is an important consideration before choosing a new sport.   There is no point generating historical rankings of, say, Uganda in field hockey if you don't have any of the results of East African hockey competitions.  In that case, if you aim was to generate reliable hockey rankings worldwide, you would need to first reduce the data gap.

Quote    Reply   

#16 [url]

Jul 3 17 4:57 PM

abramjones wrote:
TheRoonBa wrote:
Friendlies are not really necessary - and in most cases can complicate things, as teams use squads of variable strength.  In international tournaments, teams will more often than not use their best available players.
i agree that friendlies can complicate things in several situations. not only do they sometimes use lesser quality rosters as you mention (this can happen in tournaments often as well), but friendlies are often played by countries in the same region. .... i do not include friendlies that are played against countries in the same region. having said that, i still feel using certain friendlies will still provide a more accurate rating.
Personally I think if you aren't going to include friendlies, then there is no point compiling a list of matches (or a ranking) as it is clearly incomplete a priori.  Likewise if a ranking system isn't able to accommodate varying conditions of friendlies then in my view it is too simplistic for its purpose. 

Mind you, most rankings don't incorporate home advantage or winning margin, and are hence inadequate anyway.

Quote    Reply   

#17 [url]

Jul 13 17 6:01 AM

Well my point more that you haven't "completed" it as such, because there are valid historical matches you don't have in your list (and nor does anyone else except perhaps a couple of researchers such as those on this forum). Granted, in football the percentage of unknowns is smaller than in other sports, but this is precisely why it is an important consideration before choosing a new sport. There is no point generating historical rankings of, say, Uganda in field hockey if you don't have any of the results of East African hockey competitions. In that case, if you aim was to generate reliable hockey rankings worldwide, you would need to first reduce the data gap.

this is partially inaccurate, because this is not how mathematics works. i understand what you're saying to a certain point, in extreme cases... and in many other cases it may or may not negatively impact the rankings. however, we are dealing with international sports here... and due to the few number of games the rankings are automatically going to be somewhat inaccurate in most cases, which is why i refer to them as estimates. hell, in some cases with the current international situaiton accidentally missing a match will actually make the ratings more accurate in regards to reality by complete luck. but in a scenario where you have 100 games for a single team... chances are missing 10 of these matches randomly aren't going to have a negative impact 9 times out of 10.

Mind you, most rankings don't incorporate home advantage or winning margin, and are hence inadequate anyway.

i understand this sentiment, but it is a tricky topic. home field advantage is different in every sport, and is also nearly or virtually completely removed in leagues with proper management and rules. but this is a whole other topic.

Last Edited By: abramjones Jul 13 17 6:03 AM. Edited 1 time.

Quote    Reply   

#18 [url]

Jul 13 17 1:32 PM

It is part of the same topic - adding additional factors into a ranking system, and extending it to other sports. Home advantage is one such factor, a major one. Missing data is a major factor in certain sports, for certain countries. It's not a case of "random" holes in the data, it's a case of blind spots meaning that certain countries or regions (or certain eras) have much sparser data than others.

Quote    Reply   
Remove this ad
Add Reply

Quick Reply

bbcode help