APCR (ASL Player Comparative Ratings)

Aaron Cleavin

Elder Member
Joined
Sep 27, 2004
Messages
3,095
Reaction score
573
Location
Sydney
Country
llAustralia
uhm, that isn't correct :)
I have 152 games for an 1864 rating with area rating
http://www.asl-area.org/Parcours/SSB.html

Perhaps a name issue ?
Could be the same with Paul Sidhu and others. Perhaps an open record of games (like Area) will enable players to review their own results ...
The majority of your game are invalid for APCR as they have no scenario code, they are part of the 10% that are not processible

If you can send me the games with a scenario code then I can process
 

Aaron Cleavin

Elder Member
Joined
Sep 27, 2004
Messages
3,095
Reaction score
573
Location
Sydney
Country
llAustralia
Aaron,

First off, I would just like to say thank you for the tremendous amount of time and effort that you have put into creating this wonderful updated rating capability. I’m envious of the skills you have!

Relative to my case of no recorded AREA inputs since 2010, I just checked the old AREA sight and saw and additional 37 tournament entries, mostly for Bitter Ender, ASLOK, and Albany.

Are you not seeing these entries?

Additionally, I have all of my games recorded for ASLOK, Winter Offensive, Albany, and Bitter Ender for the last two years (all except one victory over Stan Jackson from last year’s Bitter Ender where I don’t have the scenario name). Will you accept these entries direct from me?

Thanks again!

Bill
Yep happy to accept direct.
It is not unlikely many of those 37 didn't have (any or good) scenario codes on them, if you have then we can process
 

Aaron Cleavin

Elder Member
Joined
Sep 27, 2004
Messages
3,095
Reaction score
573
Location
Sydney
Country
llAustralia
Aaron,

First off, I would just like to say thank you for the tremendous amount of time and effort that you have put into creating this wonderful updated rating capability. I’m envious of the skills you have!

Relative to my case of no recorded AREA inputs since 2010, I just checked the old AREA sight and saw and additional 37 tournament entries, mostly for Bitter Ender, ASLOK, and Albany.

Are you not seeing these entries?

Additionally, I have all of my games recorded for ASLOK, Winter Offensive, Albany, and Bitter Ender for the last two years (all except one victory over Stan Jackson from last year’s Bitter Ender where I don’t have the scenario name). Will you accept these entries direct from me?

Thanks again!

Bill
Some of the Data for later years is in part of my staging but hasn't made it thru next filter when in fact it is correct, doing some root cause analysis now.
Bitter Ender 2011-16 had a problem the the tournie date not be set to the right year which meant those game were calculated for the wrong period/year , easily fixed :-(
Bitter Ender 2017 have scenario code but no winning side losing side information so again not processible until I have such info. (APCR is absolute on needing a scenario and a winning side)
 
Last edited:

Aaron Cleavin

Elder Member
Joined
Sep 27, 2004
Messages
3,095
Reaction score
573
Location
Sydney
Country
llAustralia
All following items are next up

  1. Quick rebuild of game data to account for errors in a seed players data, and dates on Bitter Ender 12-16
  2. Processing Data Fro Stockholm 17 and 18 and sending the outputs to Melvin
  3. Processing Chicago Open 2019 data
  4. processing some individual data for Andy Hershey
  5. Giving Enrico a personal Game report
Processing lots of other data coming my way by the time I finish these.

Next update in 24 hours or so
 
Reactions: ecz

Justiciar

Elder Member
Joined
Mar 2, 2008
Messages
4,808
Reaction score
1,379
Location
Within Range
Country
llUnited States
For the record, and in the open, on point 4. One point to process. I submitted a correction of a win to a loss vs. Joe Celebuski during ASLOK 2017. I was using a copy / paste format and while having corrected for all else in the form I failed to adjust the win / loss part of the equation. We played back to back games in 2017.
 

Aaron Cleavin

Elder Member
Joined
Sep 27, 2004
Messages
3,095
Reaction score
573
Location
Sydney
Country
llAustralia
Do you have all years for St Louis? I can get those for sure. You have Kansas City?
I have everything in AREA but games without a scenario code or/and without a winner cannot be processed. More than happy to take updates :)
 

zgrose

Senior Member
Joined
Jun 13, 2004
Messages
3,428
Reaction score
448
Location
Kingwood, TX
First name
Zoltan
Country
llUnited States
AFAIK there are two sites claiming to AREA, so people might want to be super specific about which site they are referencing to help him debug/troubleshoot.
 

Aaron Cleavin

Elder Member
Joined
Sep 27, 2004
Messages
3,095
Reaction score
573
Location
Sydney
Country
llAustralia
AFAIK there are two sites claiming to AREA, so people might want to be super specific about which site they are referencing to help him debug/troubleshoot.
AREA as run by Bruno (I don't know anyone made any substantial ASL contributions to the other AREA in the last two decades (though I could be wring)
 

bo_siemsen

Member
Joined
Mar 11, 2004
Messages
338
Reaction score
98
Location
Copenhagen, Denmark
Country
llDenmark
The majority of your game are invalid for APCR as they have no scenario code, they are part of the 10% that are not processible

If you can send me the games with a scenario code then I can process

Thanks for looking into it

There's no chance I can dig up scenario codes for 15 years worth of going to tournaments. This, to me, looks like a significant flaw in your system. Looking at other players (like fellow dane Michael Hastrup-Leth) about half his games has disappeared in the APCR rankings. I suspect that it is the same issue. As AREA never required that information it was not gathered. At least not on a regular basis. It may be standard in (some) US tournaments and VASL tournaments but not everywhere.

This is obviously the explanation for the demise of Paul Sidhu in the rankings. He's won the Albany Tournament multiple times ... but that doesn't count. And my proudest moments as an ASL player don't count either.

I would really encourage you to find a solution to that issue. I would suggest that you consider a way to input a "default" code for such results that allow for missing values in Scenario codes and such. Otherwise you lose way too much data and that would render the rankings misleading I am afraid.
 

Mister T

Elder Member
Joined
Aug 12, 2006
Messages
3,675
Reaction score
999
Location
Bruxelles
Country
llFrance
From my archives, I've just provided scenario codes for almost 30% of my games that did not survive the changeover from AREA to APCR, provided i've grasped the underlying reason (that may explain Ecz's "paradox" :)). I have no info for a few scenarios that seemingly will remain "casualties".
I feel empathy with Bo's remark.
 

bo_siemsen

Member
Joined
Mar 11, 2004
Messages
338
Reaction score
98
Location
Copenhagen, Denmark
Country
llDenmark
From my archives, I've just provided scenario codes for almost 30% of my games that did not survive the changeover from AREA to APCR, provided i've grasped the underlying reason (that may explain Ecz's "paradox" :)). I have no info for a few scenarios that seemingly will remain "casualties".
I feel empathy with Bo's remark.
:)
I am not really sure a ranking can give a fair representation of players abilities if it relies so heavily on which tournaments you attend as well as individual score keeping. I am not just thinking from my own perspective, but also as a Tournament Director. Especially since 21 years of ASO scenarios aren't included in this new list (except for those who have their own records of their games AND have discovered this thread). It just doesn't seem viable to me.

I am not knocking Aarons' efforts. It is much needed to get something we can use for the tournament scene.
 

Aaron Cleavin

Elder Member
Joined
Sep 27, 2004
Messages
3,095
Reaction score
573
Location
Sydney
Country
llAustralia
:)
I am not really sure a ranking can give a fair representation of players abilities if it relies so heavily on which tournaments you attend as well as individual score keeping. I am not just thinking from my own perspective, but also as a Tournament Director. Especially since 21 years of ASO scenarios aren't included in this new list (except for those who have their own records of their games AND have discovered this thread). It just doesn't seem viable to me.

I am not knocking Aarons' efforts. It is much needed to get something we can use for the tournament scene.
There's no chance I can dig up scenario codes for 15 years worth of going to tournaments. This, to me, looks like a significant flaw in your system. Looking at other players (like fellow dane Michael Hastrup-Leth) about half his games has disappeared in the APCR rankings. I suspect that it is the same issue. As AREA never required that information it was not gathered. At least not on a regular basis. It may be standard in (some) US tournaments and VASL tournaments but not everywhere.

GIGO. (Garbage in Garbage out), if I have really bad data the outputs of the system will be considerably compromised
Requiring correct data is definitely a feature rather than a flaw

What can be done is that any game from January 1 2019 will need winning side and Scenario Code (No Exceptions) but that I will do a one off population of Area Data with no winning side or scenario code.

I will put things in with a scenario code "Dummy" and Winning Side "Dummy" where needed. The problem with this kind of ugly workaround is that it requires all sort of kludge on summary statistics.
I will need to make all games recorded with a dummy side be excluded form any winning %'s for attackers/defenders. For the cases where a scenario code has been recorded but no winning side this sort
of issue becomes even uglier.

Currently 18456 AREA games are in roar
I think about ~3500 are not in (Will update with a more exact number shortly)

Of the bad records some are more smelly than others

Ones with PT (Playtest) in the scenario code, should these be included or excluded, or included for peoples records but not for statistics, really should peoples ratings be changed based on results of unpublished scenarios. My preference
is not to included (Supporting Fire is obviously different as the scenarios are pre-pub in this case)

Where one players name is like bob XXXXX should we include or not (This is a small number of records 10-15 and I am strongly inclined not to include).
 
Last edited:

zgrose

Senior Member
Joined
Jun 13, 2004
Messages
3,428
Reaction score
448
Location
Kingwood, TX
First name
Zoltan
Country
llUnited States
Out of curiosity, how does the scenario identify affect the rating value? I can understand player identity and date being critical to the score but I’d imagine the rating is agnostic to the scenario played.
 
Reactions: ecz

Aaron Cleavin

Elder Member
Joined
Sep 27, 2004
Messages
3,095
Reaction score
573
Location
Sydney
Country
llAustralia
The rating is agnostic but the tracking of player perfromance by scenario type, by attacker/defender, by nationality played are all compromised, as are stats around scenario/designer performance balance. It is possible to exclude dummy scenarios but coding to do so in the right way at different places is somewhat painful
 

Aaron Cleavin

Elder Member
Joined
Sep 27, 2004
Messages
3,095
Reaction score
573
Location
Sydney
Country
llAustralia
All following items are next up

  1. Quick rebuild of game data to account for errors in a seed players data, and dates on Bitter Ender 12-16
  2. Processing Data Fro Stockholm 17 and 18 and sending the outputs to Melvin
  3. Processing Chicago Open 2019 data
  4. processing some individual data for Andy Hershey
  5. Giving Enrico a personal Game report
Processing lots of other data coming my way by the time I finish these.

Next update in 24 hours or so
1. Is complete

I am switching to including partial AREA results (due to popular demand) which I should have done today.

Also will respond to an amazing idea from Mr T. of composing a number which combines rating and RD (Rating - 190 + RD is proposed) all three numbers will be displayed
but this one may be given preference. 190 is halfway between 30 min and 350 max RD
 
Last edited:

Michael R

Elder Member
Staff member
Moderator
Silver Supporting Member
Joined
Feb 4, 2003
Messages
2,628
Reaction score
763
Location
Montreal, Canada
Country
llCanada
The rating is agnostic but the tracking of player perfromance by scenario type, by attacker/defender, by nationality played are all compromised, as are stats around scenario/designer performance balance. It is possible to exclude dummy scenarios but coding to do so in the right way at different places is somewhat painful
Think about entering the DUMMY scenario codes and sides in a way such that you do not need to exclude them from processing. Maybe the DUMMY entry shows up on a report and people ignore it, or you exclude it only there.
 

Aaron Cleavin

Elder Member
Joined
Sep 27, 2004
Messages
3,095
Reaction score
573
Location
Sydney
Country
llAustralia
Think about entering the DUMMY scenario codes and sides in a way such that you do not need to exclude them from processing. Maybe the DUMMY entry shows up on a report and people ignore it, or you exclude it only there.
Not as easy as that they need to be included in Ratings processing but excluded from other processing, the tough thing is when people entered scenario codes but not sides intoAREA, in the end these also need to be excluded
from Scenario stats as without knowing the side impossible to understand the effect of the playing on the scenarios record, and coding in such a way as to include in number of times a scenario was played but excluded from
Stats as it is neither a draw nor a w/L but an undefined to add to the number of times a scenario was played but not to its performance record is just too ugly, error prone, and hard to code .
 
Last edited:
Top