It's also important to consider processes and workflows.
Oh, very true!
I believe the most important and basic question to decide first is whether the new AREA will be of the "open" type or of the "closed" type akin to the old AREA.
Each approach has its advantages and disadvantages:
Let's start with the "open AREA" concept.
Advantages of "open AREA":
- Players would create their personal account
- Players would (well, should...) be able to type their name correctly.
- Players would enter their own games
- Workload would be distributed
- As individual players decide what to enter nor not to enter, the problems with data protection should be overseeable.
- This concept could cover not only tournament games but basically any game.
Disadvantages of "open AREA":
- It will be patchwork as experience shows that only a few participants of tournaments would enter data. There would be no tournament which is entirely covered by "open AREA".
- There is no control about the quality or integrity of the data entered. Players might type their own name correctly but not so that of their opponent, maybe they don't even know it or forget it (see all these "unknown paddy" entries in ROAR...).
- Fake entries (to improve one's rating), solitaire games, and all such stuff would find their way into AREA.
- Due to the inferior quality, integrity, and patchwork nature of an "open" concept, it would defeat the original purpose of AREA as a reliable source of player ratings and a tool to cover tournaments in its entirety.
Now to the "closed AREA" concept, i.e. the one that has been used by the old AREA and which I would favor.
Advantages of "closed AREA":
- Tournaments can be covered in their entirety which is impossible (at least very unrealistic) in an "open" concept.
- Tournament directors, by definition, know the outcomes of all relevant games, they know the names of the participants, they know the pairings, they know the scenario names and IDs. This is an invaluable advantage for data collection.
- If only TDs can enter results, then there will be no "fake" games in the database.
- The value of the "closed AREA" approach has been proven in practice in the past. After all, this is why now people think about reviving AREA.
- As part of their "terms of use", tournament directors can require the consent of data on game results to be collected, solving some issues there.
- Due to the data that would be fed to a "closed AREA" being of better quality, the ratings would be more meaningful. Without wanting to dispraise the efforts of the great website "ASL Scenario Archive" I suggest to compare the "open" ranking list there ( https://www.aslscenarioarchive.com/allplayers.php ) with the "closed" old AREA list defunct for a couple of years here ( http://www.asl-area.org/Area.html ). Looking at the player's names and their respective rank, everyone can judge for themselves, which of the two rankings is more meaningful. Both systems do not use the same algorithm I think, so to some extent it might be understandable that the ranking differs. Yet for one and the same player, the relative rank might differ a lot which will mostly be caused by the disadvantages of an "open AREA" concept.
Disadvantages of "closed AREA":
- The actual work of entering data will rest on less shoulders, namely the TD's maybe augmented by a couple of volunteers that enter data provided to them.
- A "closed" concept will require a system that administers access rights, passwords, and someone who determines who will be granted access.
- Not being among those that have access may touch the sensitivities of some.
- Bruno Nitrosso told me one of the main problems he had while maintaining AREA and entering the data was that despite receiving it from TD's only (and thus already having a much better basis than otherwise), it was often slips of paper with unreadable handwriting, incomplete names, and missing info. These slips were those that TD's often ask players to fill out and hand in during a tournament. On site, it might still be easy enough to find out who has won, which scenario it was, etc. But afterwards, things get difficult and may be tedious. This work would rest mainly on the TD's and volunteers. The prime responsibility would rest on TD's or volunteers present at the respective tournaments to collect quality data right from the beginning. Bullshit in, bullshit out. This disadvantage is much greater in an "open" concept, naturally. Yet, I also list it among the disadvantages here, because some TD's might not like this sort of work.
We want to avoid the previous situation where one person was doing all the work, and eventually gives up, so the work has to be distributed over multiple people. This means user accounts and permissions.
Let's say Joe is TD for the ASL World Championship 2021. He obviously needs to be able to enter results, but what if there are players who are not currently in the system. Should he be able to create new player records for those people?
Yes, he should be able to create new player records. And he has to be very precise doing it to avoid misspelled names which would result in double accounts etc. A real problem that Bruno Nitrosso had to deal with which was tedious.
And what about the tournament itself? Should he be able to create a new tournament within the system? Or do we have super-users who can do that, then give Joe permissions to update it?
Not sure about this. But I am convinced that there would be the need for some sort of super-user / super-administrator.
Or perhaps we have the concept of a tournament series i.e. a super-user can create the "ASL World Championship" series, and Joe is allowed to create the "2021" edition, "2022" edition, etc.
I think that a TD of a tournament should be able to create "editions" of this (yearly?) event without assistance by a super-admin.
What happens if Joe doesn't enter the results? Is another TD allowed to do it, or does it have to be done by a super-user? Data entry is always messy and error-prone, so do we have a process that requires someone else to check the data before it goes live?
For big events such as ASLOK, I believe a team of people should be assigned to the job and be granted permission to enter the data. As said before, the main responsibility for entering correct and complete data would rest with the TD's. However, these are by far the best suited to do so as they have most of the correct information to begin with (names, scenarios, pairings, etc.) as they run their tournament. They can also minimize the amount of errors, if they hand out slips of papers that hold most of the info to begin with and leave the players only the task to file in sides, winner, balance, and fun rating (whereas the latter two would not be required for AREA but rather for ROAR). The TD's and their on-site volunteers are the key-persons to make the system run smoothly.
Allowing player records to be created by someone other than the player themself also opens another can of worms, because will the record contain information the player might want to change? They then need to have an account themself, and some way of connecting their account to the already-created player record. This last bit would need to be done by a super-user (to avoid abuse) i.e. a manual process :-( You could have the person send a request to the sysadmin to have the information changed, but that's more manual work.
This can of worms can be kept tightly closed if the 'terms of participation' in a tournament could require the consent to adding the player names and results to the AREA database. If some people do not want to have their real name displayed, then this real name might only be visible on the back-end while a nickname or number appears on the front end that would be visible to anyone visting the site. Anyone who has access to the back-end would need to be able to see the real names, though, I
think.
Someone else mentioned GDPR, and this is a real consideration. I don't know much about this, but my understanding is that people have the right to request their personal data be deleted. Not made private, deleted. Are game results considered to be personal data? I don't know, talk to a lawyer. Name and country? Surely. What effect does that on the system overall, not being able to have any information on a player at all?
Within certain bounds, the GDPR grants people the right to be informed about the data that is stored about them, to have it corrected or deleted. I am no lawyer in this field, so I cannot fathom the implications in every detail. Let us think the following example: Player A has 100 games logged into AREA, the results of course affect his own rating and that of others. Would it be sufficient if his name will be terminally deleted on the front- and back-end but some number or standard entry remains, so that the ratings of other players remain unchanged by Player A dropping out? Or would everything have to be deleted which would also effect the ranking of other players? Not sure. If one wanted to be on the safe side, probably the latter option. However, I think the following would not work nor could be legally enforced by Player A: First sign up for the tournament, give your consent for AREA data collection and upon the conclusion of the tournament retract the consent and demand everything deleted (i.e. what Magnus did as mentioned in his post up-thread). After all you can't sign a contract and afterwards say that you are not bound to it.
It might actually be a good thing, since it will force the design to handle it. The current AREA site has the concept of a private player, but I think that just means they don't appear in the list of players. However, their game results (including their name) still appear on the tournament pages; IOW, not really private, just somewhat obfuscated.
Maybe one could program it in a way that the entry "opted out" appears on the front end for
every player that did so, while at the back end some anonymized number remains or does not remain. Or one could program it that no AREA rating is changed by the effects of people opting out. Good question.
So OK, a player Fred competed in this tournament and had a player record created for him. He then asks for it to be removed, so we now know him as player #123, with no identifying personal information whatsoever, just his game results. But when Fred plays at the European Championships, how does that TD know that Fred is player #123, and he should use that player record, instead of creating a new one?
If Fred just wants his personal name removed for the common public, this case could be caught if he is at least fine that his name plus attached ID# or nickname can be seen in the back-end by TD's only.
But one problem might be that Fred, being very ambitious but not very skilled, starts with his first record and a rating of 1000. Then he loses a couple of games, changing his score to 900. Being an arse, he determines that it is 'best' to demand his record deleted with all information completely and then at the next tournament claims he never had one, thus resetting his score to 1000, i.e. cheating himself up by 100 points. Probably, this could not be prevented. But on the other hand, one might just put up with this because few people would to this and by this means, he could never raise above the average entry score of 1000.
Or even worse, Fred tells the TD up-front that he doesn't want to be in this new AREA. How does the TD enter his game results, assuming Fred's opponents want their results recorded?
Three possible ways to address this.
One would be without consent to AREA, no admission for the tournament. Sounds rash, but seems to be the standard operating procedure in business right, left, and center.
The second way would be to assign anyone that does not want his results recorded a standard rating which is counted against his opponent, i.e. the standard 1000. This might scew the integrity of ratings though, because if Pleva opts out and is thus assigned the standard 1000 and then beats Fortenberry who did not opt out, Fort's rating will plummet if Pleva wins. So probably this is not a good option.
The third way would be that if one or more players do not want to be rated, then the game is not rated for both.
I think all this points to a need for the system to be able to handle anonymous players e.g. the ability to record a game result like "Mike beat someone". But how can a rating system possibly work if you don't know who one person is?
This is incredibly messy, but it's a people problem, not a technical one. The underlying issue is: what do you do when one person doesn't want to have their game results recorded, but the other one (and TD) does? Decide that, and that will then drive the technical design.
Again, two ways to address this:
If one of the involved parties does not want it, then it does not happen. Just like sex. That is pretty straightforward.
The other way would be to make the consent to AREA a prerequisite to be allowed to sign up and take part in the tournament.
Do we require people consent to have their game results recorded? Would that even fly under GDPR? And while some people have said that this is heavy-handed, I can't imagine there's a single serious competition in the world that lets you compete, but also refuse to have your results recorded. If we want to say that ASL competitions are always friendly, relaxed affairs, then so be it, but some people do take their ratings seriously. What happens when one person really doesn't want to be in the system, but the other person really wants the points? Maybe we have AREA-required and AREA-optional tournaments, but that doesn't really solve the design problem, because we still have to be able to handle the AREA-optional tournaments.
I think required consent would fly under the GDPR. If the 'terms of use' of FaceBook do (after some adaptations), than whatever a TD might want to require with regard to AREA would do, I reckon. Provided that AREA itself is designed in a way that a standard of data protection is preserved.
I never really saw the point of recording who is "attacker" and who is "defender". It's possible that an algorithm might take it into account when calculating a rating, but how would it work if both players are attacking? Or not recorded?
Even more pointless is "Axis" and "Allies". We saw how people can react to this (a fair consideration, IMO), and it's hard to see how a ratings algorithm would need this information, so why bother? Just note for the record that Alice had the Laotian Communist Freedom Fighters, Bob the People's Front of Judea, and be done with it ?
The Attacker/Defender was recorded (I think) because some of the data was also fed to ROAR.
von Marwitz