Abandon ratings

[BaronHelix]

Peoples' abandonment ratings are frequently misleading. In the starting days, game crash was way more common than it is now. It still happens, but with FAR less frequency. As a result, I don't know if people who have abandons are just early adopters or actual rage-quitters. I propose that the current abandon totals be wiped clean so that we may make more informed decisions of who we play with.

[Big Head Zach]

Agreed, with the addition that Training game abandons shouldn't be counted. Sometimes I use those to test things and to take screenshots for the Guide that I'm working on. Obviously, I don't play those games through to completion ;P

[maskdbandt]

splendid idea!  also can there be a way to mark people (for your own knowledge without keeping a separate list of names) that we want to not play with because they have abandoned a game or kicked someone out without giving them a chance to get back?

[Big Head Zach]

I believe that a "Friend / Foe" list that colors names in the lobby has been suggested and is on the list, so tp speak.

[napier]

The real problem with abandoned games is that hosts can abandon games to control their ranking.  If a muler wants to game the planetmule ranking they:
   1) host their own games (get ping advantage)
   2) pull the plug on any game they're not winning

If a host abandons a game, they preserve their ranking, nobody wins, and it's been a major waste of time for three people.  Hosts that are playing the "get to the top of the ranking" game have incentive to do this. 

If a non-host abandons a game, the game still goes on with an AI and one of the other three players will win and improve their rank.  It's annoying but not egregious.

To find if player Joe abandons games that he hosts, calculate the percent of games Joe hosts (games Joe hosted/games Joe played), and the percent of games Joe hosted and abandoned (games Joe hosted and abandoned/games Joe Hosted).  If both percents are high then Joe hosts a lot of games, and Joe abandons a lot of them.  For short we could just say Joe is a "dweeb".

I did a background check on a few top-ten players, and found substantial dweebishness in several cases.  Now I know which hosts will play a good game and which are likely to run away.  It saves me a lot of time.

A Dweeb Rating like this could provide a dis-incentive for gaming the rankings page. 

[piete]

I did a background check on a few top-ten players, and found substantial dweebishness in several cases.  Now I know which hosts will play a good game and which are likely to run away.  It saves me a lot of time.

Hi napier! I don't know how I ended up in your "ranking" or how you conducted your study, but owning one of the highest abandon counts I can justify each one of them if the game chat does not justify it by itself. I haven't quit a single game as a guest or as a host due to "dweebishness".

A recent "trend" with players I play with is that if one player leaves the game for any reason during the first rounds we decide to restart because we don't want to play with ai. And yesterday we had to restart one game that had 3 very experienced players and one beginner, because it was obvious from the first round that the game would not be fun for anybody. But we always agree on the restarts in the chat. But I'm afraid all these situations add to the abandon count.

Fortunately a lot of players know me already so I don't have to prove anything to them. I for myself have reservations for new players with zero-stats (because I don't have any idea if they are going to quit in the middle and I already have a few bad experiences) but I welcome any non-zero stat player to my games, as long as they have abandon rate less than 100 %

[napier]

The good news is that most hosts I looked at play fair games.  There are few noticeable exceptions, and not all in the top ten.  I assume that we all drop out of games for one reason or another.  The average abandon rate across the board is about 30-40% which is probably due partly to bugs and network problems, but also to phone calls, kids throwing up, fires, etc. etc. 

Yeah, I'm having some fun calling it a "dweeb" rating.  Not trying to bust on anybody.  Point is: wherever there's a ranking system there will be people gaming it, and that can screw up the dynamics of the game we came here to play.

BTW your dweeb rating was under 5%     (the average was about 12)

Good muling, dude!

[hightekvagabond]

I'd love to see Dweeb ratings be part of people's profiles, but they should only rate against people's most recent 80% of games so someone could de-dweeb by correcting their behavior.

[napier]

Makes sense.  The prodigal son effect.

[urgrue]

bump.
Dweeb rating would be very useful.
I'm starting to consider not playing against anyone with more than 30% abandons, but a dweeb rating would be much better, because as mentioned abandons may be high for various good reasons.

[Pescado]

In order to get an accurate picture of someone's behavior, a mere abandonment rate number is meaningless. You have to look at the actual games that were abandoned: WHY were they abandoned? Did the players who got abandoned seem bothered by this? For instance, I have a rather high abandonment rate. Out of about 8 games to completion, I have had 3 games go belly up and die. Why did it happen? Well, if you're looking at my stats, I'm playing the same crowd over and over, so clearly it is not because I am a dirty ragequitter. In fact, most games abandon while I am "winning"...not that we are playing for individual wins. Conversely, a "dirty" host can pump everyone ELSE's abandonment rate by selectively disconnecting THEM, then finishing the game against AIs. A certain "Akire" individual is accused of this.

Of course, I maintain based on my analysis of the networking protocol that this is all far more bothersome than it's worth. If you REALLY wanted to cheat for your ranking, you could simply do it by falsifying games completely with a custom program to submit false data to the server. There isn't any authentication mechanism, and indeed, there really can't be, because the program executable is available for dissection on the user's computer (and in Java, no less!), and so is the entire game. With the game itself hosted on the user's computer, it is subject to any modification the user wishes to perform. The architecture is so massively insecure at this point that it is trivial to anyone who really wants to cheat (I could do it, but I just don't care and don't want to make this into an object demonstration just prove what I have said here on the matter), and it is only the relative obscurity of this game that prevents massive, widespread cheating. If this thing were to become popular, at this point, cheating would be rampant.

[Bertoid]

When the game first started and there were so many bugs, people quit games all the time.  I have abandonments on my record from the first few weeks, but never since then.  Wish I could get those scrubbed.

[C64 nostalgia]

Wish I could get those scrubbed.

Reset the rankings completely (Hopefully with a new ranking system, too.).