[SUGGESTION] Anti sore loser feature?

Hi "Dominick"; ("55% win percentage") who I beat with 420 ping on HUB Frankfurt tonight 7 to 5 who quit with 4 seconds on the clock...!!!

Why not record whatever the most recent match statistics were, despite sore losers rage quitting in the last few seconds?

My anti sore loser technique is to play another match against such a sore loser and if it happens again then I don't play that person any more. One time might always be a connection problem, even if it is very unlikely.

(Same goes by the way for players that don't appreciate a game they lost and start complaining and making up excuses why they had lost)

I agree with Halagone, but an interesting feature would be to simply not record X amount of winning matches for people who use such behavior to neutralize their stat manipulation. Of course such thing should only be active after a few such matches, since there are legitimate reasons to leave like crashing etc.

It's a shame that people quit against people they're losing, because when you have those games recorded in stats you can estimate ones skill based on who they lost against. Assuming the opponent is someone somebody is familiar with. The point of the statistics is not gaining highscore, it's to evaluate and estimate skill which a lot of cups/pickups depend on.

I don't think there is a need to adapt Xonstat to people that can't lose. Those people should not be the target group of Xonotic - I mean, dealing with losing a game is a skill you should acquire at ages 3 to 6 (latest), and I don't think Xonotic is suitable for that age group anyway.

I don't seriously believe in curtailing people's right to take any action that is within the rules of the game no matter how irritating. It would be foolish to try to police every bad behaviour in practice with coded rules. However, this kind of behaviour is rife to the point it negatively impacts the play experience.

As a social species shaming or fear of shaming is an effective tactic to regulate anti-social behaviour. Going back to the OP I propose that recording and showing a player's stats at the last point they were seen in a game as valid and contributing - along with the time stamp that they left (quit/spec'd etc) and keeping a counter on their profile of the number of games they did not complete will act as a permanent stain on those guys who never learned in kindergarten anything about good sportsmanship.

If the data is publicly available, then people can be informed about choosing these kinds of pointless opponents.

I guess what I'm suggesting is a basic etiquette/reputation statistic.

I don't think there is a way to reliably determine by algorithm if the other player left due to fear of elo loss. There are a number of reasons why a player would leave a game early - be it that there are connection problems, hardware/software problems, someone at the door, didn't play any serious duel anyway and was just chatting along with the other player - or even because it has not been a duel but a one-to-one tutorial that is now finished. I wouldn't want any of these cases to be "penalized" by some sort of "bad behavior counter" in Xonstat. This would have a much more negative impact than a positive benefit: people that care for such a social stats element would try to avoid such situations, effectively this would impose a socially engineered restriction on personal freedom.

I believe people that are bad losers will just ruin their online reputation, which makes their stay within the Xonotic community rather unpleasant and, in the long-run, temporary. I don't think this is a real problem that requires a solution.

Also note that it is taking stats too seriously what makes those sore losers quit before end of the match. Now I can see two reasons why that upsets the other player: 1) not getting the deserved elo, or 2) the other player being so annoyingly immature. Both reasons can be easily resolved: for 1) you only have to realize that you're actually taking stats equally serious as the sore loser, which is quite an embarrassing discovery; and for 2) you only have to realize that you're getting angry at the other player because he has exposed quite an embarrassing weakness - which is kinda pointless.

When such system brings justice:
- Players legitimately quit to avoid elo loss.

When such system brings injustice:
- Crashes.
- Connection problems.
- Blackouts.
- Kicked/banned.
- Spectating because teams are uneven.
- Leaving an ongoing game with an idle opponent(s).
- Joining an ongoing game and fumbling random stuff while assuming it's still warmup time.
- Mommy yells dinner and you have to obey with every fiber of your body.
- You have to run to work/school.

Ways around it:
- vcall restart?
- switching teams?

I know that Antibody is looking at different rating systems other than ELO.

Wonder if this is something that can be mitigated against somewhat, by using a better rating system?

IMO the fairest way to deal with this is that regardless of how long you play in a specific game, your stats should reflect how you did in that portion of the game, whether you played to the game end or not. Issue is that this would require a fair bit of code work, and the stats detail currently feeding into Xonstat may not be capable of providing all that info.

This isn't a competitive ESPORTS game.

IMO the fairest way to deal with this is that regardless of how long you play in a specific game, your stats should reflect how you did in that portion of the game, whether you played to the game end or not.

Does the quitter automatically lose and if so then what's the point of evaluating the performance from those games again? Let's say you're leading by 40 frags and you leave with 1 minute left. Is it justified to mark the remaining player as winner? Let's say you frag your opponent 5 times and leave with 9 minutes left. How would the stats reflect how well you or your opponent played? Even if you could come up with some logical algorithm to illustrate performance in numbers, there's still a lot of things that you can't factor in. Such as favorable spawns or how the common frags per minute ratio varies between maps/players. It's just pointless to calculate ELO from unfinished games, because a lot can happen within any amount of time left during which a player decided to ditch.

Sure it's a bummer if your opponent quits to prevent ELO loss, but sometimes people quit simply because they're not match for you. That's understandable and I don't think you need to be punished for coming across an opponent that you have no chances or willpower against in the first place. My point earlier was that there's more legitimate/understandable reasons to quit than dickheads who manipulate stats to warrant this project. And like Halogene said, there's no point doing that even because this community will notice and you have nothing to gain from it. I've seen people manipulating their accuracy stats by only finishing games where their accuracy is high. It doesn't matter because everyone knows they're shite and the average accuracy only sums up last 30 or so games anyway.

You gais! I'm not 100% sure where this idea of some anti-douche system has formulated and taken over. Let's get back to my main beef here;
I'm really not proposing an algorithmic "fix" to bad behaviour (see previous post where I detailed not sitting on people's freedoms) my main point was more along the lines of avoiding other player's suffering as a result of tantrums.

The reason I feel burned by this is not stat-whoring - 300 would be my average game's ping. I frequently join games already in progress or have to wait 5 minutes for a map to download over bad connection. So I'd say my ELO would be artificially suppressed and pointless to try to boost anyway. Smiley; I recall playing you before at 300 and obviously losing miserably and gone several games before a more suitable opponent shows up eh! I'm thrilled by the couple of kills I might get, but I do also like to review game statistics to compare accuracy/damage and so on, and when somebody RQ's with 4 seconds on the clock (what else but spiteful elo whoring can this be honestly?) it robs me of the opportunity as these records are not kept. Rude as, bro!

How's this for a solution: Simply record a list of timestamps for each player at join and quit points, which can be shown in stats. Record all games that had > 1 human participant regardless of whether the match completed with all players or not. Now flamDaddy can get his numerical fix, and it becomes quite apparent who the whores are, with no imposition on said whore's freedoms! Then it's up to the community to identify and call out the sore losers if they so wish, help explain weird results, or people can ignore it.

Problem?

Anyone who quits right before the match ends probably isn't worth comparing scores with, and especially not worth the extensive coding needed for such a feature.

This issue comes up a lot. I think Halogene expresses my position on this issue the best: it should be handled socially, not programmatically. Don't play the person who exhibits bad behavior and the problem will correct itself.

Tracking losses or quits within stats brings with it a whole slew of issues, some of which Smile mentions. People quit/leave do so for a variety of reasons and it is difficult or impossible to account for all scenarios within the code.

<facetious>
OK I get it I think..... So the official stance is that I shouldn't get myself into these situations in the first place? I believe the Mayor of cologne said it best with her advice to "keep rapists at arm's length".

Silly flam, playing matches with nicks I haven't seen before (70% odd of games) and implicitly trusting them!

Should I only play people who are 100% genuine Microsoft certified I've played a dozen times++? If so, it seems like an intractable chicken and egg problem arises. I.e. most of the time I haven't played somebody before and have no way to to tell if they are gunna bail or not. And we're back to suffering many many matches and keeping a monitor full of post-it notes on who the fraggots are.
</facetious>

Yes, I realise the fact that it's impractical also to expect to go check out a player's stats page every time before the match starts, but I thought the plan was to integrate stats more heavily within the game?

To clarify again; I'm not suggesting programmatically controlling people's behaviour. I never suggested that. I'm all for the social aspect. And, as I said before, if there is some kind of metric available that players could use to guide their decision or not, well, "the more you know".

I still think such a "social" stats element would, if based upon people not playing a match to the end, be likely to show wrong information rather than correct information (in relation to what it is intended to show) since it will be hard to determine by algorithm when someone leaves to prevent elo loss and you'll likely see other information than you would want to see.

Only thing that I can think of would be some sort of reputation system, but even that would easily be subject to abuse.

I can understand your frustration but I think the origin of your problem is that the individual game has a much higher value to you due to the low number of quality games you are able to play (location dependant). The whole story would become much less of an issue if you were able to play more games. Imho it would therefore be much more rewarding (both for you and Xonotic in general) to spend efforts on increasing the player base at your location rather than introducing a new complex feature with questionable benefits. Also, this would be a task where you yourself could help out with as opposed to ask developers to implement something.

Silly flam, playing matches with nicks I haven't seen before (70% odd of games) and implicitly trusting them!

This is not a triple-A game, more like a deserted island with a select group of people that try their hardest to repopulate humanity. If you play and socialize with the community actively, you'll know what's up.

This is not a triple-A game, more like a deserted island with a select group of people that try their hardest to repopulate humanity.

[off topic]Suddenly I have the feeling we really need more female players.[/off topic]

Only thing I could see to do this properly would be having this info as a separate stat table:

ie: One stat, same as current, with a second stat that reflected all played time.

Problem is, this would involve collecting a lot more data, and you have to keep it separate from the existing data, so the amount collected would most likely double (or more). Then you'd need to modify xonstats to present some sort of figure off this data that would give you some idea.

All that said, perhaps maybe we could just capture a games participated in vs games finished stat (we already have this second part). You'd have to have at least 1 frag (or flag cap) in the game to be considered as participating.

This could be displayed as a x over y with a ratio (the same way kill ratio stat is) on the players page, and would simply be advisory.

ie: ignore it for the most part, but if you play with someone and they quit without explanation right near the end of the game, you can see if this is a running trend, or if it's a fluke.

FWIW: I know I'm somewhat guilty of quitting before end of game, but I at least try and tell people why. I can get support calls all times of the day and am expected to answer my mobile when one happens. That said, if I'm really unbalancing a game, I'll often spec if I can't swap teams to try and even things up. There's other reasons too, so I understand fully that there's a multitude of reasons for quitting mid game.

Note: I'm not tied to any of my above suggestions BTW, I'm just speculating on options. If something good comes out of all that, then I'm all for it.

Halogene - you're absolutely right; I do feel a disproportionately high emotional attachment in a match where I've finally found a challenging but fairly evenly matched (especially duel) because this is a rare occurrence in my life and the feeling of being cheated of simple pleasures or some form of achievement/recognition for doing well in the face of adversity in the form of a permanent record that most players take for granted is immensely frustrating. *goes to turn on auto-record demos again and publish all RQ wins to youtube* (I don't realistically have time for that)

As for the suggestion to invest more effort into popularising the game; FYI I have personally and enthusiastically proselytised the benefits of joining the church of Xonu to literally every single person I know and have met for the past 4 years. And believe me, in my line of work suggesting a game is highly relevant as my clients are all late teens early 20's game obsessed wanna-be game developers. I spread the word to probably 30-50 of these hot targets per year. Very few actively attempt to play, most leave after a single try, 2-3 quit once they realise that either there is a steep learning curve or there's simply nobody but me to play against without ping. I am all ears to suggestions how I can contribute to getting more players beyond this, but after all this time I'm starting to think that actually the game appeals to a more mature player than this mine craft generation can produce. Short of handing out pamphlets at the train station I'm at a loss what might help.

Smile; I wonder what your and other's opinions are about this re-population effort? The age-old "how to win friends and retain newbies" question...
I do actively participate within the community, often as a lurker. Many names I recognise from stats or the forum. But my 70% unknowns still stands. A lot of random newbs are from QL or similar and have skills. So I can have a difficult, even losing match at ping and still be unaware of their personality or reputation. Not to mention smurfs! If you're suggesting I should only play brand names I know and trust - well I simply wouldn't get a decent game in a month and I'd never win competing against the top 50 in duel. (Oh, now I'm #52. Guess I'll never get to #13 again...)

I realise the dangers of displaying an aggregate as representative of the player due to the myriad reasons they could spec/quit etc. Perhaps the best suggestion I can give here is to simply aggregate an IN/OUT stat which displays an average percentage match time that they joined and quit all games. Essentially this would over time show 0% in 100% out for those who only play full games. A higher in value would mean those who tend to join a match part way through. A much lower out % might suggest they tend to do as Cefiar does and quit half way to rebalance a team, but if you see a 95%+ in the out stat it may suggest they are habitually leaving matches near crunch time. Again, since the in/out would be recorded/shown for all matches, you could investigate an individual's behaviour by examining what happened during a specific match record.

Alternatively I do like Cefiar's suggestion that if you simply record a participated/completed ratio then it will give a good rough picture to the flakers who quit a lot for whatever reasons legit or not, giving a kind of reliability rating.

Of course, it's up to the individual to decide what to do with these numbers in the end but I think behaviour would be captured within those bounds. It would be extremely hard to explain to your opponent a 99% out stat or a very low p:c ratio. And of course, as I originally suggested, display the in and out times during the match so you can examine what actually happened on a match per match basis.

I'll stop talking now.

I am all ears to suggestions how I can contribute to getting more players beyond this, but after all this time I'm starting to think that actually the game appeals to a more mature player than this mine craft generation can produce. Short of handing out pamphlets at the train station I'm at a loss what might help.

Well, it helps to be honest about Xonotic. If you present Xonotic as a "fast paced game" you should also say it's only fast paced if you're fast. And if you present it as a "shooter" you might also want to elaborate that it's not really as much of an aim game as CSGO/Halo/TF2 is, you should explain what item control is and how it affects the gameplay. Xonotic is also not very popular game and there's a specific time/location that more people are often playing the game simultaneously. As unappealing as you might make it sound, intricate details wins false expectations imo.

I'll get to work on the pamphlets then...
Thanks for the suggestions.

I wasn't aware that you actually did try apparently all within your scope of possibilities to establish a certain player base at your location - I wouldn't know what else to do. I'm sorry to read that it didn't succeed, which makes your frustration very well comprehensible. However, I still don't think a participated/completed games record would provide the information you want to see (due to diversity of reasons to quit in the middle of a game) while having a negative impact on people that strife towards a best possible stats status. And wherever there are stats, there are also people that want to reach the top. I strongly oppose any system that would impact on people's match participation behavior for the sole reason of stats optimization, meaning that they feel the need to stay in a match even if there is no reason to do so besides stats record.

flimflam's church of Xonu; population: ~2. Guess that makes me the 1% that stuck around, huh?

flimflam's church of Xonu; population: ~2. Guess that makes me the 1% that stuck around, huh?

Well since your last game was a month ago I have to warn you that your membership card will self-destruct soon...

However; if your development efforts count as an indirect contribution then I'm already absolved of pamphlet peddling responsibility!
(Full disclose: Yes, I know you were already playing before we met )