Map cycle timeout

Server setup, history & future
ankeedo
Posts: 7
Joined: Thu 24. Mar 2022, 18:51

Map cycle timeout

Post by ankeedo »

I suggest reducing map cycle timeout to less than 25 which is too much to wait for a certain map.
User avatar
Miauz55555
Posts: 1819
Joined: Sun 7. Jun 2015, 22:12
Description: https://discord.gg/X4V8THM
Location: Germany

Map cycle timeout

Post by Miauz55555 »

Moved the topic here, as it belongs to the server region.

Thanks for your suggestion ankeedo, so you have any number in mind?
It's that high that not always the same 3 maps are voted. But I agree that it's may a bit high when there are not so much games over the day.
Image
ankeedo
Posts: 7
Joined: Thu 24. Mar 2022, 18:51

Map cycle timeout

Post by ankeedo »

I think 15 is fine.
User avatar
Pegasus
Posts: 1277
Joined: Wed 4. Nov 2009, 23:37
Description: ONSWordFactory
Location: Greece

Map cycle timeout

Post by Pegasus »

Hey folks. Looks like this might be a good opportunity to offer some historical insight into the map repeat limit setting and its relation to larger facets of CEONSS content management. Assuming the following principles still inform the current administration's priorities n' decisions - I've yet to come across evidence that suggest otherwise - they might be of some help in steering this discussion to a productive direction. I'll try to shed some light on the subject both from the ideological as well as the practical angle.


Going back to its earliest days, the philosophy behind CEONSS' roster makeup, also informing repeat limit considerations, was never intended to resemble anything like that of, say, a product manufacturer's or a super market chain's sales department. The server's aim was never to identify the best-selling "products" that attract the most demand, and then proceed to mass-produce (variants of) only those at the expense of any other kind of competition and/or give 'em prime placement at eye-level shelves across the land, while half a dozen others get relegated to the "minor" category, doomed to appear in smaller quantities and less visible places nearby, all in some kinda warped "free market" caricature. Instead, since its founding, CEONSS staff has viewed content management though a much more curative lens, resembling the way that a museum's staff or a family planning their dinner table would approach selecting and iterating on their offerings, i.e. by trying to maintain a diverse, interesting, and "healthy" cultural/nutritional "diet" so that, to drop the analogy and switch to ONS parlance, regardless of whether a player's session would last for 20mins or 2hrs, they'd be able to walk away having had an ingame experience of decent, conceptually- n' thematically-varied, challenging, viable playstyle-pluralistic, balanced, skill-based, and well-moderated gameplay quality across whatever string of maps they might have happened to play through.
This approach is a marked departure from putting most of the content management effort behind tweaking the same 10-15 top maps again n' again - whether by adding ever more edits/variants thereof to the roster or by cramming more n' more custom content in those same variants, hoping some kinda meaningful diversity might arise from that - and it hinges on spreading a similar amount of time n' resources (perhaps more) across many more "fronts" and experimental projects/bets that could well fail. However, IMO there's a good argument to be made about this philosophy being a much more liberating alternative to letting the entire community remain captive to just a handful of design concepts to be replayed n' rehashed ad infinitum. On the community's side, this "liberation" merely requires a shift in perspective, namely from CEONSS players basing their daily session quality assessment on the preconception of being able to have a match in one of the few Big Hits, thus likely becoming frustrated whenever their 2-3 favourite maps happen to be locked behind the repeat limit when they join, to an expectation of a generally less rigid, but overall still well-crafted gameplay experience regardless of specific map sequence, which also affords room for surprise, (new) map discovery and staves off a sense of inescapable, monotonous routine.
In that light, a higher repeat limit is akin to a museum employing a greater number of smaller halls that allow more exhibits to be displayed in a less crowded and more contextually optimized manner for their included collections, thus letting visitors discover and enjoy 'em on their own terms while making their own way through the various halls as they please, compared to, say, just having a few big halls where masses of people would almost always be gathered around the few "top hit" exhibits on the bigger n' better lit pedestals, leaving most other artifacts by the wayside.
In other words, a higher repeat limit value helps reduce the effect of the most frequently-voted maps sucking out all the proverbial oxygen and, instead, affords some crucial breathing room to many other (still vetted and approved) offerings to have a chance at getting discovered, played, and becoming favourite standbys/alternatives to players, particularly during bigger-sized crowd times. There's value to that, both in consensus-building during divisive votes in the short-term, as well as in warding off players gradually quitting the game in the longer-term.
Point here is, this philosophy can be summarized as "cultivate lots of good options instead of a few brand names, and trust your players to discover n' enjoy them", and IMO it has worked well for the server for over a decade now in providing both good gameplay on average, as well as much-needed variety in the ways maps can challenge players, even if it can sometimes produce an incongruous series of map choices.
Since not all servers manage their content with such a mentality, however, it's reasonable to expect that this design goal might not be readily apparent to players coming from other ONS communities, identifying favourite maps only to struggle with playing them as consistently as they'd like, and then joining the msg.board to take aim at the repeat limit as it apparently causing the problem of "too much to wait for a certain map". To that, I think, the appropriate reframing response should be something like the following: "Were the matches in the maps you did play during your sessions consistently good enough to be worth your time here? If so, that's a sign the team's curation n' editing efforts are paying off and working as intended :)." Assuming most regular players' response to this question lean toward the positive, and similar to older CEONSS considerations regarding a sensible average match duration, I'd posit that the current repeat limit should likewise not be seen as some mistake to feel bad about or try to fix, but an achievement to stand by and communicate to other ONS communities.
In case there's skepticism about this possibly being just one former staffer's perspective, btw, allow me to also resort to a practice I very rarely turn to for additional corroboration here by quoting from an internal 2013 document that Heinz drafted as a sort of distillation of hundreds of discussions had during all previous years between staffers (mostly the two of us, I think), and which was meant to serve as CEONSS' formalized mission statement or ideological foundation at some eventual, future point: "[C]ontrary to demo-servers, our server lives from map variety. Of course, everyone got a pile of favorites and some even want to play only those over and over again. To the vast majority on the other side, this would be boring to death. Hence we have both, many different sorts of maps and a rather high repeat limit." Emphasis mine.
To try n' wrap this point up on some kinda constructive piece of advice, when it comes to community members offering feedback on some key, longstanding part of the server's configuration, or indeed staffers receiving and assessing the value of such, it's important to first make sure that such feedback is based on a clear understanding of what goals the server's been trying to achieve through the current arrangement, and that any suggested change is also motivated by achieving - and can be shown to be able to achieve in an even better way - those same goals instead of any other personal/subjective priority. Without some kind of "sanity filter" like that in place, the team, and likely a part of the community too, is only liable to waste its time and energy just trying to correct misapprehensions or, worse yet, might end up performing any number of pointless experimenting only to re-learn the same lesson as before after having diminished the ingame experience of a whole lot of players for no good reason and incurring a commensurate opportunity cost of doing anything more productive with that effort in the process.


There's two additional, hopefully brief n' edifying, points I'd like to touch upon before ending this, both also drawing from CEONSS' past tackling the repeat limit question.
The first is pretty much just a historical data dump, compiled as an attempt to understand and reconstruct that particular server setting's entire past course, through msg. board searches of related terms. Here's, then, what the most complete list I could produce of tracked CEONSS map repeat limit (RL) changes since its founding looks like:


launch - ~2010: repeat limit initially set to values between 15-20, apparently (unsure)
~2010 - Sep. 2012: RL at ~30; a repeat limit of 15-20 "[did] not work" (heinz, here and onward)
Sep. 2012: up to 35 for a few days, then 30 again (same link), then reduced to 25 for months
Feb. 2013: up to 30 again
Jun. 2013: reported as 35 (heinz, here), then likely reduced below 30 again
Apr. 2014: Cat mentions internally inclination to increase the RL to 30
Dec. 2014: Cat lowers RL to 20 (no elaboration on a longer-term plan, but my impression is it didn't stay there for long)
Jun. 2016: "over past few years the server's roster size went from ~100 to 74 maps, RL is set at 28" (Ema, here)
Jan. 2017: "We started from an all time low of 20 and have been increasing it ... with the aim of getting it up to 30" (Cat, here); RL lowered to 20 for 6 weeks as a test
Feb. 2017: RL "will go back to 27 soon and will be slowly increased up to 30 as popular maps are added" (Cat, same thread)
Jan. 2018: internally mentioned that RL is currently set to 29
May 2020: RL internally mentioned to be 27
Jul. 2021: RL lowered to 25

What I'm trying to get across here is that, with the exception of the server's initial period, where things were done mostly instinctively and without the benefit of sufficient ideological evolution or reflection, for the most part of CEONSS' existence, the map repeat limit has hovered around a 5-point range from the value of 30, namely within the 25 to 35 region; periods where the RL was dropped down to 20 or even lower were rare and predominantly attempted within a trial/experimental context. Additionally, regardless of map roster length, the views of pretty much every server staffer, and owner, on the subject of a sensible repeat limit value, likewise, seem to've converged on that same 25-35 region throughout the years. I don't believe that to be a coincidence, but rather a pretty telling indication in terms of earned experience as well as common sense.


The second point has to do with what we might glean from the, admittedly few, times that the topic of changing the RL has been publicly broached on the CEONSS msg. board - specifically what the typical arguments by community members tell us about how this debate usually goes.
In the rare few times where changing the RL has either been the focal point of a thread (2 or 3 times at most, this one included) or just brought up as a passing side-issue, but to an extent also in a good chunk of internal discussions as well, there's IMO two prevalent characteristics in people's arguments: for one, it's usually framed as a personal impact issue (how the server setting affects my ingame experience, compared to other subjects where the focus is on how an admin choice might be affecting the community as a whole), and, for another, the suggested remedy is likewise of an entirely subjective nature ("just use this figure which seems more right to me instead") without further elaboration that would allow for testing and comparing of competing alternatives.
What you get from such a clash is different people conveying their own sense of a balancing "sweet spot" between risk aversion and room for freshness/surprise (or, on the flip side, avoiding boredom or a sense of wasting time on bad maps), which invariably will be different from person to person - one player's satisfaction for managing to play a "brand name" map during their session might be adding to another's frustration and sense of crushing monotony for the same reason.
To my mind, this combination of a problem that is entirely subjective/perceptual in assessment, but also not possible to generalize to some broader principle on which both sides could at least agree and thus create common ground to start tackling the issue together from, is the definition of a zero sum game: for one side to win (changing the RL in one direction), the other must necessarily lose to a commensurate degree, and there's no mitigating factor to be found. On top of that, there's also the aforementioned absence in such debates of any kind of underlying criterion/standard or methodology & measurable metric(s) proposed by which to test, verify and/or falsify competing hypotheses in terms of what might be best for the community as a whole. This is pretty much the ONS version of "thermostat wars", based entirely on people's personal convictions, experiences and inherent biases. The only options left once one realizes this are either to go with the staff's collective instinct or to keep asking ingame, or periodically polling the community, and re-adjusting based on these findings, the latter likely proving both tedious and potentially incendiary.
The polling option segues nicely to the final observation I'd like to make here. Previous CEONSS head Cat did try just that about half a decade ago, being willing to go even with a marked change from the norm if sufficient community support showed itself to be there for it. Well, for all the attention the more radical departure suggestions might've received at first (then it was "drop it to 20"; now it's even further down to 15!), what's more salient to me even in the context of the total voting group being a far cry from representative of the entire CEONSS community, is the final vote breakdown that might as well be yet another instance of the oft-mentioned political trope of "a loud minority vs. the silent majority": more than double the amount of people preferred the repeat limit stay at 30 than those who wanted a substantial reduction to it. I'd like to hope this valuable lesson has still stayed with the server's active staff, and, as ever, recommend prudence and measure in all things ;).



Alright, this went much further and got far rantier than I'd initially imagined it could, so I think this is a good place to end. To avoid any misunderstandings, I'll also repeat what I said up top: what I offered above is merely an attempt at providing historical context, i.e. my account of how the issue of adjusting the server's repeat limit has been considered and approached in the past on CEONSS. Although my impression is that there hasn't been any substantial departure from that mentality within the current team administering the server, please feel free to correct me if diverging views have developed more recently on the subject.
Eyes in the skies.
Image
joeblow
Posts: 318
Joined: Sun 16. Feb 2014, 20:41
Description: male gamer
Location: tennessee,usa

Map cycle timeout

Post by joeblow »

Well ill voice my yearly opinion on this...as an active player

2 times this week some rarely used map got voted later in the evening and the server emptied.

The server is only really popular about 3-4 hours a day if that and near the end there are no "good" choices.

My solution has always been have the map vote reset somehow so ALL the maps are available right before the server gets busy.

Thank you.
User avatar
EmanReleipS
Administrator
Posts: 2948
Joined: Wed 2. Oct 2013, 23:18
Description: Pancake Fairy
Location: Germany

Map cycle timeout

Post by EmanReleipS »

For the record, changing the map repeat limit won't help much with players making bad voting decisions... ;)

I know historically Cat has always based the map repeat limit on the number of popular maps. Now that we have fewer active players and a shorter time that a group of players are online, I see the benefit of dropping the map repeat limit. In particular, some people may hop onto the server in the afternoon and play a very popular map with just 1 or 2 other players, which locks the map for the remainder of the evening. This is why I already lowered the map limit to 25 a while ago.

I'm open to lowering it further based on a new metric called "how long is the waiting time before a map will open up again".

I'm not a math person, so this calculation is a bit rough, but let's say that a map takes 10 minutes at minimum (one quick round and then a quick victory) and 25 minutes at maximum (map going into overtime). Looking at our stats database, matches below 10 minutes or above 25 mins seem less common.

With a repeat limit of 15, this would mean between 2.5 hrs at minimum and 6.25 hrs at maximum before a map can be replayed.
This would also mean that the same map would be open for voting between 3.84 and 9.6 times per day, but the higher number is very unlikely because the server doesn't actually switch maps if no one is online. (Meaning the map repeat counter does not move forward for the majority of the day!) Therefore the average number of how often the map can be played would be skewed to the lower end instead of sitting in the middle of this range. And "being open for playing" still does not mean it would actually get played that often, just that it would be possible.

Let's look at the current limit of 25:

This means roughly a minimum of 4.16 hrs between the same map or a maximum of 10.4 hrs. The same map is available between 2.3 and 5.7 times per day.

Basically the only time someone would encounter the same map on the same day would be on weekends, if the person is playing several times per day and many hours apart (noon and evening or mid-afternoon and night).

I'm open to lowering the map repeat limit further to allow late evening players on weekdays to vote for maps again that the afternoon/early evening crowd has already played. My suggestion would be to go with a limit of 18 for that, which would be a waiting time of 3 hrs at 10 mins per map, 4.5 hrs at 15 mins per map or 7.50 hrs at 25 mins per map between the same map. I'm aiming for a waiting time of ~5-6 hrs here, so that maps which were voted for before 6pm would open up again around midnight.

It should be noted that this would not help with the problem of playing the same map night after night. That could only be resolved if we put the limit so high that the map would only be available every other day or by puking on the players who voted for that map. :lol:

Honestly I wish the map voting handler would be more sophisticated so that we would have better options to adjust it, but it is what it is...

Thoughts?
Image
User avatar
Miauz55555
Posts: 1819
Joined: Sun 7. Jun 2015, 22:12
Description: https://discord.gg/X4V8THM
Location: Germany

Map cycle timeout

Post by Miauz55555 »

TL;DR nothing, but the numbers and the text arround actually, sorry.
@Peg thanks for that numbers and the detailed informations.
@Ema I think 18 is fine.

Main reason to lower it is in my opinion because there are not so many games during the day atm.
------------
And I vote to add some more maps which play fine with 4 players or even alone as "starter" maps. So with good bot paths and not to big. Sure they will be spammy with 32 players, but the good big maps are to big for lower player counts. But that's for anouther topic. =)
Image
ankeedo
Posts: 7
Joined: Thu 24. Mar 2022, 18:51

Map cycle timeout

Post by ankeedo »

the main reason is that there are not many maps there, no map or vehicle addition/modification, so the choices are very limited, when someone plays alone and choose the most popular maps there will be no maps worth playing for other players in the peak time.