dimanche 15 janvier 2017

Selecting a WoW private server : towards common quality criteria

Introduction : Many players are looking for the right private server to select, how to decide on one isn't clear and there is no systematic guidance on how to do that. Servers keep trying to improve their quality but few quantitative data can guide them on what to prioritize.
Methods : I sent whispers to 450+ K1 and K2 characters asking them about their top 3 priorities when selecting a private server and gathered their answers.
Results : 80 players answered my survey, with 76 providing priorities. 19 broad categories were then isolated. The top 3 criteria were : #1 Population (amount & type) (n=42), #2 Scripts bugs & updates (n=25) and #3 Community (friendly, active) (n=21). 
Discussion :


Many players are looking for the right private server to select, how to decide on one isn't clear and there is no systematic guidance on how to do that. Servers keep trying to improve their quality but few quantitative data can guide them on what to prioritize. My goal was to systematically collect qualitative data on evaluation criteria so as to be able to create a checklist of quality criteria that could inform server improvements and server comparisons (in the hope of informing future changes and website design).

I believe that by asking enough players the most important server quality factors and criteria will emerge. I decided to focus on actual players in-game to prevent contaminations if the answers were on a forum (eg. players see the previous answer and answer exactly the same) and so as to reflect the current playerbase (instead of potential players which might never join at all or players on forums which I believe to not be representative of typical private server players).


I sent in-game whispers to all the players appearing in the /who list from two of my level 60 Horde characters on Kronos 1 (K1) and a level 1-6 Alliance character on Kronos II (K2) at multiple occasions during weekends (from 2-15th January 2017). I stopped whispering from my low level character as I was getting muted much faster due to different communication restrictions. I had originally planned to get 100 answers (convenience sample) or stop once content saturation seemed likely but decided to stop at 80 answers due to lack of patience. Over 450+ characters were whispered (I didn't track this number with precision), with about 80% of them on K1, which leads to an answer-rate of about 17.78% (80/450). I did not apply a level restriction to answers. The following message was sent :

Greetings, quick poll, what are the 3 most important things you look for when selecting a private server? Thank you for your help! (ignore if already asked)

Answers were copied with the EasyCopy addon and pasted on an Excel sheet. If multiple whispers from the same person were received, I combined them in a single sentence. I spread my whispers over some time to prevent being muted and kicked from the server by anti-spam scripts.

Main items were then manually extracted and combined into broad categories. This process wasn't entirely objective and a second evaluator would have been helpful.


Here are the main categories of what mattered to the players I surveyed :

Category & sub-category

# Players saying it is a top 3 priority


 - "No" latency/lag
- "Not too much" / "low" latency/lag
- "Stability"
- Amount of delay (ms)


Cash shops & in-game advantages

- Not having to pay to play (free)
- No way to buy power or advantages
- Clear costs of things you can buy 


 Scripts, bugs & updates

- "Fast" fixes 
- Actively bug fixing
- Regular updates
- "Quality" of scripts
- Small amount of bugs 
- No bugs
- Will to fix small and big bugs 
- "Attention to details"
- Using other server's codes 
- "PvP scripts" 



- "Stable" amount of players
- "Enough" / "good" amount of players 
- "Enough" players to "farm"
- "Not too many/few" players
- "High" amount of players 
- Amount of "active" players
- Amount of players from different communities (english, czech, etc.)  
- Equal amount of Horde/Alliance 
- "2000-3000" players



- "Good" / "reliable" players
- "Good" guilds
- "Friendly", "helpful" players
- "Good atmosphere"
- Knowing someone / friends / IRL there
- Language (czech, english) 
- "Active" raiding guilds
- "Active" raiding guilds in _ timezone 
- "Mature" players


State of PvP

- PvP "scene" 
- "Active" PvP community
- "PvP scripts"
- "PvP activity" 
- "Balance"


 History of the server

- "Reliability" 


Server economy

- "Auction house"


Being as close to the original version (blizzlike)

- "Close" mechanic-wise to the original
- "Everything" as it was or as close as possible / authentic
- "100%" blizzlike
- "Smooth" gameplay
- Correcting bugs that existed in vanilla
- "Upgrades" over the original WoW



- "Recommended" server
- "Good" reputation


Support, server team and Game Masters

- "Active" GMs
- Dedication to the server/job
- "Good" support
- "Quick" replies
- How the server is managed
- "Reliability"
- "Serious" / "reliable" team
- Access to customer services
- "Attitude" of GMs
- Presence of a community manager



- Regular updates
- "Good" /  "Quality" content
- New content regularly released
- "Fast" releases
- Amount of content available
- "Hard" content 


Server stability

- "Good" 
- "Reliability"
- Knowing the server won't be shut down
- "Low" downtime
- Server costs / free to play


 Rates (experience, gold, etc.)

- "Low" rates (1x or lower)
- "Quick" leveling rates (5x)


Special or additional events

-  "Interesting" events
- Amount of events / frequency

Ease of use / Access

- Easy to install
- Easy to join the server
- No queues to join the server 



- Vanilla expansion
- Patch



- No spam



- Modified respec costs (5g)
-  Race / Faction change
- "Fun"


I tried my best to separate items into categories, but the process was prone to errors (eg. does "community" mean the player wants a "friendly" community, a certain amount of players or both?), it should nevertheless give a correct overall impression. Note that some categories are very likely linked, like scripts and being "blizzlike" (whatever this means), or bugs and active GMs or economy and population.

A couple players asked further questions about the goal of my survey, wondered if I was member of the Kronos team and thanked Kronos for their work and efforts.

Two (2) players complained that my whispers were bothering them. This was partly due to me sending whispers twice during the same day to the same players. I tried to apologize and explain the limitations of my technique (double-checking that I haven't sent a message to a character already was too time-consuming). A few players answered jokes or meaningless content, which I discarded and ignored. One player mentioned he would talk to a gamemaster about my behavior but I didn't get whispered by them at any time.


This is to my knowledge the largest survey about player's preferences and quality criteria when selecting a World of Warcraft private server (although I admit I didn't search much due to being more interested in my project).

Each player being different, each one has different priorities, this was very clear in the answers of players (eg. "no bugs no bugs no bugs" versus only "population"). But on average some categories were much more important to players than others (eg. Spam n = 1 and Population n = 42). While players have priorities, this doesn't mean the ones they most value is the most important. As an example Population is ranked before Server stability but there cannot be a server population if the server is always offline. Nevertheless their attention is likely to be focused on their main quality criteria and how it changes over time (eg. more or less population).

Some players wished for entirely different things (eg. "no changes" and "modified respec costs"), had sometimes unrealistic expectations (eg. "no bugs" and "no latency") or were looking for things that are highly subjective and hard to define (eg. "fun", "good server", "100% blizzlike"). Few criteria were objectively described. Among those I could find : "No queues to join the server", "Faction change (access to)" and  "Vanilla expansion".

It is not clear how much this survey would reflect the opinion of potential players not on the server (or any private server at all), players from other servers or players on forums. Players that did not answer might have had different priorities. It should be kept in mind that priorities are likely to change over time; in this survey very few players said spam was a priority but back when goldsellers were talking on the world channel and whispering players every minute this would have been different.

Interestingly to me, transparency (how decisions are made, what money is spent on what, what is happening behind the scenes, current online per faction players, etc.), clear plans for the worst and the future (DDoS/loss of players/massive increase of players/loss of GMs/cheat/etc.), vision for the future, rules, "attractive" website, server language, raiding hours, strong presence from leader/admin, active online community (Twitter, forums, etc.), impartiality from moderators, lack of trolls on forums, pathfinding script, trust and code-sharing didn't clearly emerge. This is probably due to me asking for "quick" answers, which had the downside of being broad and not very specific. But I fear that asking for more details would further reduce the amount of players answering.

Players that want to introduce a server can use the identified criteria when discussing a server. As an example, considering how Latency is high on the list, it is surprising that so few server websites display any information about it. Same goes with downtime, as servers currently only show uptime (time online) which is not the same thing.

Next step : Making objective criteria for each category.

Supplementary data

20:41:29 [W From] <39:1>: dont have lag, have balance and dont pay to play
20:42:55 [W From] <11:2>: Hello friend. Good community, good servers and more awesome things
20:45:06 [W From] <60:3>: 1) Blizzlike mechanics 2) Fast fixes of bugs 3) Lag-free
20:45:21 [W From] <60:4>: stability, and reccomendations
20:46:24 [W From] <57:5>: i am happy with the game . i have no opinion
20:49:00 [W From] <60:6>: 1 player base, 2 do they fix bugs, have active gms, 3 do I know anyone on the server
20:49:37 [W From] <20:7>: good and stable player base,staff dedication and constant updates.
20:50:04 [W From] <42:8>: well actually this is the first private server that i join in like 10 years so... i dont actually look for something specific... just heard that I can plat wow vanilla here :D, and of course I do like that it is Blizzlike rates...the closer to the original game... the better
20:49:41 [W From] <17:9>: #1 a good player base ( need a good amount of players to play the game good , #2 a bug free server( think about dungeons and quest or even server events, and dont really know a third reason xD, if these 2 things are spot on then hanging around till the end of the server
20:50:41 [W From] <27:10>: 1: No mods, everything as it was. 2: Population 3: server Stability
20:50:56 [W From] <36:11>: Friendly community, stability of servers and new content being released timely
20:55:05 [W From] <60:12>: im not selecting private servers, I hear about them by accident and thats it
20:55:27 [W From] <60:13>: pop scripts
20:57:05 [W From] <60:14>: 1. "Atmosphere" / Friends / Social Aspects 2. Pvp Scene 3. Quality of scripting
22:42:59 [W From] <18:15>: population, reliability, scripting
22:42:54 [W From] <20:16>: I dont play them very often, I had this one sent to me from a friend and have been enjoying it, I guess just one that doesnt have a ton of bug
22:43:54 [W From] <7:17>: This is my first try with a privvate server, friends sugestd i
22:47:02 [W From] <1:18>: scripting, population and speed of releasing content
22:49:02 [W From] <26:19>: czech comunity, population, vanilla wow
22:50:04 [W From] <60:20>: population, stability and content
22:53:49 [W From] <60:21>: only 1 thing, there should be my friends playing
22:53:49 [W From] <35:22>: High Population, Smooth Gameplay (as close to retail as possible) and serious admins.
22:54:08 [W From] <60:23>: support, stability, authentic
22:55:06 [W From] <22:24>: Authenticity, Volume as it pertains to the player base and some sense of security (Knowing that the server isn't under constant threat of being shut down), Those are my 3, hope I've been of some help.
22:59:40 [W From] <7:25>: 1-low rate because it's more fun to lvl up, 2-good content that i like, 3-good online
22:59:53 [W From] <2:26>: enough players, stable performance, ongoing development
23:00:05 [W From] <53:27>: 1) where irl friends are playing and that's pretty much all sorry
23:00:47 [W From] <36:28>: customer service, the willingness to fix any bug big or small, and population, or quality of customer server I should say
23:00:08 [W From] <3:29>: number of active players, ethnic composition of players(language etc.) and type of server ofc, expansion, xp rate..
09:51:38 [W From] <10:30>: Server  stability, population, and authenticity to vanilla (minus bugs)
09:51:41 [W From] <60:31>: population, active pvp scene, equaly divided sides horde / alliance, active raiding guilds in my own timezone, + interesting events dont hurt as well to keep the people entertained, + last thing not to spam good GM support and quick reply
09:55:52 [W From] <10:32>: Fun, versatility, and population. Ty for all the great work!!
09:57:39 [W From] <14:33>: latency, population, and hoping it doesn't get shut down by blizzard. A good community doesn't hurt but you said three things.
09:58:14 [W From] <60:34>: stability, scripting, population
09:59:32 [W From] <24:35>: 1 server reliaility; no lag or not much lag.2. A decent amount of players not too many bt not to few either. 3 Mature palyers.
10:02:45 [W From] <23:36>: population, bugs, content.
10:02:50 [W From] <60:37>: GM attitude, thats all it matters
10:03:28 [W From] <14:38>: mostly just scripting and population
10:08:04 [W From] <58:39>: 1. a bug-free, authentic recreation of the original wow 2. a friendly, helpful, mature community, there's no 3 :) , that's all i want :)
14:27:16 [W From] <1:40>: friendly community, quality, latency
14:27:38 [W From] <60:41>: stability, population and server staff
14:27:58 [W From] <60:42>: stability, no lag, no bugs
 14:32:19 [W From] <60:43>: one important thing is the population ofc, you have to merge :), andf event, there is no real event there :s, and the third thing i thnk is to lvl up quiote quicvkly , maybe x5 [...] they may need a real community manager, but i know that is really diffucul and they did a great great works
14:32:22 [W From] <60:44>: scipting, non toxic comunity, good guilds.
14:32:28 [W From] <34:45>: stability, AH, ppl
14:38:44 [W From] <13:46>: Population, Active Raiding Guilds, Active GMs
14:39:14 [W From] <60:47>: 1. playing on server with irl friends, so community in generel, 2. server stability and i like how some things are changed from vanilla, like respeccing talents costs 5g, so 3. must be upgrades from retail version
14:41:37 [W From] <13:48>: Greetings. I'd say the number of active members, the activity of the GM's (do they fix bugs) and the general content (which expansion/patch..)
14:38:54 [W From] <37:49>: uhhh...scripting, attention to detail (quest accuracy, etc), and community
14:51:55 [W From] <60:50>: To be honest this is the first time I'm playing on a private server so i'm not really an expert on that topic :) I'd say that not having too many bugs, having a population that make you able to farm the content and a good support from gm
20:33:17 [W From] <31:51>: hmm.. low latency, no overcrowding, and low downtime!
20:33:29 [W From] <60:52>: to have 0 lag :D, and just to have fun, noting special
20:33:39 [W From] <51:53>: really if it is just like the old game and if there are a good amount of people playing it
20:39:08 [W From] <60:54>: community (e.g. highly populated), 1x rate, no buy to win
20:39:08 [W From] <60:55>: how goods its scripted, content (vanilla prio), costs
20:39:30 [W From] <60:56>: reliable people and a good guild
20:40:14 [W From] <3:57>: easy setup, good community,free
20:39:50 [W From] <35:58>: Yello, its 1. Population, 2. Stability 3. Bug free & functioning properly
20:40:44 [W From] <2:59>: community, how well it is maintained, and population
20:45:37 [W From] <60:60>: heyo, 1) population 2) stability 3) number of bugs and how much content is available
20:45:45 [W From] <27:61>: no bugz
20:46:16 [W From] <60:62>: How many players are playing at the server, race / faction change and service from the GM:D
20:50:30 [W From] <60:63>: number of people, lag and real xp while leveling
20:50:34 [W From] <52:64>: integrity, reliablity, ease of use
20:50:44 [W From] <32:65>: population, latency and economy I would say
20:51:23 [W From] <22:66>: Hardmode, a decent sized community and no ques, I guess
20:51:35 [W From] <29:67>: population...I like it mid lvl, not too high, 1x xp...has to be this setting for xp, almost no or no change from vanilla settings, thats all I look for friend
09:51:43 [W From] <60:68>: reputation, support, community
09:52:21 [W From] <1:69>: Hi. 1. Lag/MS 2. Scripting 3. Pop (2 - 3k seems optimal)
09:55:04 [W From] <7:70>: 1. Skripting 2. Is it really Blizzlike??? 3. Population
09:59:02 [W From] <60:71>: well i used to play on rebirth but they are closing on the 26th of march so i came here cos i had a spare 60 here, it had way better scripting but they say they will release there codes so hopefully this server will use some
09:59:19 [W From] <24:72>: number of players, history of realm, and community, i play on twinstar realms for 5 years
09:59:24 [W From] <9:73>: sorry im completly new to wow im only trying this because my friend swears i cant enjoy the new wow in the same way i would if i played private servers LOL
10:05:56 [W From] <60:74>: community, latency, pvp
10:07:13 [W From] <60:75>: scripting latency 100% blizzlike
10:08:05 [W From] <60:76>: no bugs no bugs no bugs
10:08:08 [W From] <60:77>: Not rly priorities..just helpin some lowbies with an elite qst..if i dont push buttons they die lmao
10:09:09 [W From] <60:78>: population
10:09:41 [W From] <60:79>: being able to have my pet stay with me through a zep loading screen, cant get that here, other than that its great, some mobs stuck in walls
10:11:04 [W From] <60:80>: 1) pvp scripting 2) activity of pvp 3) population

vendredi 13 janvier 2017

Kronos Survey on Facebook - 13th January 2017

I just participated in a Kronos survey that appeared on Facebook. It might be interesting to see how my suggestions end up on the long-term.

What do you like about our projects?
  • Dedication to the project
  • Leaders that have been there for a long time
  • Asking your players what they think of it (this survey) and making them part of server development (to an extent)
  • Being okay with doing some difficult decisions for long-term server health
  • Friendly community in-game from my experience
What would you like to change?
  • More transparency over what the team plans, what the thought-process behind changes incoming or current are, yes secrecy has a role but too much leads to decisions that seem unfair or not thought out at all (what the team is doing = mostly a black box). Please share some thoughts, struggles, achievements, etc. I understand the need to be seen as trustable, confident at all times but see it as unrealistic and unhelpful in general. 
  • I'd like a feeling of going forward in terms of the server improving in some way or being active (population, updates, bugs, features, news, etc.), not sporadicaly but maybe daily, weekly or monthly at least (I understand that limited time & team working for free make it easy to say, much harder to do). 
  • Open-mindedness to changes that might not entirely be blizzlike for the server good (discussed beforehand). 
  • Higher transparency over server costs, donations, stars, auctions (having no idea about this sustains myths and stigma against the server, people fill gaps with anything). 
  • No longer selling mounts/character power (assuming the server doesn't depend on them) while allowing players that bought those to keep them (this is often a counter-argument to playing on Kronos).
Do you have any suggestions for improving our services?
  • In-game large surveys with automation to collect data (real players are those actually playing), published results with objectives, implementation plans and evaluations. 
  • More user-friendly bugreporting web interface (eg. checklist included, guide that shows up while writing, links to archives or sources). 
  • Perhaps addon to facilitate goldseller reporting without all the hassle of making a report, adding a screenshot, etc (most players don't bother, the easier things are the more players do it). 
  • Monthly/bi-monthly communication about current plans, acknowledgments of issues and good things, how the server is going forward, what has been done, achievements, etc. 
  • Testimonials from players on website. 
  • One-click way to install/download Vanilla for new players (it's still difficult for too many). 
  • Tweets/posts from developers or in-game announcements of changes/corrections (eg. "Frosbolt now no longer does bug X when Y (dev 1)"). 
  • Going ahead with pre-post tests of design changes (eg. "We just changed this feature to do X and are monitoring, if not okay according to Y at 1 month we'll do Z", small website changes). 
  • Guide/Youtube guide for GMs or such. 
  • Server blog with team's points of views or some player's highlight?
Anything else on your mind?
  • Trying everything at once is a recipe for failure, focus on a couple things you believe (or have data that suggest) is most important and do that well. It might fail and that's okay. 
  • Don't aim for the sky, small improvements are nice too. 
  • No need to be the one big giant in the "market" if what you do is good (I think it already is) and the population is sustains itself. 
  • Note : Some players play equaly on K1 and K2 or on many factions and this survey can't notice them. I hope my expectations aren't too high and wish you well !
That was roughly what was on my mind when I quickly answered it. Let's hope for the best!

dimanche 13 novembre 2016

Essais cliniques non publiés en Suisse


Approximativement 25-50% des essais cliniques réalisés sur des humains (des études où l’on tente d’évaluer l’efficacité et la tolérance d’interventions ou de traitements médicaux) ne sont jamais publiés, ceci en grande partie parce que les auteurs ne soumettent pas leurs travaux à des journaux scientifique, ont des résultats négatifs ou ne les estiment pas intéressants

ClinicalTrials est une base de données qui catalogue les essais cliniques enregistrés avant qu’ils ne soient réalisés et surveille si leurs résultats sont publiés. Celle-ci permet de mieux réaliser la proportion de résultats qui ne sont pas partagés et les modifications appliquées au protocole initial (le plan de l’étude). Dans certaines institutions, l’enregistrement des essais cliniques est obligatoire. Malgré cela entre 30 et 50% des essais cliniques enregistrés sur ClinicalTrials ne sont jamais publiés.

ClinicalTrials a un rôle important dans la recherche scientifique. En effet, si des études remarquent qu’une intervention est inefficace et ne publient pas leurs résultats, les chercheurs risquent de surévaluer l’efficacité de cette intervention en se fiant aux études aux résultats positifs. Lorsque des modifications sont réalisées au protocole de base d’une étude, celles-ci peuvent induire des erreurs dans l’analyse ou l’interprétation des résultats. Si des interventions ne fonctionnent pas et qu’aucun chercheur ne publie à ce propos, d’autres chercheurs risquent de réaliser la même erreur et de dépenser inutilement des ressources (argent, temps, etc.). Enfin, la réalisation d’essais clinique implique la participation de volontaires, l’utilisation de leur temps et l’exposition à des effets indésirables ; si les résultats ne sont pas publiés leurs efforts peuvent être considérés comme en grande partie perdus, personne n’en profitera.

Le 3 novembre 2016, Ben Goldacre et Anna Powell-Smith ont publié un article (pas encore révisé par les pairs) portant sur un outil de suivi des essais cliniques (le TrialsTracker) inclus sur ClinicalTrials et permettant d’estimer la proportion d’essais non publiés, en fonction de leurs sponsors. Selon celui-ci, l’hôpital universitaire de Bâle (University Hospital, Basel) situé en Suisse a sponsorisé 64 essais cliniques de phase 2, 3, 4 ou inconnue et terminés depuis 2006 (en excluant ceux terminés ces 24 derniers mois, soit depuis novembre 2014) et parmi ceux-ci, 38 (59,4%) n’ont jamais été publiés. 26 essais auraient donc été publiés.

En accédant au lien « See all completed trials on ClinicalTrials.gov » après avoir sélectionné l’université de Bâle j’arrive sur le site web de ClinicalTrials qui indique 136 essais complétés avec les critères (Completed | Interventional Studies | University Hospital, Basel, Switzerland [Exact,Lead]), cela ne correspond pas au chiffre précédemment mentionné mais les critères de recherche ne sont pas exactement les mêmes. Si je rajoute (Studies with Results), seuls 5 essais semblent publiés, dont 4 correspondants aux critères de TrialsTracker. A nouveau, les chiffres ne correspondent pas. Serait-ce une erreur de ma part ?


Curieux, j’ai brièvement tenté de trouver les résultats publiés de ces essais sur Google en recherchant le nom de l’essai, son nom  officiel ainsi que son numéro NCT (un numéro unique attribué à chaque essai enregistré sur ClinicalTrials). Je me suis limité à la première page de résultats du moteur de recherche Google. Lorsqu’un article publié dans un journal ou un document PDF avait le même nom/numéro que l’essai enregistré, que l’université de Bâle était mentionnée et que l’année de complétion et les auteurs correspondaient, j’ai considéré que les résultats avaient été publiés. En cas de résultats trop larges et hors-sujet j’ai rajouté « basel » aux mots recherchés.


Année de fin
Non trouvés
Non trouvés
Non trouvés
Non trouvés
Non trouvés

(Double ?)
Non trouvés


Sur les 20 premiers essais sponsorisés par l’université de Bâle (selon la liste obtenue au travers du lien de Trials Tracker) correspondant aux critères de Goldacre et complétés, 19 (95%) n’indiquent pas leurs résultats sur ClinicalTrials. 6 essais (30%) restent introuvables après ma brève recherche sur Google. 12 essais (60%) sont trouvables et leurs résultats ont été publiés mais n’ont pas encore été introduits dans ClinicalTrials. 1 essai (5%) est partiellement trouvable (résultats brefs disponible sur le site web de l’hôpital). 1 essai (5%) est trouvable et ses résultats ont été rapportés dans ClinicalTrials.

Trials Tracker indiquait 38 essais aux résultats manquants sur ClinicalTrials. Parmi mon échantillon de 20 essais, j’en ai identifié 19, dont 12 (63%) avaient déjà publiés leurs résultats dans un journal scientifique. 

Bien que de nombreux essais n’ont pas rapporté leurs résultats à ClinicalTrials, mes chiffres ne s’accordent pas à ceux de Trials Tracker, j’imagine que je ne réalise pas la même recherche ou que nous partons de données différentes ?