Please implement TremStats for the GrangerHub Public server. Recently I have heard certain people talking about what the majority of players want, but both they and their opposition were unable to provide hard, statistical FACTS to support their argument. In 1.2, on the Official US1 server, I was able to track my stats and look at statistics about which maps are played most, which maps are voted for the most, and which maps’ votes pass most frequently, among other statistics. I think TremStats would be a valuable resource for the community to make informed decisions about the server moving forward, as well as provide a fun way to compare oneself to one’s friends.
[size=18]Would you like to see a TremStats page for the GrangerHub Public server?[/size]
GrangerHub will have statistics about servers/mods/maps/players/clans on the GrangerTracker that it is developing.
But I would like to add that as the current map selection process is fundamentally flawed on GrangerPub and on GrangerClub now, knowing which maps are played the most, which maps are voted for the most and which maps’ votes pass the most frequently will not tell you anything about what the majority of the players do or do not like. Statistical results will tell you nothing of value if your data is skewed, as it is said in computer science “garbage in, garbage out”.
Before statistics of how often maps are played, changed, voted on, etc can tell us anything of any value, the various systems related to map selection and the decision to or not to change a map (i.e. voting enhancements, PlayMap, and Pre-Game Warmup) need to be completed. Those systems are still a work in progress requiring a lot more feature implementation & finetuning, and would not result in 100% accurate information on the preferences of maps by the majority of players, but already it is substantially better than the current system on GrangerPub and on GrangerClub.
Is there some reason why the old TremStats could not be implemented temporarily until this is complete?
Howso? It seems to work fine. It already requires a supermajority to pass, and everybody on the server has the option to vote. If they choose not to vote, their vote is not counted. If they are AFK or do not know the map, they are not forced to have their vote counted as a “no”. There is no situation when someone’s vote is decided for them simply because they chose not to vote. If someone fails to show up for a vote in Congress or for a presidential election, their vote is not counted; it is not counted as a “no” or for a specific person simply because they didn’t show up.
It obviously would not definitively say what the majority of players do or do not like, but it would show statistical trends of votes for certain maps passing/failing, as well as showing which maps are most frequently voted for. The question isn’t what a majority of players like, it’s a question of what the majority of players ONLINE like, and thus the opinions of players who are online more frequently are favored. If I spend 3 times as much time on Tremulous as another player, my opinion matters 3 times as much when it comes to the server because it will affect me 3 times as much. While this obviously isn’t true in normal votes, it makes complete sense for things such as maps on the server, server map rotations, etc. So while it is not representative of the majority of players, it will, given enough time to settle down, reach a point where it is representative of the opinions of players based on the amount of time they spend on the server. No data is insignificant, it is only insignificant if the conclusions drawn from it are irrelevant to the data or based on incorrect assumptions about the data. I’d rather have the data and not have it change anything than not have the data at all.
This is incorrect. The statistics would undeniably be more valuable than not having the data at all. The voting system of test7341 is probably even more flawed than that of vanilla Tremulous, and the warmup would only change whether or not the map was considered “played” or not based on when the map was skipped, which is currently irrelevant because any map that is skipped was skipped while being played.
It would be better than the current system because you currently have NO good way of determining the preferences of the majority of players besides word of mouth and personal supposition. Both of these are quite clearly skewed towards your personal preference. Perhaps you choose to argue against statistical analysis because it would prove wrong your own opinions of what people want from the server? Or perhaps you are ignorant enough to believe that no data is better than some data. I don’t know, but either way, it is unhelpful to deny the community a way to come to their own conclusions about the server from undeniable statistical data.
For example, looking at GameTracker shows that ATCS has been played more than ATCS_2015 in the past week, in spite of the fact that ATCS_2015 is in the rotation multiple times (and has been the full week) while ATCS has not been in the rotation for the whole week. In spite of that, ATCS has been played more frequently than ATCS_2015 (15% to 14%). That’s useful data because it shows ATCS is a very highly voted for map. (And, in fact, it is the most played map in the past week, in spite of it not being in the rotation.)
Irrelevant information, incomplete information, and misinformation are not good to use for making decisions. The statistis on map selection you want would not be helpful in deciding what maps should and shouldn’t be in the rotation, since as I said the current selection process is fundamentally flawed on GrangerPub/GrangerClub. Deciding on a map rotation based on the behavior of that flawed system would at best improve nothing, and at worst amplify the negative results of that flawed system.
The current approach I’m taking for the map rotation is very imperfect (even the rotation system on its own is fundamentally imperfect), which is why we are working on these new systems to drastically improve that process.
@Menace13 , please link to the source of TremStats, and list all of the features it includes. I can’t guarantee that GrangerHub development will be able to find time to implement it as it is not currently a high priority, and there are many other items on the TODO list (one of which is the GrangerTracker).
Oh and we don’t have tremstats because it only can do stats for ONE server and is grossly inadequate for our needs.
Now if you want to debate what is good or not good for trem without making a fuckton of assumptions and twisting people’s words, please continue.
P.S.: I read your entire post, comprehended every point you tried to make, and determined that most of it didn’t dignify a lengthy response. If other admins didn’t respond to you in the way you wanted I’m sorry, but how about trying to understand where we are coming from before attacking any of us or criticizing what we are doing?
“everyone hates a new idea, until it’s done.” -some dev at Valve, talking about Half Life 2.