Re-Organizing DSB

[falconeer]

The Averages are wrong. Ent's comparing the average of 1000 games west to the average of 49,000 games east. After 2011 the population dropped to 50 people. Nothing determines how many kills I can get in a sector as proven with 45c's 44 kills in West. The power of averages works when you have data, such as 50k practices. East is accurate. West, and south, and north cannot be compared at all because the gaps in data is huge between all of them. None of them should be compared to east most of all.

Tell me, whats the average of East if you use the first 1000 practices? Is it 16 like west? or is it still 21?
Why doesn't Ent make a table of East averages. Is it because my point will be proven?
1000 data entries from 2010 in order.
2000 data entries
3000 data entries
10,000 data entries
20,000 data entries
30,000 data entries
40,000 data entries

Do them all in order, average the kills, and lets watch the average spike up and down? If it stays the same, giving you the benefit of the doubt of not cheating. Then you've proven the averages are the same no matter the data entries. But I suspect the kill average will be smaller than 21, and increase with more data.

[jim the chin]

Ent could do that, though I think an average from a sample of 1000 won't change much if you increase the sample size to 50000.

There's also a reason for the averages being higher in some sectors. If you push 3 people into a small space (east) then they'll get more kills between them. That doesn't necessarily make one of them MVP. If it's an even fight, they won't be MVP; but there's still a greater chance of them getting MVP in that sector. Any of us who've looked at MVP stats down the years will know that most come from east. That's what averages are about. Extreme cases will always happen though.

Say you're playing upper west or lower west against someone who is much worse than you, you'll effectively be dueling them for the whole game. For example, I played Gurliver in west back when he was a complete newb and went 81-27 I think. That made me the highest points scorer on my team. It also allowed me to attach in north to help lock that sector, bringing even more kills.

In general, you can MVP in any sector if you're playing against bad opponents. So citing individual cases is pointless. All I know is you can probably trust the averages, because if there's more space in a sector, there's more time spent flying between the different flag poles, and less time spent fighting. Even when I was MVP in west, 81-27 meant I only had 108 battles. Many east players gets scores around 70-50 if they win. They wouldn't be MVP over me, but they would have had 120 battles. If they'd had opponents that were as easy as mine, they might have gone 90-30 and got the MVP.

[falconeer]

It's not about space, 2 vs 2 happens everywhere especially in east. You rarely get a 3vs3 going, because for one, the third player is already dead. 2 vs 2 lower south. 2 vs 2 RN, and 2 vs 2 LN. There is no big difference between sectors, the space isn't different in west than it is south or North, it's all about the same space. And in fact South is the biggest sector with it's own duel scenario. All that matters is the play-style of the player, their personality, that's all that affects kills. If I'm going to chase an enemy down, he will turn and fight. If I run he will chase and we will never fight.

Prove East has a majority of MVPs, get the stats from league.nu, and show me the MVP was east for 60% of games. Even if it was, where was the MVP the other 40%? In west? I bet you MVP is almost evenly distributed, and that West saw the majority of MVPs. I've never seen a west player on the bottom of list, they were always top 3 in league. I'll check out league MVPs later. A bit busy over the holidays..

[Tembest]

Could you Ent change the value to 90 and 3 (bonus)? I know how to change base value but I'm not sure how to change the bonus.
I am extremely busy for a couple of weeks more so I cannot test the gameplay myself, but I'm sure others will. Perhaps it would be a good idea to announce
it somewhere as well so people actually there has been a change.

Edit: Oh and, I'll try to look up for the discussion for you Sever, but not today.

[Entr0py]

Could you Ent change the value to 90 and 3 (bonus)?

Yeah, I'll try to get to it later this week. I'll show BH/Eridu how to do it for the future, too.

The Averages are wrong. Ent's comparing the average of 1000 games west to the average of 49,000 games east.

That's a legit thing to do. Learn about how averages work.

Tell me, whats the average of East if you use the first 1000 practices? Is it 16 like west? or is it still 21?
Why doesn't Ent make a table of East averages. Is it because my point will be proven?
1000 data entries from 2010 in order.
2000 data entries
3000 data entries
10,000 data entries
20,000 data entries
30,000 data entries
40,000 data entries

Do them all in order, average the kills, and lets watch the average spike up and down? If it stays the same, giving you the benefit of the doubt of not cheating. Then you've proven the averages are the same no matter the data entries. But I suspect the kill average will be smaller than 21, and increase with more data.

Here you go. Averages for all sectors, in groups of 1000 player match entries, in order by date played. This includes the average for the current batch of 1000 player match entries, as well as the running average for the cumulative amount to that point.

Code: Select all

East:
Group    Group Count    Cumulative Count    Group Average    Running Average
    1           1000                1000      21.52177287        21.52177287
    2           1000                2000      22.06927528        21.79552408
    3           1000                3000      21.63997357        21.74367391
    4           1000                4000      22.52762958        21.93966283
    5           1000                5000      21.99766312        21.95126288
    6           1000                6000      22.43791477        22.03237153
    7           1000                7000      22.89823217        22.15606591
    8           1000                8000      22.33407107        22.17831655
    9           1000                9000      21.93317269        22.15107835
   10           1000               10000      21.45833799        22.08180431
   11           1000               11000      22.04474555        22.07843533
   12           1000               12000      21.96261862        22.06878394
   13           1000               13000      20.90978313        21.97963003
   14           1000               14000      21.48318347        21.94416956
   15           1000               15000      21.58327045        21.92010962
   16           1000               16000      21.39842016        21.88750403
   17           1000               17000      20.78741086        21.82279267
   18           1000               18000      21.60649204        21.81077597
   19           1000               19000       21.9061503        21.81579567
   20           1000               20000      21.99757255        21.82488451
   21           1000               21000      21.53128761        21.81090371
   22           1000               22000      21.56964277         21.7999373
   23           1000               23000       21.3722529        21.78134233
   24           1000               24000      21.28512635        21.76066666
   25           1000               25000      21.29312506          21.741965
   26           1000               26000       20.9278721        21.71065373
   27           1000               27000      21.08981771        21.68765981
   28           1000               28000       20.6986842        21.65233925
   29           1000               29000      20.88992888        21.62604924
   30           1000               30000      21.07341105        21.60762796
   31           1000               31000      21.28387125         21.5971842
   32           1000               32000      21.17020284        21.58384103
   33           1000               33000      21.07080266        21.56829441
   34           1000               34000       20.5644988        21.53877101
   35           1000               35000      20.82874564        21.51848457
   36           1000               36000      21.12130817         21.5074519
   37           1000               37000      20.98577684        21.49335257
   38           1000               38000      20.92899908        21.47850116
   39           1000               39000      20.89705327        21.46359224
   40           1000               40000      19.23854168        21.40796598
   41           1000               41000      21.03410178        21.39884734
   42           1000               42000      20.44682134        21.37618005
   43           1000               43000      20.67201101        21.35980403
   44           1000               44000      20.12775117        21.33180283
   45           1000               45000      19.81547679        21.29810669
   46           1000               46000      19.56845862        21.26050565
   47           1000               47000       19.3075406         21.2189532
   48           1000               48000      19.56247489        21.18444324
   49           1000               49000      20.01444643        21.16056575
   50            649               49649      19.44622083        21.13815624


North:
Group      Group Count      Cumulative Count      Group Average      Running Average
    1             1000                  1000        21.05338306          21.05338306
    2             1000                  2000          20.939374          20.99637853
    3             1000                  3000        21.22748723          21.07341476
    4             1000                  4000        20.88183903          21.02552083
    5             1000                  5000        21.07361803          21.03514027
    6             1000                  6000        19.85103419          20.83778926
    7             1000                  7000        19.31846901          20.62074351
    8             1000                  8000        19.57260282          20.48972592
    9             1000                  9000        19.60056077          20.39092979
   10             1000                 10000         19.6981418          20.32165099
   11             1000                 11000        18.78470358           20.1819285
   12              592                 11592        19.85905092          20.16543924


South:
Group      Group Count      Cumulative Count      Group Average      Running Average
    1             1000                  1000        18.51185353          18.51185353
    2             1000                  2000        18.28303994          18.39744674
    3              602                  2602         18.5670407          18.43668408


West:
Group     Group Count     Cumulative Count     Group Average     Running Average
    1            1000                 1000       16.43306507         16.43306507
    2             173                 1173       18.77182159         16.77799676

Feel free to do your own analysis and make your own interpretations.

What I see is that the averages remain pretty consistent from the very beginning, with all sectors having a different average. The one trend I DO see is that in east and north, the kill averages were increasing for the first few groups, before falling linearly with all groups thereafter. I'd wager that this is due to most of the best players stopping praccing, reducing the kill average accordingly.

[Entr0py]

It's not about space, 2 vs 2 happens everywhere especially in east.

Actually, space has a huge effect. The easiest way to see this is to consider an extreme example - suppose you put a 3v3 on a totally empty map, position them randomly, and tell them to duke it out for 20 minutes. I guarantee you that there will be far fewer kills over that 20 minutes in that scenario than if you were to place the 3v3 in south, and again fewer in south than if you were to place the 3v3 in east, and again fewer in east than if you were to place the 3v3 in BAB. It takes time to chase and kill an enemy, and what we're really talking here is kills per minute.

[falconeer]

Looks good, We need more West and South data =/

Wish people would practice south more.

Basically your idea of what happened to north implies good players change the average by 2. That's why West at 19 isn't such a bad idea. It's very possible for west or south to have a 19 kill average. Every sector at 19 was a good idea. East and North even have 19 as some of their kill averages.

I said it's not about space. Which means space has an effect, just not a big one. Space doesn't have a big enough effect to reduce kill averages in west to 16. What has more of an effect is if all the good players don't play west. A bigger effect is when we only have 1129 practice entries. Etc. Like I said 2 vs 2 happens everywhere, and its not from across sectors. Tons of 2 vs 2 happening at the core flags for example. One of the tightest spaces in the deathstar. Your extreme example has nothing to do with the deathstar. Which shows between sectors the space is not all that different. And south is the biggest sector, yet it has a higher kill average.

Your argument is that its kill based on minute. But flags bring players together to kill. This is true for all sectors. Whats also true for all sectors is that if you put me LN, South, or Even east, I could run indefinitely, use my specs to avoid death, attach to a teammate not to engage enemy. Especially LN, South and West I can run for the whole 20 minutes and not die. Will I do this? No. Because IN EVERY sector, flags bring players together. We fight for core west flags, we fight for LN flags, we fight for south flags. You can't even say it takes me forever to fly to the flag, because we have portals. This argument on space is very very weak. If all you want to do is flag touch, you'll have no engagement, but if you want to bell flag, you need to kill, have the advantage/upper hand, lock out the sector, defend the flags you have, you have to fight.

[Tembest]

Yeah, I'll try to get to it later this week. I'll show BH/Eridu how to do it for the future, too.

I know actually how to change the base kill value but I'm unsure of the bonus.

Is the kill bonus what bot calls death bonus?
DSB-PracBo> Death Bonus: 5

If it is, I can change it, no problem.

[Entr0py]

Yeah, I'll try to get to it later this week. I'll show BH/Eridu how to do it for the future, too.

I know actually how to change the base kill value but I'm unsure of the bonus.

Is the kill bonus what bot calls death bonus?
DSB-PracBo> Death Bonus: 5

If it is, I can change it, no problem.

Not sure - that might actually be something else. I'll have to look at it. But just changing the values is not a good idea - first I need to create the new league for these new game settings - Indi League 2 or something, and I need to require registration. It will take ~20 minutes to set up. I'll try to do it tonight.

[Tembest]

I was hoping we could actually test those settings for a while instead of assuming they will work fine.
Once a little testing has been done, reset all lifetime stats, and then we can start a new league, called League 1 as Sever proposed.