Elo score boom

[mcompany]

Hmm... I took a second look at the stats, and it seems as though mumpsimus is the only player with a notably higher level of activity. The majority of other players actually still doesn't seem to be all that much better than what they were before the change to matchmaking

[mcompany]

13 bronze (+2)
75 silver (+4)
14 gold (+2)
12 platinum (-4)
17 diamond (+7)
15 master (+5)
13 grand master (+6)

[mcompany]

Better late than never
16 bronze (+3)
115 silver (+40)
12 gold (-2)
12 platinum (±0)
13 diamond (-4)
10 master (-5)
12 grand master (-1)

[pier4r]

Note on the score activity. Since alpha 12.x we have the ghost players (that are well done so far) and almost every time - aside little coincidences - a player plays against a ghost.

So if before there were 300 multiplayer matches, each requiring 2 active deployments (or so), now with the same number of deployments the game gets 600 multiplayer matches.

And that is not bad, since everyone, anytime, even with two days of difference, can compete (if the active player and the ghost are within 400 score points). For this I say it is well done. In particular the ghost choice seems not dumb, or better, seems that the ghost uses the latest active deployment of its related player that resulted in a score within 400 points from the player deploying. That is neat.

[mcompany]

I'm just getting lazier by the day. I really need to get back on a lot of stuff
20 bronze (+4)
137 silver (+22)
18 gold (+6)
18 platinum (+6)
15 diamond (+2)
10 master (±0)
12 grand master (±0)

Full positive week. Nice

[pier4r]

So using a test account after having reinstalled gladiabots I can say that to reach 1100 one has to try a little. The next wall is 1300, then 1500. Passing 1500 requires finally a bit if effort. The elo works well.

I wonder how many hit 1500 and quit.
I remember a self proclaimed great player that copied simple bot of pew, reached 1700 without even understanding why and then said "this game sucks". Lol (eril or something like that was the name)

[pier4r]

A bit late. (though I thought about a scraper of the official stats page. The official stats page has a lot of info if one processes them a bit. Neat)

Code: Select all

9  bronze (-11)
85 silver (-52)
12 gold (-6)
7  platinum (-9)
9  diamond (-6)
10 master (±0)
15 grand master (+3)


So somehow the game was spammed around (reddit or social networks?) then people - even with the new awesome matchmaking where people don't have to wait much anymore, plus the awesome ai editor and so on - behaved like before. I guess the game should point to keep the few that love it, because those few that love it are not that common.

Alpha 12 is amazingly better than the one before for me, in terms of fun that I can extract every half an hour. But the game is still not trivial, requires effort and you know, idle games are the games that makes the most as many people don't want to think.

I remember my colleagues working in IT "I don't want to think much, I want to relax". And I also have to say that the game is super addictive, so if one plays gladiabots without discipline, he does less of all the rest (reading, writing, whatever active activity).

[mcompany]

6 bronze (-3)
94 silver (+9)
9 gold (-3)
10 platinum (+3)
9 diamond (±0)
15 master (+5)
14 grand master (-1)

[pier4r]

3 bronze (-3)
89 silver (-5)
11 gold (+2)
7 platinum (-3)
13 diamond (+4)
12 master (-3)
12 grand master (-2)

[mcompany]

9 bronze (+6)
102 silver (+13)
10 gold (-1)
8 platinum (+1)
13 diamond (±0)
9 master (-3)
14 grand master (+2)

[pier4r]

2 weeks later

12 bronze (+3)
106 silver (+4)
9 gold (-1)
7 platinum (-1)
9 diamond (-4)
6 master (-3)
12 grand master (-2)

[mcompany]

I guess never leave a ghost to do your housekeeping
8 bronze
75 silver
10 gold
5 platinum
12 diamond
6 master
17 grand master

I will say, don't expect me to update this too frequently, since I've mostly retired from this game for now

[pier4r]

woah. Would be nice if other people could collaborate too!

[pier4r]

8 bronze
55 silver
27 gold
9 platinum
14 diamond
8 master
20 grand master

[ElvenMonk]

Ok, are multiple ghosts created for each player? If so, why? Why not to have single player's ghost always updated with latest deployment. Also to prevent inflation the ghost's score should never be replaced with player's current score.

[pier4r]

I am running (very slowly) a simulation. Even without ghosts (but with active players) inflation is unavoidable.

Actually it is not inflation (inflation is when more points are pumped in the system, ghosts allow that), it is more like "the rich get richer".

I am using for starters 20 players starting with an score of 1000, while the matches are decided by their defined "true strength" having at least one player dominating all the others (although not by a large margin). I am running iterations until the top score stabilizes. So far, already with as little as 20 players, the top player reached elo 4000.

For example here

Code: Select all

  modifiers { 0 25 50 75 100 125 150 175 200 225 250 } with possible sign change, uniform.
  7: :s: 206534.7228
  6: :macroIterations: 50.
  5: :innerIterations: 800.
  4: :true_base_strength: { 850. 871. 886. 910. 948. 953. 971. 975. 976. 989. 1022. 1034. 1038. 1047. 1071. 1102. 1131. 1136. 1172. 1214. }
  3: :minScores: { -2233. -2004. -1560. -1335. -903. -602. -556. -148. -106. 229. 564. 936. 948. 1052. 1553. 2181. 2741. 2663. 3221. 3806. }
  2: :maxScores: { -1240. -1187. -440. -440. 136. 480. 458. 704. 719. 1219. 1763. 1937. 1936. 2060. 2272. 3071. 3500. 3712. 4087. 4743. }
  1: :lvAvgScores: { -1692. -1594. -1201. -887. -267. -34. 52. 290. 376. 581. 1172. 1317. 1491. 1690. 1909. 2647. 3156. 3165. 3708. 4120. }


So forget fixing about inflation, that won't be solved (unless - as in economy, scores are much like a sort of hard won currency - "redistributing taxes" are introduced). Focus about distances between scores that are and were still valid.

Before ghosts the point was that players stopped playing (that will be another simulation to do) and so they hold points outside the system and that was a sort of little tax. With ghosts this is not possible at least yet, and I wouldn't even call it a problem.

Why is it not a problem? because people wants to be fixed on the metric "highest absolute value" instead of "biggest distance between players". In my view this means that people try to force the system to follow a poor metric because they did not understand (n1) how the system work. (the system works with distances rather than absolute values)

n1: I am the first to say that I need to learn a lot of nuances about the system - the system is made by scores and matchmaking - but the average understanding of it in my view is dismal.

[pier4r]

anyway stats

16 bronze +8
44 silver -11
28 gold +1
22 platinum +13
16 diamond +2
12 master +4
24 grand master +4

[ElvenMonk]

Where I can read how, when and with what score new ghosts are created?

[pier4r]

Eh it is scattered on the forum. The page https://wiki.gladiabots.com/index.php?t ... king_rules is still not updated by GFX.

The closest I got is this:
viewtopic.php?f=4&t=1316&start=20#p15358

[pier4r]

Ok no one helped me and mcompany to keep track of the player stats. This is something that we will regret as historical records are hugely important to make future decisions. /shrugs

Anyway now that there are ghost, thanks to the kind db export from gfx (for everyone, there is a link in the unavoidable tournament meta) I just checked something about net flow of points. The same I did in the past for the flow of points between score ranges (pre ghosts). See here: http://pier4r.wikidot.com/pierworks:art ... lo-inights

So in short there is a massive trade of points towards the players.

For example checking the week 2018-08-18 to 2018-08-25 one has.

sqlite code

Code: Select all

/*

CREATE TABLE matches (
  id INT UNIQUE ON CONFLICT REPLACE,
  season INT,
  type TEXT,
  resolutionEngineVersion INT,
  creationTime TEXT,
  resolutionTime TEXT,
  missionID INT,
  player1ID INT,
  player2ID INT,
  player1InitEloRating INT,
  player2InitEloRating INT,
  player1JoinTime TEXT,
  player2JoinTime TEXT,
  player1CommitTime TEXT,
  player2CommitTime TEXT,
  player1Ghost TEXT,
  player2Ghost TEXT,
  expectedResult REAL,
  actualResult REAL,
  player1EloDelta INT,
  end_missing_column TEXT
);

*/

select
  sum(player1EloDelta)
from
  matches
where
  player1Ghost like 'y'
  and
  resolutionTime between '2018-08-18' and '2018-08-25'
;

select
  - sum(player1EloDelta)
from
  matches
where
  player2Ghost like 'y'
  and
  resolutionTime between '2018-08-18' and '2018-08-25'
;


result

Code: Select all

sqlite> .read net_flow_players_ghost_points.sqlite
-16882
-12245


So 29127 points moved from ghost to players in the week 2018-08-18 - 2018-08-24.

That is ok anyway. What matters are the score differences.

week 2018-08-11 2018-08-17

Code: Select all

sqlite> .read net_flow_players_ghost_points.sqlite
-27223
-23650


50873 points (there was more activity)

In general the trend is that ghost release points in the system and, as I said, it is ok. One should not be obnoxious about top scores.