Return of Reckoning

Hey all,


Since I was thoroughly bored this evening, I figured I'd take a look at the Ranked leaderboards and compare some numbers.


What I ended up doing was take the amount of players of each class that as per 30-07-2021 either,

a. had a positive MMR
b. had 0 MMR but at least a single victory.

This number I divided by the total amount of players of that class on the leaderboards.


Disclaimer: there is an error in the ranking list of the leaderboards, whereby every page it skips five places. (First page ends at #15, but the next page starts at #21). As a result the rankings shown on the leaderboards are incorrect. I rectified this, however if you notice any errors please let me know and I will change the numbers accordingly.


Here are the results:

1. Ironbreaker: 61/77 = 79.2%

2. Runepriest: 76/104 = 73.1%

3. Warrior Priest: 95/133 = 71.4%

4. White Lion: 33/47 = 70.2%

5. Knight of the Blazing Sun: 79/114 = 69.3%

6. Slayer: 69/101 = 68.3%

7. Blackguard: 55/81 = 67.9%

8. Swordmaster: 43/64 = 67.2%

9. Chosen: 76/114 = 66.7%

10. Disciple of Khaine: 84/129 = 65.1%

11. Zealot: 73/114 = 64.0%

12. Choppa: 46/74 = 62.2%

13. Marauder: 45/74 = 60.8%

14. Shaman: 64/106 = 60.4%

15. Black Orc: 39/66 = 59.1%

16. Witch Elf: 43/73 = 58.9%

17. Squig Herder: 30/53 = 56.6%

18. Shadow Warrior: 24/43 = 55.8%

19. Witch Hunter: 29/53 = 54.7%

20. Archmage: 33/64 = 51.6%

21. Engineer: 16/39 = 41.0%

22. Sorceress: 12/31 = 38.7%

23. Bright Wizard: 13/37 = 35.1%

24. Magus: 6/23 = 26.1%

_______________________________________________________________________________________________________________________________________________



Here is the same measurement, but it excludes all players who have played less than 10 games. This was added as response to database developer Kabocu's comment.


1. Ironbreaker: 39/77 = 50.6%

2. White Lion: 23/47 = 48.9%

3. Runepriest: 50/104 = 48.1%

4. Slayer: 48/101 = 47.5%

5. Blackguard: 38/81 = 46.9%

6. Knight of the Blazing Sun: 38/81 = 46.5%

7. Swordmaster: 29/64 = 45.3%

8. Warrior Priest: 56/133 = 42.1%

9. Zealot: 47/114 = 41.2%

10. Shaman: 40/106 = 37.7%

11. Chosen: 41/114 = 36.0%

12. Discipline of Khaine: 45/129 = 34.9%

13. Witch Elf: 24/73 = 32.9%

14. Archmage: 21/64 = 32.8%

15. Marauder: 23/74 = 31.1%

16. Black Orc: 19/66 = 28.8%

17. Choppa: 21/74 = 28.4%

18. Witch Hunter: 14/53 = 26.4%

19. Shadow Warrior: 11/43 = 25.6%

20. Sorceress: 7/31 = 22.6%

21. Squig Herder: 12/53 = 22.6%

22. Engineer: 8/39 = 20.5%

23. Bright Wizard: 6/37 = 16.2%

24. Magus: 2/23 = 8.7%




What does the forum community think? Is there anything that can be gleaned from this?

Please share your thoughts.

Only the best intentions here, truly.

wargrimnir wrote: ↑Fri Jul 30, 2021 8:22 pm Only the best intentions here, truly.

Indeed! Nerf order.

Lies, damned lies and statistics.

But, seriously, why did you combine positive MMR with just a single victory? Seems like a weird split as it conflates reasonable success (positive MMR) with any success (one victory) with participation (tried it at all win or lose).

Uh, kinda strange evaluation. IB place 1? Best IB is place 48 lol...

Here is my ranking:

bad: most Magi, most Sorcs, most Engis, most AMs, most SWs, most WHs, pretty much every BW
superior: slayer
good: all other classes

how unexpected, lots of blue in the top

how strange, when each order class is superior to its destro counterpart

Grasthard wrote: ↑Sat Jul 31, 2021 3:21 am how unexpected, lots of blue in the top

how strange, when each order class is superior to its destro counterpart

Sorc>bw
Msh>wl
Rsh>rsw
Msh>asw
Shammy>am
Dok>wp
We>wh

Mara><wl
Magus><engi
Ib><bg
Zealot><rp

Sm>borc
Slayer>choppa
Kotbs>chosen

Aethilmar wrote: ↑Fri Jul 30, 2021 11:37 pm
But, seriously, why did you combine positive MMR with just a single victory? Seems like a weird split as it conflates reasonable success (positive MMR) with any success (one victory) with participation (tried it at all win or lose).

I have been watching the leaderboards for a while and noticed the large amounts of people who basically could not participate, not being able to get above 0 MMR or even take a single win. I find that interesting since it seems to heavily correlate with class, and also worrying because player participation (or the lack thereof) is what has stopped Ranked from being a success.

If the match making was working well, one would expect that these low skill players (MMR =/= Skill, but you get my meaning) would eventually accumulate at a certain MMR, where they would be able to find equal matches. This doesn't seem to happen, however.

The choice for positive MMR or at least a single victory was because some players have 0 MMR, but still participate and take many wins (they take more losses, but I still count that as participation).

But the measurements could be repeated with different parameters to see if it results in a different outcome. This is the only measurement I did, but I'm genuinely interested in more statistics concerning Ranked.

Caduceus wrote: ↑Sat Jul 31, 2021 2:16 pm

Aethilmar wrote: ↑Fri Jul 30, 2021 11:37 pm
But, seriously, why did you combine positive MMR with just a single victory? Seems like a weird split as it conflates reasonable success (positive MMR) with any success (one victory) with participation (tried it at all win or lose).

I have been watching the leaderboards for a while and noticed the large amounts of people who basically could not participate, not being able to get above 0 MMR or even take a single win. I find that interesting since it seems to heavily correlate with class, and also worrying because player participation (or the lack thereof) is what has stopped Ranked from being a success.

If the match making was working well, one would expect that these low skill players (MMR =/= Skill, but you get my meaning) would eventually accumulate at a certain MMR, where they would be able to find equal matches. This doesn't seem to happen, however.

The choice for positive MMR or at least a single victory was because some players have 0 MMR, but still participate and take many wins (they take more losses, but I still count that as participation).

But the measurements could be repeated with different parameters to see if it results in a different outcome. This is the only measurement I did, but I'm genuinely interested in more statistics concerning Ranked.

It's not about class, it's about correct spec.. This statistic show nothing, for example 90% SW trying ranked are skrimish speced, when they should be assault, AoE BW/sorc/mara etc, this is normal.
Engi / magius, which can be one of the best classes for rankings in the right hands are in bottom in your statistic.

if you add all classes that have 1 victory together. 571 order 573 destro, the only useful information from this statistics.

Anyway ranked is dead so who care:D

And it ain't the right spec either..at least exclusivly. Especially with Solo Ranked there are to many variables.In group ranked you could say something about the overall team that is queueing .. There is nothing of scientific worth in this. I'm sorry having to neglect the time spend on crunching these numbers. Must have taken some time

Just starting at the actual player/you would keep me busy listing the individual variables they bring into the ranked scenario.

Starting from a fairly broad..


.. knowledge of the game(detaunts, immunities, guard, ap cost, casting animations) by individual.
.. over knowledge of their class and the grouped players classes OR knowledge of the enemy(one might know the sorc being faced off against well but not the SH or Magus)
.. over the mood the player is in.
.. over the way the grp syngergises
.. gear
.. renown spec in regards of friends as well as hostiles
.. using voip to communicate
.. individual ability to adapt towards each others playstyle
.. aso.
.. aso.
..overall balance of classes.
.. general viability of a class for a certain scenario.
.. aso.
.. aso.



Just one deviation by one player might cause a loss or the resulting flame in chat i.e. 1 heal does not know the value of a detaunt or both healers start prematurely res at the same time which both get interrupted/staggered/friend killed on res @20% as for the other way around a tank/dps that does not know the res animation wont ever taunt it to prevent

I wont spend more time on this... Since you just posted the numbers and did not detail what you think of them and the goal you had in mind when gathering the data I'ld say that these results did achive one thing -> Filling the time you had to waste with something.

No offense. Just evaluating the data you posted and putting my perspective on it. The effort is appreciated but pointless..

It is possible to take general conclusions from these numbers -> overall server/class population and choices. Still you need another value to hold it up against.