Let me get you a few extra processed stats, along with comments on what each one
means.
Effectiveness of each faction
Victory Points per player:
Kurita: 20512/2092 = 9.805
Liao: 5420/685 = 7.912
Rasalhague: 13288/1812 = 7.333
Steiner: 19315/2661 = 7.259
Marik: 8544/1221 = 6.998
Davion: 11236/2019 = 5.565
IS average: 78315/10490 = 7.466
Falcon: 19427/1516 = 12.81
Wolf: 34350/2956 = 11.62
Jaguar: 12632/1110 = 11.38
Ghost Bear: 19157/1862 = 10.29
Clan Average: 85566/7444 = 11.49
The obvious disparity between the VP/player numbers between the IS and Clan factions indicate that on average, Clan players would have dropped 53.9% more often than IS players. To conclude anything as to whether this would have influenced the effect mercenary premades had on the overall results requires more stats from Tina re: how many 12man drops vs how many non-12man drops there were per house. The effects of [228] alone on Kurita would probably have contributed to its massive victory point-per-person advantage.
Let's just say that you probably want to stay away from Davion and Ghost Bear, though, while Falcon and Kurita probably have the most dedicated players, between their mercs and their house loyalists.
Overall results
The Clans earned 85566 victory points while the Inner Sphere earned 78315 victory points, therefore the Clan:IS win ratio is about 33:30. However, the final result was 36 Clan : 27 IS, which indicates the current way of determining the overall win/loss is still too dependent on the last few hours, and not reflective of the overall situation.
Given that a Counterattack favours IS (Clan loses ground if they lose, they win nothing if they hold territory) and an Invasion favours Clan (IS loses ground if they lose, they win nothing if they successfully defend), there being 8.56% more Counterattacks than Invasions means the mode favoured IS more - which is probably due to the system allocating Counterattacks and Invasions by the ratio of nodes on the map (supported by how on average the clans held 10% more nodes than IS - reasonably close to the 8.56% mode bias) - the more nodes you possess, the more Counterattacks occur and the fewer Invasions occur. While on paper this is a good thing as it results in closer end-results, in matchmaking this is an unfair thing if one side ONLY attacks and the other ONLY defends, as whichever side is leading will be given an exponentially increasing disadvantage.
Incidentally, as more Clan mechs were destroyed and IS players dealt more damage, the main reason for their greater victory ratio in spite of this was Clans having exclusive access to the Invasion mode. In Counterattack, to win, both sides have to have a kill lead (attackers also need to destroy Omega in addition to this) - but in Invasion, kill leads are irrelevant. Given more clan mechs were destroyed, the Clan victories were not due to base domination, but due to the destruction of the Omega objective.
It might be arguable that in the absence of an Omega kill victory condition the Clans would have had a greater kill ratio than they have at present due to the partially suicidal nature of Omega rushes, but the high deployment of Stormcrows and the kill ratios indicate that Invasion Omega Stormcrow rushes were a decisive factor behind the overall Tukayyid win, in spite of a matchmaker system which favoured the IS - and if the matchmaker was more fair, the overall number of victories by the Clans would have been even greater. The 'final' decision of win-loss, however, was not linked in any way to the overall victory rate, and is dependent entirely on the events of the last few hours of the event.
Ghost Drops
Player-Drops per faction:
The number of mechs deployed in total divided by 4 is the number of players involved per side multiplied by the number of matches, as each player can bring 4 mechs.
Clan = (114478+205358+304425+44751)/4 = 167253
IS = (134890+144825+240195+149310)/4 = 167305
Given that there should be 12 players per side in a match, dividing these, we get a total of:
Clan = 167253/12 = 13937.75 matches.
IS = 167305/12 = 13942.08 matches.
Possible reasons why these numbers are not integers is that some matches start with less than 12 players per side because of disconnects, and one side may not even have any players at all in a ghost drop. As the number of players are about 52 different, we might be able to assume that at maximum, 4 ghost drops occurred out of 13942, or an insignificant rate of 0.03%.
Also, the earlier official base stat declared a total of 334558 drops, where 334558 also happens to be the total number of pilots dropped if we add the Clan and IS numbers above. For obvious reasons, the number of pilots dropped is not the same as the number of drops - it's off by a factor of 24.
With a total of 13942 matches the total number of destroyed mechs is 961974, which indicates an average of 69 mechs destroyed in total per match between both sides. Given a total of 96 mechs are brought in, 72% of mechs used in a match are destroyed.
IS-Clan/Mech Balance
Also, the large dominance of Thunderbolts, Stalker-4N, Raven-3L, Stormcrows, Hellbringers and Timberwolves should probably signal to
some people than
some changes should be made. Just look at the numbers of mechs used, bring up the quirks for mechs like the Kintaro and Gargoyle, and bring down some of the quirks for the top few, hmm?
The EXTREMELY SMALL number of Clan Lights and Assaults (especially Clan Assaults) deployed should also send certain signals regarding the effectiveness of Clan Assaults and Lights in an environment where only overall tonnage matters.
Given that close to zero ghost drops occurred from earlier estimation, we can assume that the damage parity between IS and Clan signals that the tech trees' relative damage per second in practice for the most used mechs are in balance when IS is given a 1/24 tonnage advantage. Works well for CW, but it would conversely indicate the Clan variants are still more powerful per ton when used in a public drop context for the most used mechs. Nobody can conclude anything about UrbanMech vs Gargoyle really from these results given that nearly nobody used them.
Given that 5% more friendly fire occurred for Clan pilots... you draw your own conclusions. Could be a weapon duration thing, could be a lack of fire discipline thing, could be a lack of piloting discipline thing.
Map Balance
Given the map rotation Cold=>Hot=>Purple=>Black=>Green on 63 slots, and assuming all slots have an equal chance to be contested, the number of times a map is played is inversely proportional to the time it takes per map. Correcting for the number of slots each map possesses, we thus have the time per map ratios:
Grim Portico = 65788/12 = 5482
Boreal Vault = 71222/13 = 5479
Emerald Taiga = 65149/12 = 5429
Sulphurous Rift = 66285/13 = 5099
Hellebore Springs = 66114/13 = 5086
Remember that a higher number means faster games - which implies the games are less balanced (as balanced matches conclude with timeout or close to it, while 1-way stomps end quickly). These numbers indicate Sulphurous Rift and Hellebore Springs are the two most balanced maps, while Portico, Boreal and Taiga are the most one-sided maps - which should surprise exactly nobody who played in this event. Would be interesting if Tina revealed the Invasion:Defense:Counterattack:Hold territory win/loss ratios on each of these maps, though I'd already predict in advance that Boreal and Portico favour defenders of both types, while Taiga favours attackers of both types.
Weight Class Balance
Weight class longevity:
IS Light: 102854/134890 = 76.3%
IS Medium: 112512/144825 = 77.7%
IS Heavy: 182823/240195 = 76.1%
IS Assault: 112456/149310 = 75.3%
Clan Light: 89270/114478 = 78.0%
Clan Medium: 156883/205358 = 76.4%
Clan Heavy: 239852/304425 = 78.8%
Clan Assault: 36113/44751 = 80.7%
Given the numbers of mechs taken includes the whole dropdeck (which is supported by how all the total numbers per side sum to exact multiples of 4), this % indicates both the relative willingness of a player to use the mech as well as how likely it is to be survive - high percentages mean these mechs are MORE likely to die.
Note in particular how 80.7% of all Clan assaults brought were destroyed, and how the destruction rate numbers are higher for every class on Clan relative to IS except mediums... then look at the rates at which Stormcrows were used. You can draw your own conclusions from there.
Interestingly there is a stat irregularity here as all of the mech classes are destroyed at a rate greater than 72%, but the overall number of mechs destroyed by match is 72%.
New Player Experience For CW
The challenge distribution was as follows:
8166 players got up to 10 points
2416 got 11-20 points
1214 got 21-30 points
1108 got 31-40 points
2767 got 41-50 points
For a total of 15671 players with points.
Since the total number of players was 17226, 1555 players played in the event and got zero challenge points (ergo, zero victories. In other words, 9% of all players got no points and decided to stop playing in the event from the outset, and 56% of all players stopped within the first 10 victory points - which means the
minority of players are those who actually went further in the event. That indicates some work may have to be done to keep players actually playing CW - some factors, possibly the queue, possibly the matchmaking and possibly the game mode itself is turning off more people from playing than not.
Further Stat Requests For Ms. Tina Benoit
- Number of perm contract players vs number of lone wolves vs number of merc corps on temporary contracts for each faction
- Number of IS players vs number of Clan players per Challenge point tier
- Number of 12-man vs 12-man drops vs number of 12-man vs non-12 man drops vs number of non-12 man vs non-12 man drops, and the victory ratios of each
- Victory ratios per map per objective type
- Average kills per team per map (low kill numbers on a map on an Invasion team for instance may indicate a map is liable to be stormcrow/ice ferret/IS light-rushed)
- Complete list of all Merc Corps involved in each faction, the number of victory points contributed by each, the total number of games played by each as well as the number of active players per merc corp (where active = at least 1 game played the whole weekend).
Tl;dr Section
- Nerf SCR, TDR, TBR, STK-4N, etc. Buff AWS, VND, KTO, GAR etc.
- Stormcrow rush was major tactic here in the same way IS Light rush was major tactic in last event. Consider buffing OGen and Omega HP.
- Clans outperformed IS on the whole.
- Clans had much lower waiting time; more data needed to conclude whether this meant Clan premades had more relative power than IS premades. but it definitely made it easier to get 50 CP on Clan.
- More than half the population quit within the first 10 CP.
- 9% quit after losing their first few game(s).
- Grim Portico, Boreal Vault and Emerald Taiga are unbalanced.
- Clan Assaults are underpowered.
- Davion... sigh.
- Falcon actually performed best on Clan, not Wolf.
- Nearly no ghost-drops on Tukayyid.
- The last-hours effect method of determining a win is completely unrelated to overall performance, and should be scrapped.
Congratulations to the Clans for winning Tukayyid overall.
Additional edit: Turret/Dropship potential
Ace Selin, on 03 May 2015 - 11:02 PM, said:
This is a false stat. IS were defending so should have found kills easier for 2 reasons, they had the help of turrets and there were Clan rushes on Omega where Clans died 12-0. That means during the actual fights, the Clans won more often than not.
Actually this is probably the source of the stat irregularity I was wondering about. My best guess is that PGI counted 'number of mechs destroyed' by the number of player kills, while the number of destroyed per class is by the actual number of mechs destroyed.
Hayashi, on 02 May 2015 - 02:34 AM, said:
With a total of 13942 matches the total number of destroyed mechs is 961974, which indicates an average of 69 mechs destroyed in total per match between both sides. Given a total of 96 mechs are brought in, 72% of mechs used in a match are destroyed.
Weight class longevity:
IS Light: 102854/134890 = 76.3%
IS Medium: 112512/144825 = 77.7%
IS Heavy: 182823/240195 = 76.1%
IS Assault: 112456/149310 = 75.3%
Clan Light: 89270/114478 = 78.0%
Clan Medium: 156883/205358 = 76.4%
Clan Heavy: 239852/304425 = 78.8%
Clan Assault: 36113/44751 = 80.7%
Interestingly there is a stat irregularity here as all of the mech classes are destroyed at a rate greater than 72%, but the overall number of mechs destroyed by match is 72%.
So we have a new processed statistic here I missed on my first pass. The number of clan mechs destroyed by players is 481,187, but the total number destroyed is 522,118 - so 92.16% are destroyed by players and 7.84% are destroyed by turrets and dropships combined. The number of IS mechs destroyed by players is 480,787, but the total number destroyed is 510645 - so 94.15% are destroyed by players and 5.85% are destroyed by dropships alone (clans had zero access to turrets).
Attackers always have access to dropships with large lasers while defenders always have access to dropships with medium lasers, so regardless of which side you're on the dropship kill rate can be assumed to be approximately the same. That means turrets had an overall effect of destroying 1.99% of all players - and since Counterattack has no turrets, correcting for the overall occurrence rate of Invasion maps, the estimated potential of turrets is to kill 4.15% of targets in an Invasion map, or 2 mechs per match. It probably shouldn't surprise anyone that turrets are less effective than dropships.
Edited by Hayashi, 04 May 2015 - 01:13 AM.