Failure to Match: 100 PvP Matchups from Divine Champions

2»

Comments

  • Unknown
    Unknown
    edited February 2015
    The problem with the system is that it's a measure of how strong you are plus how much you bribed the game during a particular event. As such, this is not a reliable number to carry over to the next event because there's no way to predict how much money you'll spend on the next event. Stuff like shields/boosts basically amount to bribes and that's something most even P2W games don't do, so it's not surprising that any rating that accounts for bribes doesn't really work. That is, take MTG which is very P2W in certain format for sure. But the P2W happens before the event starts. If you sign up for a Vintage tournament you'd be well advised to spend a lot of money on the cards that are worth hundreds you need to be competitive, but this is before the event. Once the event starts you don't get to call a timeout by slipping $10 to the judge and then secretly sideboard some cards that wasn't already allowed to win. The P2W is supposed to be in the preparation of building the best roster, not coming up with about $5 worth of shields that you don't even necessarily have to pay any money to accumulate. People certainly have spent hundreds or more on their roster, but in the end none of that even matters very much without that extra 500 HP of shields.
  • Stax the Foyer
    Stax the Foyer Posts: 941 Critical Contributor
    edited February 2015
    A lack of appropriate targets for newer players might be because there's not a critical mass of newer players playing throughout the tournament. You're always going to have some people joining late and steamrolling through the lower point ranges, but if you have enough lower-level players bouncing around, the targets will be there.

    PvP can be intimidating to new players, and in an essentially zero-sum point battle, it's probably tough for newer players to feel like it's worthwhile. Some incentive to get people to regularly participate in pvp, especially at lower levels, could help a lot.

    If there was a daily goal to play in 10 pvp matches (or whatever appropriate number), maybe with a progress bar and a trivial reward for accomplishing it, people would do it, because you'd be able to fill that progress bar even if you were getting smacked around. Even a standard token or a small amount of ISO would be appealing to new players, and it could significantly increase the pool of available targets for lower leveled players.
  • simonsez
    simonsez Posts: 4,663 Chairperson of the Boards
    Vynyv wrote:
    Describing the state of the current system does not justify its merits.
    As I said in the post prior to the one you quoted, I'm by no means a PvP defender. And if you're looking for the game to change the point structure, also as I said in the previous post, this is not an argument around matchmaking, it's an argument against the way PvP is designed.

    Aside from that, I also believe our issue is an inability to agree on what it means to be at equilibrium.
  • Unknown
    Unknown
    Vynyv wrote:
    In response to the Upcoming Versus Matchmaking Test thread, I decided to keep track of a string of 100 PvP matchups in the current Divine Champions PvP event.

    My in game name is Vyn, and the end time for my slice is Wednesday (Feb 4th) at noon (EST). The data below was recorded around 5pm (EST) on Monday (Feb 2nd).

    A screenshot of the first page of my roster can be found here.

    Due to boosting, the actual top level characters on my roster are as follows:
    • L111 5/1/5 Thor 2*
    • L111 4/4/3 Ares
    • L80 0/2/1 Loki
    • L74 3/5/5 Daken
    • L74 1/0/1 Elektra
    • L60 3/1/4 Storm 2*
    • L60 0/1/0 Thor 3*
    • L53 5/3/5 Storm 1*
    • L50 3/5/4 Moonstone
    • L45 4/1/4 Wolverine 1*
    • L45 2/2/3 Human Torch 2*
    • L40 5/5/3 Iron Man 1*
    The average level of my top 3 characters is 100.7. Since it would be rather inefficient to use both 2* Thor and Ares at the same time, I normally play with Daken/Ares/Loki instead, with average level 88.3 in this event. Keep in mind that my Loki has very few covers too, so even that level vastly overstates my roster strength.

    Now for the 100 matchups, the columns in the table below are:
    • No.: Match number
    • Pts: My point total at time of match
    • Rnk: My rank at time of match
    • Rwrd: Point reward for victory
    • L1/L2/L3: Level of enemy's three characters
    • Avg: Average level of enemy's three characters
    My point total drops periodically because I had to tank battles to conserve iso-8.
    No.   Pts	Rnk	Rwrd  L1	 L2	 L3	 Avg	  Notes
001	229	143	022	141	094	141	125.3	
002	229	143	032	141	127	153	140.3	
003	229	143	024	120	080	120	106.7	
004	229	143	027	141	134	141	138.7	
005	229	143	035	125	249	191	188.3	
006	225	146	033	237	156	156	183.0	
007	225	146	034	228	209	166	201.0	
008	225	146	022	052	040	043	045.0	Attacked
009	247	126	026	094	141	141	125.3	
010	247	126	034	249	201	220	223.3	
011	247	126	025	094	156	142	130.7	
012	247	126	020	052	040	043	045.0	Attacked
013	266	114	037	230	166	166	187.3	
014	266	114	020	141	166	141	149.3	
015	266	114	019	094	141	117	117.3	
016	266	114	029	215	166	252	211.0	
017	266	114	037	141	141	094	125.3	
018	262	118	023	141	094	094	109.7	
019	262	118	029	141	152	094	129.0	
020	262	118	024	143	153	153	149.7	
021	262	118	035	141	075	141	119.0	
022	262	118	044	228	270	405	301.0	
023	260	120	029	141	141	141	141.0	
024	260	120	025	405	225	270	300.0	
025	260	120	026	141	094	141	125.3	
026	260	120	024	143	190	094	142.3	
027	260	120	036	141	117	141	133.0	
028	256	123	028	141	141	104	128.7	
029	256	123	037	094	209	141	148.0	
030	256	123	035	141	131	094	122.0	
031	256	123	023	198	166	210	191.3	
032	256	123	036	249	166	166	193.7	
033	251	128	028	168	141	161	156.7	
034	251	128	022	141	094	094	109.7	
035	251	128	039	141	177	141	153.0	
036	251	128	026	110	166	270	182.0	
037	251	128	034	141	141	141	141.0	
038	246	134	029	141	094	114	116.3	
039	246	134	027	129	099	075	101.0	
040	246	134	029	105	094	127	108.7	
041	246	134	028	094	141	141	125.3	
042	246	134	023	141	117	102	120.0	
043	241	136	029	114	141	094	116.3	
044	241	136	027	141	094	094	109.7	
045	241	136	028	141	156	153	150.0	
046	241	136	033	141	091	094	108.7	
047	241	136	030	141	104	094	113.0	
048	235	141	029	111	146	141	132.7	
049	235	141	025	249	141	166	185.3	
050	235	141	036	141	094	105	113.3	
051	235	141	022	086	060	081	075.7	Attacked
052	257	127	032	141	094	122	119.0	
053	257	127	026	141	094	094	109.7	
054	257	127	029	156	140	153	149.7	
055	257	127	022	127	094	094	105.0	
056	257	127	023	106	141	141	129.3	
057	250	133	029	166	141	249	185.3	
058	250	133	023	105	141	114	120.0	
059	250	133	032	141	094	141	125.3	
060	250	133	028	131	153	094	126.0	
061	250	133	033	141	094	141	125.3	
062	245	137	030	141	141	078	120.0	
063	245	137	025	189	077	116	127.3	
064	245	137	026	141	156	078	125.0	
065	245	137	040	249	229	166	214.7	
066	245	137	030	141	141	141	141.0	
067	237	144	028	116	141	141	132.7	
068	237	144	026	141	134	154	143.0	
069	237	144	033	249	230	213	230.7	
070	237	144	021	094	141	063	099.3	
071	237	144	025	105	105	105	105.0	
072	231	153	032	156	152	130	146.0	
073	231	153	031	153	135	220	169.3	
074	231	153	026	127	134	131	130.7	
075	231	153	023	156	130	229	171.7	
076	231	153	022	141	100	141	127.3	
077	224	157	023	117	141	114	124.0	
078	224	157	033	171	141	117	143.0	
079	224	157	022	140	141	140	140.3	
080	224	157	022	050	050	053	051.0	Attacked
081	246	141	022	105	063	120	096.0	
082	246	141	023	141	141	117	133.0	
083	246	141	038	249	166	213	209.3	
084	246	141	043	206	141	153	166.7	
085	246	141	029	141	141	141	141.0	
086	240	145	025	087	095	141	107.7	
087	240	145	027	094	141	141	125.3	
---	232	155	025	249	166	176	197.0	Retaliation node
088	232	155	029	141	094	114	116.3	
089	232	155	023	141	086	094	107.0	
090	232	155	030	147	166	249	187.3	
091	227	159	026	141	141	166	149.3	
092	227	159	025	122	094	131	115.7	
093	227	159	027	141	127	117	128.3	
094	227	159	027	141	094	094	109.7	
095	227	159	027	141	114	141	132.0	
096	220	166	040	180	127	141	149.3	
097	220	166	030	191	166	166	174.3	
098	220	166	024	141	094	094	109.7	
099	220	166	030	141	094	141	125.3	
100	220	166	033	094	122	141	119.0
    The number of matchups with average level less than or equal to the average level +X% of my top 3 characters was...
    • Top 3 Avg Lvl +00%: 6
    • Top 3 Avg Lvl +10%: 20
    • Top 3 Avg Lvl +20%: 33
    • Top 3 Avg Lvl +30%: 54
    • Top 3 Avg Lvl +40%: 62
    • Top 3 Avg Lvl +50%: 77
    The number of matchups with average level less than or equal to the average level +X% of the 3 characters I actually play with was:
    • Team Avg Lvl +00%: 4 (I attacked all 4)
    • Team Avg Lvl +10%: 5
    • Team Avg Lvl +20%: 9
    • Team Avg Lvl +30%: 22
    • Team Avg Lvl +40%: 34
    • Team Avg Lvl +50%: 55
    In summary, I had a total of 4 matchups with opponents that I completely outleveled and just rolled over. Aside from that, there were only 2 or 3 of matchups where I had any chance of victory, and only if I used some serious power boosts. Needless to say, this makes for a frustrating PvP experience when one has to spend far more time skipping and tanking battles than actually playing in one.

    Another point worth noting is that during this entire experiment I was only attacked once. If I was given difficult matchups due to over performance, then I ought to be hit more often so I can drop my points/rank and receive more reasonable matchups.

    Based on this record of matchups, I think it is fair to say that the matchmaking algorithm for Divine Champions is mostly ineffective. Hopefully this event was still running the old algorithm, because the results here sure look no different than that from previous PvP events.

    Edit 1: The average level of the matchups was 140, when I ran a team with an average level of 88.
    Edit 2: Added explanation of Rwrd column


    Without having time to read everyone's comments, mid 200s seems about right for a roster that doesn't even have maxed 2*s

    Even 3* rosters aren't getting 1000 now, probably closer to 750 as 4* have started to take over

    As you build up more you'll get higher scores with fewer attacks.

    Also keep in mind the alternate pvp's with weird setups always don't' get accurate readings for difficulty. - see Nefarios foes, Lonestar, BoP etc
  • dkffiv
    dkffiv Posts: 1,039 Chairperson of the Boards
    Vynyv wrote:
    My Team:
    L111 4/4/3 Ares with 6,878 health
    L74 3/5/5 Daken with 3,100 health
    L80 0/2/1 Loki with 2,913 health
    Total Health: 12,891
    Max damage efficiency: Onslaught at 267 damage per green
    Green match damage: 61
    Max damage per turn: (61 + 267) * 3 = 984

    Enemy Team (33 point reward):
    L371 5/3/5 4Thor with 22,491 health
    L201 5/3/5 Loki with 7,152 health
    L248 5/5/3 X-Force with 10,069 health
    Total health: 39,712
    Average match damage: 89.5
    Average match-3 damage per turn: 268.5

    You aren't supposed to be fighting that team. That guy is in the process of climbing. Use the skip button until you find a decent team or have to pay the skip tax. Instead of paying 10 ISO to skip at that point join a match and suicide (if you haven't read up on proper tanking, your team has to die, not just retreat). If you're doing that properly you should eventually find 2* rosters.
  • dkffiv
    dkffiv Posts: 1,039 Chairperson of the Boards
    You should also take into account that the PvP season and a major PvE just ended and most people with 2* rosters are probably doing those rather than an off-season PvP with terrible rewards.
  • Unknown
    Unknown
    simonsez wrote:
    And if you're looking for the game to change the point structure, also as I said in the previous post, this is not an argument around matchmaking, it's an argument against the way PvP is designed.
    We are probably just splitting hairs over terminalogy here, but I am very much making an argument against the match making system as implemented in Divine Champions. My position is that players should always have access to reasonable matchups. I believe this reflects the intended design of the PvP system, though it is certainly not what is happening in practice.
    simonsez wrote:
    Aside from that, I also believe our issue is an inability to agree on what it means to be at equilibrium.
    I have taken equilibrium to represent the point value or rank that accurately reflects a player's ability to win battles. It may be that you are using equilibrium to represent the point value or rank that a player would naturally acquire under the current system. Unless you have a vastly different conception of equilibrium however, these definitions are functionally equivalent under the current context. Where we seem to differ is that I believe players at equilibrium should still have access to reasonable matchups, whereas you seem to disagree with this.
    LoreNYC wrote:
    Without having time to read everyone's comments, mid 200s seems about right for a roster that doesn't even have maxed 2*s
    The criticism here is not a complaint on poor placement, but the lack of reasonable matchups to keep players entertained. Please do read the other comments when you have more time however.
    dkffiv wrote:
    You aren't supposed to be fighting that team. That guy is in the process of climbing.
    Exactly! Why match players up with a team they are not supposed to fight? Let the climbers hit low level players to their heart's content, but don't ask low level players to skip through 100 battles just to find 4 even lower level players to roll over.
    dkffiv wrote:
    Use the skip button until you find a decent team or have to pay the skip tax. Instead of paying 10 ISO to skip at that point join a match and suicide (if you haven't read up on proper tanking, your team has to die, not just retreat). If you're doing that properly you should eventually find 2* rosters.
    As I described in my original post, this is precisely what I was doing during the experiment, and the results were abysmal. I spent far more time skipping and tanking than fighting in battles, and the battles weren't even that interesting because they were against players much lower level than me.

    Edit: I will add a TLDR summary to the original post for folks with limited time.
  • dkffiv
    dkffiv Posts: 1,039 Chairperson of the Boards
    The comment was in regards to if you knew how to properly tank or not. A lot of new players think you just need to join a battle and then lose (the difference between actually dying and just retreating. Also remember that its a measure of how many hit points you lose so tanking with a level 1 or 15 isn't very effective). When tanking you need to constantly lose, not just stop as soon as you see what you perceive to be a reasonable target. You need to lose to those teams too and then you'll start getting a string or 4 or so decent matches at a time.

    If anything you sound like you prefer the old system where high level teams can just beat you whenever. As is you're mostly invisible to 3* rosters so any points you earn are kept until you hit 400+.
  • Unknown
    Unknown
    Since we're talking about equilibrium state which implies no shields and that you also have to walk away from the game at some point (constantly playing is not really what I think of as equilibrium unless you're supposed to play continously for the duration of an entire event), I'd be very surprised if someone can have a score of 900 in an equilibrium state. That dictates what everyone else's score can be. I'm estimating there are at least 2 tiers of 'no chance of winning' above the original poster. That is, there's some guy he has no chance of beating who in term has no chance of beating the top roster. So where are you supposed to put these 3 drastic levels of performance? You can't even have 500 points because that'd imply you've a point spread like:

    500: not even maxed 2*
    700: some guy with max 3* that can trivially defeat the guy at 500
    900: the top roster at equilibrium who can trivially beat the guy at 700

    Does this structure even make sense? I don't think 200 points is enough to separate a tier of 'trivially win'. In most events, when the points have time to settle, I find guys within 200 points within me are definitely not trivial.

Categories