Jump to content


  • Content Сount

  • Joined

  • Last visited

  • Battles

  • Clan


About IVemo

  • Rank
    Leading Rate
  • Insignia

Recent Profile Visitors

187 profile views
  1. IVemo

    Skill-related matchmaking for more balanced games

    Matter of how much you let players' skill ratings vary in a single game. I advocated as small variation as possible. Ironically, random matchmaking will achieve that potatoes and the rest are evenly spread in games, and almost never will there be a game without a potato. Every game is a similar kind of mess.
  2. IVemo

    Skill-related matchmaking for more balanced games

    Blizz poster says long waiting times are bad and that they cannot meet people's high expectations. Did I miss something? They are themselves using skill-based matchmaking, with 50% matches.
  3. IVemo

    Skill-related matchmaking for more balanced games

    This is going a bit too much out of scope but Google artificial neural network or machine learning. It is a black box. You cannot or would not want to adjust for any external considerations after it has been taught. I guess that is an eye-opener for many. About "where skillbased mm has failed" this I cannot comment on as I do not know these games you refer to or what has happened there.
  4. IVemo

    Skill-related matchmaking for more balanced games

    "Probability to win". They would have to teach the neural net with data from previously played games first. After that: IN: [Ships, players, maybe map&mode] -> OUT: [Probability for team 1 to beat team 2]. It would attempt to find matchups as close to 50% as possible.
  5. IVemo

    Skill-related matchmaking for more balanced games

    A few comments on these: 1. If one breaks apart your sentence you are essentially saying that players cannot necessarily be put in order of skill or that a team's total strength cannot be evaluated correctly. I disagree with this, ELO enhanced with a neural net will do it without a shadow of doubt. It is a common mistake to think that there are some other dimensions, like "how organized your team is", in play when determining the outcome of a match. There are not. A single number, that how likely you are to win, can describe it all. High skill play will look more organized, better, than low skill play by definition. Thus saying, even in chaos there is order. 2&3. A good point, I suppose many would not. They would cry, be as loud as possible on Reddit, threaten quitting, try to punch the devs if they could, etc. It is just a whole lot of nothing. Everyone knows those who perceive themselves as good, part of whose self-identity comes from being good in X are the least likely to quit. More likely it would be an opportunity for WG to monetize this sudden perceived drop in performance, give people an easy way out. Personally, I have played with both systems, skill-based and non-skill-based, and vastly prefer the skill-based matchmaking as that makes me play better, faster. Non-skill-based matchmaking cannot do it because the variety of opponents are more than likely unable to exploit the flaws in my play. I guess this is why I have recently only been playing clan battles.
  6. IVemo

    Skill-related matchmaking for more balanced games

    WG is losing players and money (and not only due to a single reason, by the way) when they are not using a skill-based matchmaking system. I suppose they are aware of this now but were not when the game was made. The why: 1. Assume players gravitate towards games where they have the most "fun" or at least stop playing games where they are not having any. Assume win rate of players correlate with this "fun". 2. If you pair players randomly the result is highly variable win rates. Those good at the game can reach 60%+ win percentages while it is not unheard of to get less than 40% if your play is actually helping the opposing team. 3. With highly variable win rates, what do you think happens next? Yep, starting with those with a 40% win rate decide sooner or later that this is not a game for them and quit playing. This starts a cycle where those at the bottom are always leaving the game, upping the skill requirement for those who are still staying to enjoy the game, to have fun. So, you see, there is a constant push for players to quit the game with a system where players are paired randomly, not according to their skill. With skill-based matchmaking that would change. Bad players' win rates would get increased, up to 50%, and instead of thinking about quitting they might keep playing.. especially if the other games they are playing are not using skill-based matchmaking and are offering them 40% win rates. Such a system should also be beneficial for those skilled at the game, by the way. Even though their win rates would collapse, down to 50%, their games should instantly become more organized, logical, interesting as now they would only be playing with their equals. There are other advantages as well. Like, it would have been leaps easier to balance the CVs if you only had players of equal skill playing against each other. Another example, with skill-based matchmaking WG would have got a more diverse gameplay offering, making the game look less repetitive, which would be another reason why player retention should be higher.
  7. IVemo

    Concerning the upcoming CV rework

    I suppose the reason for the unpopularity of playing CVs was more widely those exaggerated needs for micromanagement. To be able to play CVs a player had to pass these "tests" = learn these things: Be able to control planes sufficiently well, so that they do not auto-suicide Learn to drop manually, which should take some 100 or more games to put in muscle memory Learn what kind of anti-drop maneuvers ships are able to perform and how to counter them Figure out what drop patterns are best for each target Learn AA capabilities of ships Look at ship AA values in the loading screen Figure out what level of engagement / attempt to shut down is best vs. each opposing carrier And when you are finally doing all of that you have to ask yourself, where is the fun for doing such a big amount of work? I admit it used to be in sinking boatloads of ships in return, that had no way to stop you, but WG aggressively put a stop to that, wonder why :D. Anyway, my argument is that if a player did not learn all of those things in the list they would have probably quit, which should explain the low popularity. WG could have retained the core idea of the old CV play and do away with all those details requiring too much micromanagement, but they chose to chase a different kind of target audience.
  8. IVemo

    Fara on carrier rework

    Which one would you choose, 1 or 2? 1) Unending chip damage at whatever target, usually the same until it is sunk. 2) Single, bigger strikes, usually at different targets. I don't know why they "listened" to the common CV feedback and chose 1 because I am pretty sure it is not an improvement. Core gameplay loop with 1: Chip damage target -> repeat Core gameplay loop with 2: Single strike target -> Stop enemy CV's strike -> repeat That there are so many obvious issues with 1 kind of tells they may be desperate to make something work and are quietly wishing for tapping the Fortnite/kid audience.
  9. IVemo

    Holiday Lottery 2018 - Try your luck !

    I would like to participate in the lottery.
  10. IVemo

    Should Radio Location be removed?

    Radio Location is fine but Priority Target is not. Or fix Priority Target to only include ships that are sighted.
  11. Hiryu, Taiho, Essex etc. are the most iconic of CVs. Taiho, possibly the best looking ship in the whole game :), will surely become available as a premium if the CV change is liked. I suppose that was half of their intention behind the tech tree change, to produce new premiums cheaply. However, what is more important I am afraid that CVs are likely to stay in the margin if this CV change goes live because the new arcadey take is a leap too big towards its own imaginary ship-planes combat while not offering an addictive enough gameplay in return. Namely, CVs in reality were not about chip damage but single massive strikes, which actually failed at least as much as the fantasy chip damage idea but just for different reasons. After all, CVs' advantage that lead to their dominance in WW2 was not their superior striking power but their striking distance (up to a few hundred kilometers) most of which should be negated here. What I would have liked to have seen was a dumbed down RTS, enabling higher plane counts and more realism tuck in. Likely it would not have brought in the masses to play CVs but that should not have been the main objective in the first place, to offer more of the same. Instead, CVs could have been used to spice things up, e.g., to introduce a new danger to the usual gameflow or to offer a different kind of a game experience where multiple CVs were put in a single game. The old CV design did something like that but I believe it altered the game in wrong places and was just unbalanceable with the lack of a proper skill-based MM. Now thinking about it, maybe CVs were just wrongly tiered, Taiho for example actually being an equivalent of a tier XI ship.
  12. IVemo

    How would you improve the skill of player base?

    Everyone is assigned a rating, describing how well they play the game. Say there are players with ratings 800, 900, 1000 and 1100 looking for a game. The matchmaker would create a match where 800 and 1100 would face the 900 and 1000, giving both sides about an equal chance to win. After the match, winners get their ratings adjusted upwards and losers downwards. However, nowadays as I said it is possible to use machine learning to estimate player skill levels and change their ratings by large amounts when no certainty about player skill has been reached. With a random matchmaker you could have the same match with 1100 and 1000 battling versus 800 and 900 which kind of explains why stompfests are a common occurrence.
  13. IVemo

    How would you improve the skill of player base?

    ELO-based matchmaking is the current game industry standard solution to this problem which, naturally, is not endorsed by WG. Any other solution as per current knowledge is just inferior, though many developers tend to want to hide the fact ELO is utilized. Often because it is believed that players want to see progress when in reality there rarely is any. I would argue that that just makes it that much harder to improve when you do not see what works and what does not. To go on a bit about ELO-mm, the practice of not using it is costing WG money due to many reasons. The main ones are because a pressure is created on the bottom of the player skillpool to exit the game and secondly because game diversity is reduced due to "every game is similar" (there are no high or low skill-only games besides the ranked mode(s)). More elaborate reasoning why there is pressure on the bottom: 1. Assume everyone chooses to play games they excel the most at. 2. If random matchmaking, some people are extra happy with win rates well over 50%. But those high values come with a cost: An equal amount of people must have well below 50% win rates and are not happy playing WoWs. Because there are games that offer at least that 50% to them, they will quit playing WoWs. Thus, a cycle is created where there is always a pressure on the least performing players to say bye bye to WoWs, over time increasing the skill threshold required to be happy playing WoWs. 3. With ELO, everyone's win rate is adjusted to 50%, so none is too happy or too unhappy. This is optimal. TL;DR: I can see why WG does not want to implement ELO-based matchmaking but any reasons, e.g., "more hectic games when high-skill players are kept separated" or the clever "it does not work with 12 player teams" etc. are simply false. To speed up the process, one could just use machine learning to estimate player skill levels.
  14. What WG is planning to do with CVs is not necessarily a range limitation because that would be highly unrealistic. Another fitting but totally realistic change, by the way, would be to alter plane spotting mechanics so that targets far away (without surface detection) cannot be seen and thus targeted so easily. In reality, planes often were not able to identify targets nor were able to see ships covered by clouds.
  15. IVemo

    Whining about CVs :)

    The game could easily put 2 bad CVs against a good one at higher tiers. On what to do with the CV class, I would personally like to hear what WG has in store for them in advance to avoid Graf Zeppelin type of disaster happening again.