Jump to content

Lokistics

Members
  • Content Сount

    14
  • Joined

  • Last visited

  • Battles

    982

Community Reputation

18 Neutral

About Lokistics

  • Rank
    Seaman
  • Insignia

Recent Profile Visitors

The recent visitors block is disabled and is not being shown to other users.

  1. Lokistics

    @WG: Fair Fights Are More Fun

    The current matchmaker tries to match each ship on Team A with a ship of the same class & tier on Team B; that part you probably knew. From what I've seen, it works something like this: For each ship in the queue in order of wait-time, find a matching queued ship (or wait a little while for one) as follows: If it's a premium ship: 1. Try to match it with an identical premium ship (Belfast <-> Belfast), if possible, else: 2. Try to match it with another premium ship of the same class & tier, if possible, else: 3. Try to match it with a tech-tree ship of the same class & tier, if possible, else: 4. I'm not sure, it rarely gets to this point. If it's a tech-tree ship: 1. Try to match it with a tech-tree ship of the same class & tier, if possible, else: 2. Try to match it with a tech-tree ship of the same tier, if possible, else: 3. I'm not sure, it rarely gets to this point. Then add both ships to the match, one on each team. Obviously divisions are a little more complicated; I'm not quite sure how it handles them. It does seem to try to match them against other divisions, and it probably follows the above rules for each ship in the division. This is how premium ships are currently "balanced", as far as I can tell. They're not balanced against tech-tree ships, they get their "balanced" win rates by special matchmaking (putting 1 on each team, when possible).
  2. Lokistics

    @WG: Fair Fights Are More Fun

    So, quick probability/statistics lesson; judging by last page of responses, I think it's needed. If you flip a fair coin, it has an equal (50%) chance of being either heads or tails. Flip it twice, and there is a 25% chance both will be heads, 50% chance for one of each, and 25% chance both will be tails. Flip it 3 times, and there's a 12.5% chance all will be heads, 37.5% chance for 2 heads and 1 tails, 37.5% chance for 1 heads and 2 tails, and 12.5% chance all will be tails. The exact probabilities aren't important; pay attention to the pattern instead. The highest probability is that it will be close to 50% heads and 50% tails, and the more times you flip it, the smaller the probability of getting all heads or all tails. Now, what if someone hands you a coin, and you don't know whether it's fair or if it's a trick coin? The only way to find out is to flip it, and the more times you flip it (the larger your sample size), the more confident you can be that the results you get accurately represent the true probability of the coin. If you only flip it twice, there's still a 50% chance of getting either all heads or all tails, which doesn't accurately reflect the coin's probability. Though the probability of getting a non-representative result shrinks as the sample size grows, it never goes away. There will always be some margin of error. Obviously, ships in this game, and whether they win or lose, is more complicated than a coin flip. Things like: player skill, captain abilities, ship upgrades, etc etc all play a part. Fortunately, that stuff can be mostly ignored when figuring out if a ship is balanced. With a large enough sample, the average player skill should be pretty close to the average across all similar ships, unless there's some reason better/worse players tend to be playing the ship more. The same goes for the other factors. Overall, you have the average player with the average captain, average upgrades, etc. Now, noise also plays a part. Say you have a bad game, you screw up or whatever and you lose, but not because of the ship. This is a bad data point (noise), that doesn't reflect the ship's true chances of winning. The smaller the sample, the more that 1 bad data point can skew the results. Since noise is random, it should even out and eventually you'll have a game where you win, but not because of the ship. So with large enough samples it's not really a problem, but the smaller the sample the more normal variation you can expect to see. Balance conclusions between ship classes can't be drawn from win-rate, but that's because the matchmaker tries to put an equal number of ships from each class on each team. However, DPS is also the wrong metric; you're forgetting a DD's main advantage: stealth, and the main thing they contribute to a team: scouting. I'm not ready to say BBs are OP yet, because (if you didn't figure it out from the USS Cod videos WG put out) submarines are coming, which could provide not only a good counter to BBs and CVs, but also the ship class that DDs should be strongest against. I think you're on the right track; making the Premium Account subscription more appealing is probably the most viable way for the game to move away from pay-to-win. It would need some immediate (non-combat) incentives, as well as some recurring rewards for keeping your subscription active. I shall resist the urge to be insulting. Please learn statistics.
  3. Lokistics

    @WG: Fair Fights Are More Fun

    That's an interesting case. As they all have worse win-rates than Myoko herself, my guess is that it's a combination of the following: The ARP ships haven't been available for a while, so people playing those ships are probably veteran players. Players see an ARP ship and think: "Ooh, fancy, I should shoot at that." Being focused in a cruiser is usually bad. Losing a veteran player hurts the team more than losing a potato. If you look at ARP Kongo clones, it's actually the opposite: they all have better win-rates than Kongo herself. I think the difference is, being focussed in a battleship is good for your team. However, it's important to keep in mind that the sample sizes for ARP ships are an order of magnitude smaller than for the base ships. There will always be some noise, which gets amplified as the sample size gets smaller, so I'd expect the win-rates for the ARP ships to fluctuate a lot more than those of the base ships. Other possibilities: The base ship got some buff/nerf that wasn't applied to the ARP ship. The sample of players who play the ARP ships are just a little better/worse than average, as most players probably don't have both the ARP ship and the base ship. Statistical noise; it's a thing. It's why you always see a margin-of-error in polls.
  4. Lokistics

    @WG: Fair Fights Are More Fun

    Nueve de Julio is in the "Pan-America" nation, and as such can't share captains the way Boise can, so players probably run a less-skilled captain overall.
  5. Lokistics

    @WG: Fair Fights Are More Fun

    The game has always been a rock-paper-scissors balance between battleships, cruisers, and destroyers; let's leave tactics out of this particular discussion, as they vary by map and evolve with the meta. However, you two bring up a good point: we need a definition of "fairness". I'll call a ship fair/balanced if it has a 50/50 aggregate win rate without any special matchmaking. While this is not a perfect metric, if we can get all ships to at least that point, I think it's a good 1st step. I don't personally have enough data to take a position on the subject, but regardless it's off-topic. If you want to talk about how you feel betrayed by WG, etc, please make your own thread for that; they haven't censored you so far.
  6. Lokistics

    @WG: Fair Fights Are More Fun

    Good info, thanks. I'm a Software Architect in my day job; I write algorithms all day long. I'm not "clinging" to anything; I'm simply trying to figure out how the matchmaking algorithm works based on the available evidence. It's helpful to know that the matchmaker doesn't try to even out premium ships between the teams overall, and doesn't try to even out premium ships between the teams based on class; thanks, that's a good data point. I'm not sure why you're so convinced that it doesn't do so based on class + tier, or based on ship, as the evidence available to me suggests this happens much more often than simple probability would allow. I'm pretty sure the matchmaker will, for example, happily put all of the Zaos on the same team, as long as there are an equal (or at most +/-1, in rare cases) number of T10 cruisers on the other team. If the matchmaker won't do the same for premium ships, that means premium ships get special matchmaking treatment. If all players had access to T10 ships (I'm not suggesting this), then I'd agree, it wouldn't be a problem. The issue is that the players who are most likely to be bottom tier are newer players, which is the opposite of what you want (assuming you want the newer players to stick around). Perhaps a matchmaker rule where a player is always top-tier in their highest-tier ship would help. That way, you'd never face T10s until you have your first T10. I'll add that to the list of suggestions.
  7. Lokistics

    @WG: Fair Fights Are More Fun

    I'm aware the game has a progression model; this thread is about how to improve it. How about the next tier? T10 doesn't have stock hulls. Some T10 ships do have module options, but they aren't pure upgrades the way the B/C hulls, etc are for T2-T9 ships; they're trade-offs. You know what are supposed to be fair? Games. I'm not talking about strategy within a match; I'm talking about matchmaking.
  8. Lokistics

    @WG: Fair Fights Are More Fun

    If you can screenshot a match with 2+ division-less premium ships of the same class & tier on the same team, and none on the other team, I'll agree that I'm seeing patterns in the noise.
  9. Lokistics

    @WG: Fair Fights Are More Fun

    And when you played those ships, what percent of the matches had 2+ division-less premiums of the same class & tier on your team and none on the other team? I can and I have, and I still don't want to play matches where I'm T8 against a bunch of T10s. The fact that you're worried enough people would pick that option tells me that you know most players don't like it. You sound like a thoughtful person, so please think about this. Player retention can never be 100%. In order for the game to stay viable, we need to constantly recruit new players to replace the ones who leave. This game has a pretty steep learning curve; the skill gap between new players and veterans is massive (which is OK). Why compound that by putting veterans in ships 2 tiers higher? The newer you are, the more likely you are to be bottom tier, which is the opposite of how it should be. I'd be completely fine with it if the matchmaker up-tiered the veterans and down-tiered the newer players, but that's the opposite of what usually happens. New players also have the least time/etc invested in the game; if it's not fun for them, they aren't gonna stick around.
  10. Lokistics

    @WG: Fair Fights Are More Fun

    Every time I've played the Yubari in a random battle, there's been exactly 1 other Yubari, and it's been on the other team. I'm not saying the matchmaker won't create a match for a premium ship without one of the same class & tier on the other team, just that it prefers to, when possible. If you don't believe me, track your matchmaking for a week and then come tell me I'm right. Ignore divisions though, I don't have enough data to know how the matchmaker handles that. Says the Beta Tester. If some people want to do that kind of thing for an extra reward or something, fine, make it optional, but don't force it on everyone. For people trying to grind to their first few tech trees, it's not fun. Designing the game based primarily on what your most skilled/oldest players want is the exact problem I'm talking about.
  11. Lokistics

    @WG: Fair Fights Are More Fun

    Yeah, I saw that shortly after I posted xD Updated the original post accordingly. The only changes to the matchmaker I suggested were: Balance premium ships so they don't need special matchmaking (simpler matchmaker). The current matchmaker tries to match a premium ship on Team A against another (preferably identical) premium ship on Team B. Limit the matchmaker to 2 tiers (simpler matchmaker). I like the idea of a matchmaker ranking, so you'd get matched against similarly-skilled players, but I don't think the game has enough of a player base to support that currently. To make that possible without glacial wait times, the game would need to actually be able to recruit and keep more (new) players... which would require un-tilting the playing field so new players don't face the double-whammy of playing against better opponents in objectively better ships.
  12. Lokistics

    @WG: Fair Fights Are More Fun

    In the game's current state, that's probably true. Perhaps we as a community need to figure out a way to move the game away from pay-to-win, while still being economically viable for WG.
  13. I'm a casual player; I have a job and a life. What I want is a game where I can log in a few hours a day, several days a week, and have some good PvP games where the outcome is determined by the skill of the players. I actually want to pay real money to support games I like (having a job does come with perks), but I can't justify paying to support the game in its current state, and the original NTC proposal would have taken it in absolutely the wrong direction. I see WG taking features that would be great for any PvE game, and putting them into what is, at its core, a PvP game. In PvE games, the more you play, the more advantages you get/the more powerful you become, and it's not a problem because you're not fighting other players. Putting that crap into a PvP game is a great way to make sure new players can never catch up to a competitive level, and quit as soon as they realize it. The reason everyone had their hair on fire about the original NTC proposal is that it would have taken what is already an overly-tilted playing field and made the problem even worse. I'm absolutely fine with other players being better at the game, but it should only be skill. Being a veteran player is already enough of an advantage. Things WG could do to actually improve the game: Balance premium ships to the point where they don't need special matchmaking to get a 50/50 win/lose rate. Always put players top-tier in their highest-tier ship, or don't put players against ships higher-tier than their highest-tier ship (new matchmaker rule). This way, new players would never face T10s until they have their first T10. Disable 3-tier matchmaking; fighting ships 2 tiers higher is no fun. I'd rather wait longer for a match where I can have some impact. Remove premium consumables; WG already plans to do this, and this is a good change. Remove stock hulls and modules that are purely improvements; nobody wants to play a nerfed version of the same ship. Let rental ships for ranked/clan battles equip camo; why wouldn't you? Overall, design the game around the philosophy that fair fights are more fun. Sure, some veteran players might run out of things to grind and go do something else for a while, and that's OK, because the game will actually be able to recruit and keep new players.
×