Random and ranked games are unrelated

Back from anti-conspiracy theory duty! I hope some people sleep better, knowing that neither a traitor has the US nuclear codes, nor a bunch of traitors are working to claim it. But more important things are calling, namely video game strategies.

I’ve always seen World of Warships random and ranked battles two completely different games, attributing it to random being rigged for maximum profit. But the reason is irrelevant. The fact matters, because it creates actionable advise: ignore your random statistics, because it’s irrelevant, just maximize XP gains and play ranked for challenge.

I kept looking for more evidence of this claim and here I have a fat one. I compared the random and 9th season T8 ship statistics. Not players, but the total ships. At first, we can see that players relatively use the same ships in ranked as in random:

Now let’s see how the ships perform in different battles:

Good correlation again. So the players use the same ships with the same performance, which clearly provides the same results:


Absolutely not! The correlation between random and ranked winrate of a ship is a mere 0.25. This cannot be explained any other way than the strategy of the games is very different. I tried to figure out how, by ranking the ships by ranked battle winrate and also by random winrate and calculating the difference. There are many ships with shocking differences, like the Hsienyang being #29 out of 35 in ranked battles, but #4 in random battles:

Unfortunately I don’t know enough of the ships to see any reason what makes them perform so differently in random and ranked battles, but they clearly do. Maybe someone smarter can figure out why. But that’s secondary. What really matters is that these game modes are fundamentally different and getting better in random doesn’t help you with getting better in ranked.

Author: Gevlon

My blog: https://greedygoblinblog.wordpress.com/

10 thoughts on “Random and ranked games are unrelated”

  1. How about “Ranked and Random games are different” instead of “unrelated”?

    You just found a good correlation in Ranked and Random damage but low correlation in Ranked and Random win rate. You immediately disregard the damage data to declare them “unrelated”.

    This is a clear confirmation bias. You have some pet theory and fish for data until you find something you like and then immediately interpret it as a “proof” of your pet theory.

    How about this: Random games are random, unrelated events. No matter how well you do in one of them and how often you play, your next random game will be roughly the same as the last one.

    In Ranked the battles are related. If you play better you end up playing against better opponents. If you play more you keep up with the curve and play on your own level. If you play fewer games you stay behind the curve and play against worse opponents (because the better ones left for higher ranks). This means that winrate is based on more than just individual skill, but also on time played and rank bracket played. These variables are absent in Random games but it doesn’t mean that skillset required to succeed is very different.

    Another difference: random stats show people levelling their ships. Every ship is played there because it has to be to unlock the next Tier. In Randoms people pick the ship based on their preferences and the power of the ship. Therefore something like LoYang is vastly overrepresented because it has the best hydro in the game and never bumps against its natural counters from higher Tiers (Z52 and YY).
    Look at HsienYang. It is a pretty good ship in Randoms: a copy of a Benson with funny torpedoes. It can win artillery duels with most other DDs (including some higher tiers) and torp the clueless BBs in the chaos that is a typical Random game. Also, because it’s a rather new addition to the game, the people playing it are mostly experienced players (the newbies haven’t caught up to PA DDs yet and play IJN), giving it better stats on average.
    In Ranked it is almost guaranteed to play against a LoYang because they are from the same nation. It is vastly worse than Lo because it doesn’t have the great hydro. Every good player knows this before they even pick their ship and simply doesn’t play HsienYang. As a result, only poor players (or those playing “for fun” who disregard their winrate) drive the ship in Ranked and give it horrible stats. Even worse, because the ship is unsuited for Ranked, there is a number of players who play it Ranked to level towards ChungMu (they pick their ship based on their levelling needs rather than ship’s qualities). Not only do they play an inferior ship but also they don’t have the experience to get the best of it.

    And the last one: you are comparing ships, not players.
    Quote: “Good correlation again. So the players use the same ships with the same performance, which clearly provides the same results:”
    Wrong. There is no reason to think they are the same players. For example, I have played and XPed through all T8s but I only played 1 of them in Ranked per season (Martel and Kutuzow).


  2. @Stawek: no, the two games are unrelated BECAUSE their damage numbers are so well correlating. It shows that it’s not the ships behave differently (which would make sense due to only facing same-tier instead of uptiered/downtiered). It’s the meta/strategy that is different.

    You are now contradicting your old pet theory of “low ranked is bad players”. Now you assume that sub rank 10 players (who played T8 in ranked) are making wise decisions about ships and strategies. Some fun fact for you: the HsienYang has 0.4% popularity in random and 0.8% in ranked, twice as high. Why?

    Please form a coherent idea about the happenings. Mine is simple: these are completely different games with different matchmaking, played by the same tools (like Black Jack and Poker), so we should ignore randoms all together.


  3. No, I am making an exactly opposite statement: bad players in R11 make bad decisions and play HsienYang, good players play other DDs, therefore Hsien has terrible statistics in Ranked.
    In Randoms it’s reversed: Good players play HsienYang because they are farming towards the much better YueYang. Poor players (and those who play fewer games irrespective of skill) don’t play HsienYang because they are still levelling the more popular, older lines.
    Of course the strategy is different, I’ve been saying this from the very beginning. It doesn’t mean they are completely different or unrelated. Most importantly Ranked stats are much less reliable because they can’t take into account some hidden variables and have some feedbacks absent in Random. Random stats are unreliable because they contain divisions.
    In the end, a good player will be good in both Ranked and Random games. Skills like shooting, map awareness and positioning are important in both. Just like somebody with good intelligence will be good at both Poker and BlackJack (to the point where skilled players have been banned from BlackJack casinos for counting cards). Idiots will lose in both.
    Besides, you are slowly getting better win rates in Random games so it’s time to get convinced that the win rate in Randoms actually means being good.


  4. Gevlon: Other than a single line about going all in to tier 7,8,9, and 10 japanese destroyers and spamming torps from long range… any other advice?


  5. @Stawek: Poor players (and those who play fewer games irrespective of skill) don’t play HsienYang because they are still levelling the more popular, older lines. OK, but then why do they suddenly play it in ranked in 2x numbers.

    I’m not getting better winrates. I was at the same winrate before I purposefully ruined it with the useless Mahan test. I still get complete nonsense results, like 3-4 kills defeats and zero damage wins. (damn, I lost a Kraken because the enemy reached 1000 points 2 seconds before the torps hit the 5th target)

    I’m 100% sure that the outcome of the random battles are decided at loading screen and if I was placed the other team (replaced by the equal tier DD player), still the same team would win. On the other hand in ranked I could always make difference.

    @Phil: advice on what?


  6. There weren’t more Hsienyangs in Ranked than there are in Randoms. First of all, only a 3rd of the players even tried Ranked. Second, a lot of them played only 30 games or so, lastly, most games were played in T10s. Sure, it’s twice as high percentage but from a vastly lower base number.
    Also, I don’t know how is the popularity in Randoms calculated and what is the 100%? is it all ships, is it all T8s, is it last few weeks, is it all time? Does your Ranked number include the previous season, when Hsien was a new addition?

    There are 250 or so ships total, 0.4% is just about level. There were some 50 ships in T8 Ranked, 0.8% is less than half of the number expected.


  7. Phil: advice?
    The best learning ships are German and Japanese cruisers. They will teach you the basics and punish you instantly for mistakes. Quick feedback is good for learning. You will not regret having Zao or Hindenburg.

    IMHO the strongest line overall are PA destroyers nowadays. They are very good up until T9 with smoke and OP in T10 with radar, but they can be really hurt by any radar nerfs (YY, in particular, has already been affected by modules nerf).
    Unlike Gevlon I think you should immediately sell lower tier ships and transfer the captain to the next tier. I see no point in keeping Yugumo while you can play Shimakaze. If you don’t have premium and the credits to play higher tiers start earning those in a side line, preferably a different class so you learn the thinking of your enemies. For BBs and DDs I advise US ships for learning purpose. Do not play CVs.

    As for game tactics, focus on staying alive and make sure you only take fights where you’re at an advantage. That’s the only advice I can give you that applies to all ships, everything else is too specific.


  8. On how to play, specifically. IE:

    Something goblinish like:

    With the Japanese destroyer at tier 10, you have a generous 20 km range. So use it, completely ignore your guns and keep range at all costs.

    Anything sneakier than that?

    Also: When you get a container, I always choose resources, since you can trade those for limited edition ships. In theory, everything else (XP and credits) can be ground out


  9. @Phil: ranked is the only “game” here. Randoms are just a grind to get ships for it and to get a basic hand-eye coordination with them. Select one type of ship and play only with that one. The Wiki pretty well describes their strengths and weaknesses. I like IJN destroyers for being able to use torpedoes outside of detection range, but any line is fine.

    Ranked seasons happened in many tiers, not just T10, so you should have ships and basic skills for all tiers (besides really low tiers). In retrospect I regret having 5 captains, but that’s crying after poured milk. You should have only one well trained captain and retrain him to the proper rank when ranked season comes.

    Most important advice: never expect random teammates perform above the level of bots. Never sacrifice anything for them. Don’t spot for them, don’t tank for them. Let them do it for you and just farm damage which gives the XP

    @Stawek: Yes there were. The data is T8 ships in random (that’s the 100%) and T8 ranked ships from season 9 only.


  10. One piece of data to look at for your ships that rank well in ranked vs low in random is that kutuzov, tirpitz, prime eugen and atago are all ships that must be bought with cash or doubloons. These ships are played in ranked during those seasons, and seen there often in my experience. Tirpitz has the nickname “Derpitz” as players new to the game wanted a cool ship right away and bought it to play, with often cringeworthy results. Atago is a very good ship, but is like many ijn cruisers unforgiving in steering errors. Eugen is similar, and is also a ship that people new to the game could buy and try to jump in at t7 games to theirs and their teams typical detriment.

    Kutuzov is a ship no longer for sale, and she is a monster at tier or +1, due to having smoke, long range and accurate, fast guns. She is considered an unbalanced ship in the meta, able to shred destroyers, outgun other cruisers, and be a flamethrower for battleships.


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s