Jump to content

Winter_Man

Members
  • Content Сount

    40
  • Joined

  • Last visited

  • Battles

    6386
  • Clan

    [RBMK]

Community Reputation

14 Neutral

About Winter_Man

  • Rank
    Seaman
  • Insignia

Recent Profile Visitors

514 profile views
  1. Winter_Man

    PSA Regarding KOTS Mission for NC camo

    Has anyone gotten the mission yet? 0.10.10 dropped but I still don't see it.
  2. Sorry to hear about your ranked experience! Just remember that you earned all the low-hanging prizes and there's no shame in stopping at 12 (I do it often). As for the Anchorage, I'm with @RainbowFartingUnicorn that it's a bad idea to sell premium ships. If you want it back I think you can submit a ticket to WG and ask.
  3. Oh man that sounds rough! I've never done more than 130 ranked battles in a season and I thought that was crazy haha (although I never made rank 1 )
  4. The summary graph was generated by simulating 52 separate win rates (48%, 49%, ... 99% 100%). For each win rate 50,000 players were simulated and the program counted the number of battles each player took to reach rank 1. The median battles for each win rate was shown on the summary graph. By definition, half the players will take less battles than shown on the summary graph and half the players will take more battles. Since there is no certainty (i.e. playing ranked battles has inherent randomness), the best way to show the randomness is with confidence bounds. Typically people use 95% confidence bounds. As for the inputs, the only 2 inputs are win rate and first place chance. Since I control these inputs and simulate each possible win rate, there is no deviation associated with them. The deviation is only in the output.
  5. Yeah my personal opinion is that getting to rank 11 or 12 then stopping is the best move. It gives you most of the bonuses without wasting too much time. The program I wrote can also simulate how long it takes to reach a specific rank (doesn't have to be rank 1). So if people are interested I could make a graph of how hard it is to stop at rank 11 starting from 18.
  6. You're right. There should be a correlation between win rate and the chance of you getting first place. I could not find any data to correlate the two, so I assumed it was random (since there are 7 players, a 1 in 7 chance is just under 15%). In general, as your win rate gets higher you will reach rank 1 quicker. So much quicker that your first place chance doesn't really change how quickly you reach rank 1 if you have a high win rate. First place chance has the largest impact on lower win rate players.
  7. By deviation, I assume you mean the confidence interval. That is next on my "to do" list. I'd like to update the summary graph so that it shows the 95% confidence bounds. In general, as your win rate gets lower, your "deviation" increases. So People with 48% win rate could take 700 battles to reach rank 1 or they could take 3,000 battles to reach rank 1. While people with a 60% win rate will almost never take more than 300 matches to reach rank 1.
  8. You have some good suggestions. I'll address them each... Do you know that ranked win rate is not the same as random battle win rate? Yes. Win rate in the simulation refers to your ranked win rate. There is no data publicly available that allows us to correlate the rank WR to random battle WR unfortunately. Otherwise I would try to make it related to your random battle win rate since it is easy to see your random battle WR. Why doesn't the graph end at 70% and start at 40%? The program simulates how many battles it takes to reach rank 1. For every win rate between 48% and 100% it simulates 50,000 players trying to reach rank 1. It counts the wins required for each simulated player and determines the median battles required for a player to reach rank 1. As a player's win rate approaches 100%, it becomes much easier for the program to simulate the player (since they typically reach rank 1 much quicker). As a player's win rate approaches ~50% it becomes painfully slow to simulate. If I had a super computer I could simulate lower win rates, but as you drop below 48% the required battles quickly exceeds 1,000. It is not feasible for players below ~48% WR to attain rank 1 because it will take too long to play enough battles. Why don't you highlight the 45% to 55% block clearly since most players fall into this category? In my original post I only included the summary graph. There are more detailed histograms for each win rate available here: https://github.com/jcorvino/wows-ranked/tree/master/example-regular-season. I've included an example image for 50%
  9. Now that ranked season 18 has started, I thought I'd share this simulation with everyone. Here's how many battles it takes to reach rank 1 based on your win rate: The simulation is written in Python 3.8 and is open source. You can view the codebase here: https://github.com/jcorvino/wows-ranked
  10. This assumes that the drop rate is proportional to the number of soviet token bundles in the armory. Also I do assume that you buy them in order. If you get >1200 tokens, you can buy bundles 1, 2, and 3 in that order. I edited the main post and incorporated some of your comments.
  11. Absolutely correct! The 300 token container is there to skew the average up so that some lucky players can actually finish the grind. I made a previous post in this thread that shows the average is probably closer to 9.75 tokens/container if you can't get the 300 token container.
  12. Winter_Man

    Cruiser Tallinn for Token, a pipe dream?

    Right now it seems like the missions will repeat 4 times (once each week). This would give 840 tokens for the P. Bagration, 840 for the Ochakov, and 420 for the Mikoyan. The grand total would be 2100 which matches the premium shop descriptions. It hasn't been confirmed yet because in-game nobody can tell if the combat missions repeat each week for certain.
  13. Winter_Man

    Directive 1 tokens...

    I wrote a post here (link below) explaining the expected rewards from the containers. Basically, you can expect ~22 tokens/container. However, the 300 token container skews this average a lot since it is such a high value. If you ignore the 300 token container, the average is ~9 tokens/container. This agrees with what a lot of players are experiencing right now. Basically, get more containers and pray to RNG for a 300 token container.
  14. Winter_Man

    Soviet containers

    I wrote a post here (link below) explaining the expected rewards from the containers. Basically, you can expect ~22 tokens/container. However, the 300 token container skews this average a lot since it is such a high value. If you ignore the 300 token container, the average is ~9 tokens/container. This agrees with what a lot of players are experiencing right now. Basically, get more containers and pray to RNG for a 300 token container.
  15. There's another topic discussing that here: Basically, we expect that the missions will repeat each week. Therefore, the missions will repeat 4 times and give a total of 2100 possible tokens. It hasn't been confirmed yet, since the event has not been up for a full week.
×