Jump to content

renalove

Members
  • Posts

    82
  • Joined

  • Last visited

  • Days Won

    2

Posts posted by renalove

  1. 13 hours ago, 430i said:

    @renalove The expected score numbers in the league table do not match the ones in the booster simulator table.

    9 hours ago, OmerB said:

    I saw something really strange in the league (on PSH) - I have the same team on two slots, but I got different predictions for each on of them (although they are the same)!

    Actually, the game showed me different stats for the two teams, although they have the same girls... 😮 So it's not a script thing, but something that was broken in the game.

    I didn't know where to report this (probably on the Bugs sub-forum when I think of it).

    On PSH:

    ImageImage

    For some reason the team selection page is broken after the Labyrinth update. The first team on the team selection page is equipped with boosters, but the other teams are not. My script was using the data from the team selection page so it broke together. The league table simulator has already been fixed, but there are no plans to fix the team selection page simulator.

    On 11/28/2023 at 3:03 AM, 430i said:

    @renalove Is it possible to extend the booster simulator (for a given player) to include the simulation with the opponents' unboosted stats? Currently only the current data is taken into account, which is boosted more often than not, but the optimal booster config might be different between boosted and unboosted stats.

    I added a simulator for fighting an opponent whose boosters have expired. This simulator only supports legendary chlorella, cordyceps, and jujubes. It is inaccurate because it does not support ginseng, but it is still useful.

    image.png.4480d17223b1147f86cb6e08c6efe5d0.png

    • Thanks 5
  2. The booster simulator that was broken in the game update has been fixed.

    This bug fix was painstaking. The game update has changed the calculations of Endurance and Harmony. Ginseng and Club's bonuses are added together when calculating Hardcore, Charm, and Know-how, but multiplied when calculating Endurance and Harmony. In other words, Endurance and Harmony are calculated from higher Hardcore, Charm, and Know-how than their actual value. Additionally, due to a bug, Club's Endurance and Harmony bonuses are partially applied twice. Also, the Endurance bonus from Harem level has been changed, and I could not figure out the formula in the end, so there is a slight difference from my booster simulator predictions.

    These are the formulas I use in my script. (in the case of Hardcore Hero)

    Endurance:

    var endurance = 0
    endurance += 4 * hero.hardcore * (1 + club.hardcore)
    endurance += equipments.endurance
    endurance *= 1 + club.endurance
    endurance += 4 * equipments.hardcore * (1 + club.hardcore)
    endurance *= 1 + club.endurance
    endurance += 4 * (hero.hardcore + equipments.hardcore) * (1 + club.hardcore) * ginseng.hardcore
    endurance += harem_endurance * (1 + club.endurance) * (1 + club.endurance * (1 + club.endurance))

    Harmony:

    var harmony = 0
    harmony += 0.5 * hero.charm * (1 + club.charm)
    harmony += 0.5 * hero.knowhow * (1 + club.knowhow)
    harmony += equipments.harmony
    harmony *= 1 + club.harmony
    harmony += 0.5 * equipments.charm * (1 + club.charm)
    harmony += 0.5 * equipments.knowhow * (1 + club.knowhow)
    harmony *= 1 + club.harmony
    harmony += 0.5 * (hero.charm + equipments.charm) * (1 + club.charm) * ginseng.charm
    harmony += 0.5 * (hero.knowhow + equipments.knowhow) * (1 + club.knowhow) * ginseng.knowhow
    • Like 1
    • Thanks 5
  3. 2 hours ago, Horsting said:

    @renalove Something seems to be wrong with the booster simulator in leagues:

    image.thumb.png.b45e7801596ad13e734b66280b59dc84.png

    I currently have an AME and 3 Chlorella + 1 Cordyceps equipped. The E[P] on the opponent's side is matching also the one in team pages and the battle page, but the booster simulator shows lower results on all booster combinations. The same is true for all other opponents (the 5 or so I checked).

    It seems to be related to this bug fix. But why are endurance and harmony up over 10% higher than before the update? I have no idea what calculations they are doing... It will take some time to fix.

    Quote

    Bug fixes:

    • The calculation of the Club’s endurance and harmony bonus has been corrected to be taken into account in the player’s stats.
    • Thanks 1
  4. 2 hours ago, renalove said:

    What a surprise! A lot of numeric data that was string type has been replaced with numeric type. This applies not only to boosters, but also to player IDs, quest IDs, equipments, books, gifts, girl poses, and much more. And it has not been replaced in Nutaku. It could just be that the update is delayed, but it is also possible that they decided to use a different data structure for HH.com and Nutaku.

    HH.com (Chlorella)

    image.png.be40fe33ee7156860b68aed5e36d5bf7.png

    Nutaku (Chlorella)

    image.png.be19fa2a15a6e5164f93efcae4ee041b.png

    This change could affect almost all scripts. If you find a bug, please report it to us. I will also check each one to see if it affects Battle Simulator.

    This change has been reverted. It is no longer necessary to modify the script. 😂

    • Hearts 1
  5. 9 hours ago, Master-17 said:

    Icons for booster simulators in the Season in all games are again superimposed on the "!" icon.

    Thanks. The fix is simple.

    17 hours ago, Horsting said:

    the booster timeout cycles in league are broken since today. It seems that time data which was previously given (and expected by the script(s)) in hours, is now provided in minutes, or instead of minutes in seconds. The cycle is full for all timeouts of 24 minutes or more, instead of 24 hours:

    image.png.eece2a2612211861ce44cdd4b9118a5b.png

    What a surprise! A lot of numeric data that was string type has been replaced with numeric type. This applies not only to boosters, but also to player IDs, quest IDs, equipments, books, gifts, girl poses, and much more. And it has not been replaced in Nutaku. It could just be that the update is delayed, but it is also possible that they decided to use a different data structure for HH.com and Nutaku.

    HH.com (Chlorella)

    image.png.be40fe33ee7156860b68aed5e36d5bf7.png

    Nutaku (Chlorella)

    image.png.be19fa2a15a6e5164f93efcae4ee041b.png

    This change could affect almost all scripts. If you find a bug, please report it to us. I will also check each one to see if it affects Battle Simulator.

    • Like 3
    • Thanks 1
  6. Script updated. Added booster and skill simulator to season.

    2 hours ago, Horsting said:

    Makes sense 😅. I remember you wrote something when adding GS5 support the first time. Makes sense to handle it differently than crits, i.e. not splitting simulations every round/turn, as this would additionally double or quadruple (stun vs stun) the possible outcomes and hence simulation efforts additionally every round.

    But it expects it to be triggered in the first round then, and you use 7%*lvl as chance for the whole fight? I mean the chance that stun is triggered anywhere in a battle is quite high. E.g. with 5 rounds and stun level 5:
    sum{i=0;i++;i<5}(0.65^i*0.35)=1-0.65^5=88.40%

    But if it is triggered in the second last round, its effect is ~halved, and if it is triggered in the last round, it has no effect at all. And due to GS4, if it is triggered in the first round, its effect is a little smaller than if it it is triggered in the second round. So without really doing the additional battle course split in the simulation, I see no chance to get an accurate result. But doing the exact calculation will lead to an endless calculation.

    In practice, that expectation calculation is done for every round until stun is triggered. As you say, this can take 4 times longer. Assuming 23 points if stun triggers, 20 points if it does not, and a 35% chance of stun, the expected value is calculated as follows.

    expected_value = 0.35 * expected_value_with_stun + 0.65 * expected_value_without_stun
    
    Round 1
    0.35 * 23 + 0.65 * 20 = 21.05
    Round 1 + 2
    0.35 * 23 + 0.65 * (0.35 * 23 + 0.65 * 20) = 21.7325
    Round 1 + 2 + 3
    0.35 * 23 + 0.65 * (0.35 * 23 + 0.65 * (0.35 * 23 + 0.65 * 20)) = 22.176125
    Round 1 + 2 + 3 + 4
    0.35 * 23 + 0.65 * (0.35 * 23 + 0.65 * (0.35 * 23 + 0.65 * (0.35 * 23 + 0.65 * 20))) = 22.46448125

    Stun at the end of the fight has a little effect on the expected value because the probability is so small. The probability of stun being triggered in each round is as follows.

    1st: 0.35 = 35%
    2nd: 0.65 * 0.35 = 0.2275 = 22.75%
    3rd: 0.65 * 0.65 * 0.35 = 0.147875 = 14.79%
    4th: 0.65 * 0.65 * 0.65 * 0.35 = 0.09611875 = 9.61%
    5th: 0.65 * 0.65 * 0.65 * 0.65 * 0.35 = 0.0624771875 = 6.25%

     

    • Thanks 1
    • Hearts 1
  7. 17 hours ago, Ravi-Sama said:

    I like that the girls' attributes are visible now w/ little icons at the bottom, while hovering over them.  Useful to check if a player is actually making proper use of girl skills.  Also, to double check the synergy of your own teams.  Wish it was like this throughout the whole game.

    image.png

    Thanks for posting the screenshot. Yesterday I added the girl's traits to the tooltip. I noticed that some pages did not have them added, so I fixed that today.

    13 hours ago, Horsting said:

    please correct me if I am wrong, but AFAIK the stun skill is calculated as if it was triggered always after first attack. While theoretically, due to GS4, it would have a larger effect when being triggered later, it can in practice also be triggered not at all, or before the last round, so that half of its effect is lost. So I think it is currently overrated by the simulator.

    If stun is always triggered on the first attack, then level 1 stun and level 5 stun would have the same result 😅

    The script calculates both the expected value if the stun is triggered and the expected value if the stun is not triggered, then multiplies each by its respective probability and adds them together.

    expected_value = expected_value_with_stun * stun_percentage + expected_value_without_stun * (100% - stun_percentage)

    I do not believe there is a bug in this calculation. If you find a specific strange case, please let me know.

    • Like 1
    • Hearts 1
    • uwu 1
  8. 16 hours ago, Master-17 said:

    If, when choosing an opponent in Leagues, Season or Pantheon, you have a team that gives better average results, then a button would appear that switches the active team to this one.

    16 hours ago, Master-17 said:

    In theory, this should not be considered cheats in any way, because in essence, only information is read, as in the case of all other presented scripts, and changing the command simply saves us 2 transitions with loads.

    It is not easy to store all teams because raw team data is surprisingly large. And I think the function to automatically switch to the best team with a single click may violate Terms of Use. I do not plan to implement such a feature. However, it is good to discuss such ideas.

    16 hours ago, Master-17 said:

    P.S. For the Season, the team is selected for the currently selected opponent (1 of 3)

    Added a simulator for the last selected opponent in the season to the team page.

    3 hours ago, Horsting said:

    Currently the all opponents booster simulator does not respect the "skip fought opponents" flag. Could you add this?

    Added a filter to the booster simulator in the league table.

    • Like 2
  9. I proved the multi-fight bug on the test server. My weak team had the skill Reflect. My opponent's team did not have Skill 5. My sim prediction that supports Skill 5 was 4 points. Zoo's sim prediction that does not support Skill 5 was 3 points. The result of the multi-fight was 4-3-3.

    image.thumb.jpeg.487e4900bfbf74b31397c55643b56b27.jpeg

    image.thumb.jpeg.5a0dfb5e748757a796fed13d806ba330.jpeg

    This result indicates that Skill 5 was only triggered in the first fight. Skill 5 should be triggered once per fight, but in fact it seems to be triggered only once in all the fights. The developers probably forgot to reset the flag in the loop process. This would also explain why single fights are not affected by the bug.

    • Thanks 9
  10. As you may have noticed, several features have been added to the script. Most of them are slow and are disabled by default. You can enable them in the config if you need to.

    image.thumb.png.7417bfdf2a91f1e316b07f402d668d1f.png

    Booster and skill simulators simulate your booster and skill combinations. Not your opponent's. The skill simulator only changes tier 5 skill, tier 3 and 4 skills are not changed. The simulation is started by clicking on the icon next to your team hexagon.

    image.thumb.jpeg.c1261ca1db9d69bbd07e722e18bf4272.jpeg

    Clicking on the AME icon in your row on the league table will simulate booster combinations for all players.

    image.jpeg.2a1a181d814d4b3b5c99fff368eb338f.jpeg

    Added feature to show the probability of each league result. This is usually fast but sometimes slow.

    image.png.27836bb0a01f28c11ae659020c99a1e0.png

    Script
    https://github.com/rena-jp/hh-battle-simulator-v4/raw/main/dist/hh-battle-simulator-v4.user.js

    GitHub
    https://github.com/rena-jp/hh-battle-simulator-v4

    • Thanks 4
    • Hearts 2
    • uwu 1
  11. 24 minutes ago, Tom208 said:

    It's just due to the end of the daylight saving time this night.

    Oops, I forgot about DST. Okay, it's actually 6 hours because it adds 1 hour when DST ends. Thanks.

    • Like 3
  12. On 10/24/2023 at 12:47 AM, DvDivXXX said:

    PS: Rena, the one thing that's a bit inconvenient is that I need the "Go" button to work in order to access your Booster Sim, but unfortunately @430i's Hentai Heroes +++ script still breaks the "Go" button. I love using it to have an overview of the field's top past performances, but I need to disable it to use your new sim. It's a problem with 430's script, not yours though.

    OK. Now my script adds sims to @-MM-'s HH Leagues++.

    On 10/24/2023 at 6:27 AM, Master-17 said:

    Does not work in CxH and PsH. I visited the Market, went to edit all 6 commands, and still a message appears that there was an error with a suggestion to repeat all the steps. Tried it several times.

     

    Booster icon overlays the "!" icon

    1.jpg.6ff2b4c47640bf2dc61a6a67656a533c.jpg2.jpg.95fb068c23c79f1e11ef2f98a5fa008b.jpg3.jpg.f0db567e6e91fd98423d19612945dae5.jpg

    Thank you for your report. I fixed an error that occurred when not joining a club and changed the position of the booster icon. If you are still having problems, please let me know.

    4 hours ago, Basniowy said:

    I have observed that the booster simulator is having some issues with updating, when i am switching between different sets of player gear. Do i have to repeat the whole process (every single time) for the script to update?

    I checked and found that you need to reopen the team editing page after you change a hero gear. So in the current version, yes. This behavior is different from what I expect. I will look into having the script automatically load the team editing page and update the data.

    • Thanks 4
    • Hearts 1
  13. Battle Simulator v4 is released. It includes Booster Simulator.

    Shield vs Reflect
    image.jpeg.eda02c66667a27994c0d29a33212431b.jpeg

    GitHub:
    https://github.com/rena-jp/hh-battle-simulator-v4

    Script:
    https://github.com/rena-jp/hh-battle-simulator-v4/raw/main/dist/hh-battle-simulator-v4.user.js

    How to use:

    1. Remove my old simulator
    2. Install v4
    3. Go to the market page
    4. Go to every team editing page (not team selecting page)
    5. Go to Pantheon or League opponent page
    6. Click on the booster icon nearby your team

    Please let me know if you find any bugs 😉

    • Thanks 8
    • Hearts 1
  14. 15 hours ago, mdnoria said:

    Is it just me or the computing of PvP results from the BSDM script are extremely slow since today, especially noticeable in leagues since there is more fights to compute (to the point of crashing the tab) ?

    If you are using 4 chlorella and shield skill since today, that is why. Chlorella and shield prolong the battle and take longer to simulate.

    The script simulates two cases for each attack: a normal hit and a critical hit. For each additional attack, these combinations are doubled; For each additional round, these combinations are quadrupled, since there is one player attack and one opponent attack per round. So if the battle is extended by 5 turns with chlorella and shield, the simulation time will be 4 * 4 * 4 * 4 * 4 = 1024 times longer.

    An easy way to avoid this problem is to reduce chlorella. And cordyceps help to end the battle faster, thus reducing the simulation time.

    • Like 3
    • Thanks 1
  15. 11 hours ago, 430i said:

    Is there a way to somewhat correctly estimate the win probability with a headband booster before quipping one? AFAIK MM's script does not consider girl skills and renalove's script does not take potential headband into account.

    I was not aware of that. My script runs before MM's script, so you need to change the order. Please open "Settings" of my script and change "Run at" to "document-end" and "Position" to after MM's script. This should work.

    image.png.7b8eb6276d8bcb81272877303744f317.png

    I also plan to add a booster simulator to my script. The current development version requires hero stats and some bonuses to be calculated manually and written directly into a script. Once some formulas are figured out, this can be automated.

    image.jpeg.de0900ae3bd2976e9ac9b875838855f4.jpeg

    • Thanks 3
    • Hearts 3
  16. I updated Harem++, Liliat's script. It now supports girl skills 😁

    GitHub:
    https://github.com/rena-jp/haremplusplus

    Script:
    https://raw.githubusercontent.com/rena-jp/haremplusplus/main/release/haremplusplus.user.js

    There is one big change, the quick harem appears on the waifu page instead of the home page. When you open home, the game may receive rewards from cards or open a bundle shop popup. These conflict with Harem++ and cause a forbidden error. I decided to move the quick harem to the waifu page to avoid troubles. The waifu page does not automatically reload when the timer ends, unlike the home page.

    You may need to reopen the quick harem after the first load as the cache data format has changed. The second load should work fine.

    • Thanks 8
    • Hearts 3
  17. To be honest, I don't think most players understand shield skill mathematically. In league battles, you get points based on your remaining ego if you defeat your opponent. If your ego is over 90%, you get 25 points. If your ego is between 80% and 90%, you get 24 points. To get 25 points, you need to keep reduction of your ego below 10% of your max ego.

    Level 5 shield gives you 40% bonus. To get 25 points, you only need to keep reduction of your ego and shield below 50% of your max ego. The difference between 10% and 50% is 5 times. This is better than defeating your opponent one turn earlier and reducing the number of times you are attacked from 2 to 1. It is also better than using stun to skip two your opponent turns and reducing the number of times you are attacked from 3 to 1.

    Shield gives +40% bonus only for 16 points or less. It gives +50% bonus for 18 points, +100% for 22 points, +200% for 24 points, and +400% for 25 points. This is a mathematical feature of shield.

      Reduction of ego (and shield)  
    Points No shield Level 5 shield The difference
    25 below 10% below 50% +400% (50% / 10% = 5)
    24 below 20% below 60% +200% (60% / 20% = 3)
    23 below 30% below 70% +133% (70% / 30% = 2.33)
    22 below 40% below 80% +100% (80% / 40% = 2)
    21 below 50% below 90% +80% (90% / 50% = 1.8)
    20 below 60% below 100% +67% (100% / 60% = 1.67)
    19 below 70% below 110% +57% (110% / 70% = 1.57)
    18 below 80% below 120% +50% (120% / 80% = 1.5)
    17 below 90% below 130% +44% (130% / 90% = 1.44)
    16 below 100% below 140% +40% (140% / 100% = 1.4)
    • Like 1
    • Thanks 3
×
×
  • Create New...