I'm sidetracking a little this week.  Understanding how magazine detonations can be mitigated and/or exploited hinges on the ability to land hits reliably against the magazines themselves.  Yet this game provides no direct feedback to say when a given module is struck.  How can you determine if a given weapon attack is successfully damaging a magazine or not?  The best we can manage is to understand how various attacks behave when they strike a ship and extrapolate from that.  Specifically, I wanted to better understand just how much of an area-of-effect high explosive attacks from HE shells, bombs and torpedoes had and potentially uncover how much damage they did to modules.   Data-mined sources have long listed a statistic called "Explosion Size" for all HE attacks.  Minotaur's torpedoes have a 1.2 rating.  Hood and Warspite's 381mm HE shells have a 2.23 rating while a USN 127mm shell has a 0.38 rating.  All of these statistics are easily drawn up from third party sites like gamesmodels3d.com.  Without knowing what these values were measuring, they didn't mean much.  Yes, it's nice to know that a 381mm Battleship shell has almost 6 times the size of a 127mm USN shell, but what was that size even measuring?  Radius?  Diameter?  Area of a circle?  Area of a cube?  Area of a sphere?   To test this, I selected a ship with a rather cluttered deck with low-health modules -- specifically the C-Hull Izumo.  Her decks are strewn with tight clusters of single-mount 25mm guns that, according to data sources, have a mere 200hp.  Practical tests show that a direct hit from a single 203mm IJN shell will kill them every single time, with multiple being killed if the shell lands near enough to them.  By taking a Furutaka with her single-gun turrets, it should (in theory) be possible to get an approximation of how large an area these 203mm shells were affecting.  So I loaded up the training room and began shooting at Izumo and taking measurements.   Bow anti-aircraft mounts on C-Hull Izumo -- the perfect test bed for AOE blast radii.  The deck is similarly clustered across the length of the ship, giving plenty of test-area.   Example of 203mm HE shell damage from a single hit.  Two AA guns were destroyed yet those around it were not.  Repeated tests helped define what the minimum blast radius had to be.   From the Q&A with Sub_Octavian, we know that all modules within the blast radius are equally affected, short of armour providing some kind of protection.  For deck element AA guns, there exists very little to no protection at all.  Simple test firings in this manner gave an indication of how big the blast had to be and how big it could be.  This was, of course, assuming that the blast was circular or spherical in shape (so help me, if it's a cube or a cone, I will lose my head).    After littering the decks of several Izumo with 203mm impacts, I then moved up to 381mm guns on Hood.  I got around the problem of two-shell impacts in the same area muddying results by standing off at range and lobbing HE shells onto Izumo's decks.  Dispersion ensured the shells never landed near each other.   It was immediately shocking how much larger the area of effect was from 381mm HE shells.  This indicated that the data mined value were not calculating area, but rather the radius or diameter of the blast.   With all of these test results, I went back to the data mined statistics and tried to make sense of them again.  The breakthrough came from a pair of derp moments.  The first derp was that I knew I had seen a similar style of measurement, in non-descript units in other stats and that came up from AA guns.  Sure enough, the ranges of the AA auras were described in these units and they could be quickly converted to their known ranges.  This would have saved me some time had I realized it sooner.   5.0km = 167 3.0km = 100 2.5km = 83 2.0km = 67 1.5km = 50 1.2km = 40   This gave us a unit of measure -- 1 unit = 30m (or 0.03km).  From these, I could then plot out what the in game values.  So I began converting and plugging away.  This gave my 203mm Furutaka guns with 0.84 value a 25.2m radius.  This gave the 381mm off Hood an enormous 66.9m radius.  Now, the ships in World of Warships are twice the size they should be (Yamato measures over 520m in length in game, but she was only 260m or so in real life), but this still seemed a little big to me.  Using dseehafer's wonderful ship-scale comparison, I began plotting the hits onto these scale models and found my second derp immediately.   Furutaka in red, Hood in yellow.  The outer rings are from my initial calculation.  The inner rings are the corrected values.  Blue are Minotaur torpedo AOEs (at proper scale) from my tests in Part Three at the specified impact points.   My circles were twice as big as they should be.  The problem was that I had assumed the AA values were a radius when they appear to be a diameter.  When the values were halved, from 1 unit to 30m down to 1 unit to 15m, everything began lining up all hunky dory.  Is this actually correct?  The test data seems to support it.  From this I was able to make scale models of all of the blast sizes for the values I had and begin overlaying them on ships.  Some of the data was initially quite exciting, especially when it came to ideas on how to specifically target modules.   Maybe we should start throwing around more battleship HE at destroyers?   My early elation was quickly marred with frustration.  Initial tests using this new found data to try and pop the torpedo armaments of Gneisenau with Furutaka's 203mm guns (and thus try and isolate how much damage was being done per shell) yielded unexpected results and brought into question just how certain modules are damaged and destroyed.  Per Sub_Octavian's explanation, damage is initially fixed and then modified by the amount of armour protection surrounding a given module.  Thus, close range hits all striking the same place should (in theory) produce the same results.  Instead, I saw inconsistency.   It took as few as 3 hits to destroy Gneisenau's torpedo mounts.  It took as many as 39.  Between 6 and 8 were most common.  I wasn't missing -- I was causing critical damage on the torpedoes with my first shot in many cases.  So what gives?  This points to another significant gap in our understanding with how modules (specifically those with a working-damaged-destroyed mechanic) interact with high explosive damage.  The information we have in the wiki describes that as soon as a module is reduced to zero hit points, it is destroyed -- so why the disparity between three and eight hits needed to destroy it when the same place was being struck (never mind thirty-nine hits)?   This kind of inconsistency extended to deck elements too, like AA guns.  A hit that should have destroyed 4 AA guns instead destroyed 3, but instead of one of the outer guns surviving, it's one of the inner guns that somehow managed to dodge the attack despite clearly being inside the radius.  I can attribute some issues with modelling and/or collision errors, but talk about demoralizing.  If I can't accurately predict when external modules are being damaged (and with visual aids and critical hit indicators to assist), how can I accurately predict and calculate magazine damage?  I need more information about the hit-boxes of modules and how their critical hits work.   This, of course, calls for more experimentation.