Y'a plus de poneys. Ni de worst, storm, midnight, last squadron, supernova, consortium, arakeen, cinis, complétez la liste de tous ces gens sur qui on aimait bien taper... et qui ont été vaincus par Arcann et ses sbires. Je vous propose de faire ce topic un genre de mémorial à nos ennemis disparus

Dernièrement, lorsque le bonus d'xp de commandement est sur le jcj, on a le droit à des BG 70% noob avec des bus qui parcourent la map comme à la grande époque. Et puis on voit des fails curieux, par exemple des gars qui désactivent leur propre point sur odessen, des mecs qui "glissent dans le vide" sur quesh ou sur voidstar au terme d'un tp en vitesse de force, et puis également des joueurs expérimentés qui n'en ont rien à faire de gagner le BG mais qui veulent leurs frag solo (surtout les agents) Il y a aussi des epic win dans un sens comme dans l'autre (sur Odessen c'est n'importe quoi des fois). On retrouve des fufus tout mignons qui viennent mez + tag + flash +tag (ou qui essayent). Et puis des gens qui sont pas au courant pour le gaz dans les arènes, qui fufutent et qui crèvent dans leur coin.. Du coup je me permet de ressuciter ce topic (mon topic) car je suis sûr que plusieurs d'entre vous auront des anecdotes sympa à rigoler ou des bonne tranches de racontade à rapporter

Hi! Here is the original post from KeyboardNinja. Can't get the spoilers back, but see #1 for tables. --------------------------------------------- Update (Mar 12): Kinetic Bulwark/Dark Bulwark calculation has been revised and made much more accurate (the value is slightly lower than assumed; special thanks to dipstik!). Self-heal contribution to survivability has been corrected to post-mitigation (rather than pre-mit, as before). Self-healing is no longer overvalued on bosses with single-swap mechanics. Corrupter Zero's Sweeping Slash has been corrected and is now considered a force/tech attack. Defense minima for PvP gear are now based on Obroan rather than Dread Forged. Update (Dec 26): Relics are now part of the fundamental calculation; you no longer need to include them in your stat budget! Riot Gas has been dropped from the Vanguard bonuses (reasoning long and dull). Defense minima have been reworked, resolving a serious bug where vanguard minima were underestimated while shadow minima were overestimated. --- Using a slightly different approach from dipstik, I modeled the three defensive stats against each other and computed optimal distributions at various stat budgets from 600 to 5000 (for reference, augmented Arkanian has a stat budget of 2181; Dread Forged a stat budget of 2721). These stat values should provide accurate itemization targets for guardians and vanguards, and mostly accurate targets for shadows (Endurance is not considered here, which is a significant omission where shadows are concerned). The fundamental function that I maximized was as follows: mrke * (6/23) * (1 - (1 - defense[x*a + relic] + 0.5*0.1)*(1 - (1 - crit)*shield[x*b]*absorb[x*c])) + mrke * (1 - 6/23) * (1 - (1 - defense[x*a] + 0.5*0.1)*(1 - (1 - crit)*shield[x*b]*absorb[x*c])) + ftke*(1 - (1 - resist)(1 - (1 - crit)*shield[x*b]*absorb[x*c])) (where a + b + c = 1 and a,b,c >= 0) (note: m/r attacks are assumed to have 100% accuracy. low-accuracy attacks are modeled via a defense addend of 5%. bosses cannot crit) The constants in this equation, mrke and ftke, are the weighted algebraic mean computed damage ratios in Dread Fortress or Dread Palace. The damage ratios are sufficiently different between these operations as to merit separate stat distributions (which are given below). Note: if you're running the old operations (S&V or TfB), you can still use these distributions. To optimize for S&V, use the values from Dread Fortress. Likewise, to optimize for TfB, use the values for Dread Palace. The damage distributions are fairly close between the respective operations, or at least close enough that it doesn't matter. Shield, defense and absorb ratings are all constrained to within ranges as a function of your current stat budget. These ranges were computed by elidion in an attempt to accurately approximate the discrete itemization minima provided by BiS ear, implants and enhancements for all classes. Note that these minima are actually a function of the class in question, since the BiS implants for shadows carry absorb, while vanguards and guardians get defense. With these calculations, I generated a set of realistic (achievable with itemization) target ratings for defense/shield/absorb at all stat budgets 600-5000. Additionally, I plotted the ideal stat distributions for all three tanks. Enjoy! Using The Tables To determine your target itemization, add up the defense, shield and absorb ratings from your character sheet including stim. This is your current stat budget. Using this, find the nearest entry in the appropriate table below for your class and stat budget. Your goal is to get your defense, shield and absorb rating as close to these recommended values as possible by swapping mods, enhancements and augments. This will maximize your achievable mitigation from defensive stats. Note that these tables were all generated assuming use of the best in slot relic types: Reactive Warding and Fortunate Redoubt. These tables are not fully accurate with other relic combinations! They will be close, but not quite correct. Because these relics are included in the optimization process, there is no need to consider the relic proc value in your stat budget (as was previously required). Simply use your character sheet values; no need to calculate! As an example, imagine that my shadow's character sheet shows a defense value of 520, a shield value of 1100, and an absorb value of 900 (with stim activated). Thus, my total stat budget is 520 + 1100 + 900 = 2520. Rounding to the nearest value in the "Average" table below for shadows, we find the following entry: {2500,{defense->455,shield->863,absorb->1182}} My current defense rating is 520, which is too high by 65 (almost exactly 1 mod). My shield rating is 1100, which is too high by 237. Since the only way to increase shield rating over the minimum is to stack shield augments, this should be a very easy problem to fix. Finally, my absorb rating is 900, which is too low by a whopping 282 points. If I swap one mod from defense to absorb, and then swap all seven of my shield augments to absorb, I should be just about perfect. I'll end up a little low on defense and a little high on absorb, but it should be good enough. In most cases, it is impossible to achieve precisely the optimal stat ratings. Just get as close as you can. Average M/R+K/E: 71.7760% F/T+K/E: 24.3432% F/T+I/E: 3.8806% DtPS: 6336.47 For those who don't want to carry around two gear sets… Shadow Graph Guardian Graph Vanguard Graph Dread Fortress M/R+K/E: 74.6510% F/T+K/E: 20.4436% F/T+I/E: 4.9040% DtPS: 5987.66 Shadow Graph Guardian Graph Vanguard Graph Dread Palace M/R+K/E: 68.5082% F/T+K/E: 28.7756% F/T+I/E: 2.7163% DtPS: 6685.28 Shadow Graph Guardian Graph Vanguard Graph PvP M/R+K/E: 30.6% F/T+K/E: 54.1% F/T+I/E: 15.3% DtPS: 7000 These numbers are based on a spreadsheet tabulated by Ilmar. As he is a Scoundrel DPS and not a tank, the exact percentages are going to be a bit different for tanking. However, these still represent the most authoritative values we have. Assuming a 25% natural crit chance (on average) and adding the auto-crit attacks found in Ilmar's spreadsheet, we assume a 33.25% crit rate overall (thus reducing the value of shield). The 7k DtPS value roughly approximates two geared DPS focus firing with all cooldowns available over a span of roughly 12 seconds. Average values over an entire match are clearly much lower than this, but also much less interesting since the healer (and tank) is primarily concerned with who is being bursted and by how much. Shadow Graph Guardian Graph Vanguard Graph One somewhat hilarious thing you will notice in both the stat budgets and the graphs is that the ideal defense rating actually decreases as your stat budget increases at a certain point. This is true for shadows and vanguards, but vanguards show the most precipitous drop. This is not a typo or a mistake, and has been verified formally and justified with informal arguments. It does make sense from a mathematical standpoint, but it's a little bizarre. Just go with it. :-) For reference, this is the Mathematica notebook used to generate this post. Overall Survivability Note This section previously counted self-healing pre-mitigation. This is inaccurate, as any self-healing mechanic is applied to damage taken post-mitigation. Thus, self-healing mechanics were previously undervalued by a substantial margin. This error has been corrected, resulting in a slightly different balance picture than before. What follows is the comparative tank survivability including all buffs, damage and self-heals in each operation (higher is better). In other words, you should be able to predict your net external healing required by multiplying an operation's DtPS by 1 - # where "#" is the survivability value from below. For PvE survivability, I'm assuming full 78 armorings and a stat budget of 2721 for each tank (in other words, full min-maxed Dread Forged). For PvP survivability, I'm assuming full min-maxed Obroan with maxed augments and a stim. Expertise is ignored as it only functions to cancel out the damage bonuses from another max-geared player. Defensive cooldowns are not included in the calculation! Dread Fortress Shadow: 73.4956% Guardian: 74.4393% Guardian (hybrid): 75.1821% Vanguard: 72.8545% Dread Palace Shadow: 73.3933% Guardian: 74.1169% Guardian (hybrid): 74.9887% Vanguard: 73.1754% PvP Shadow: 57.4008% Guardian: 58.0441% Guardian (hybrid): 58.0374% Vanguard: 58.1033% Or, if you prefer the graphical approach, here is what the relative balance looks like. Note that this is a relative graph, meaning that the best tank for a given piece of content is taken to be 1, and all other tanks are a percentage of that value. If all tanks are balanced to within 5%, then all of these bars should be at 0.95 or higher. Overall, balance is surprisingly tight. Guardians sit on top for current content because of how low damage levels are and the non-scaling nature of self-healing mechanics. Increasing boss damage by 15-20% (generally what is seen in Nightmare Mode) puts all three of the tanks on almost exactly even footing. Shadows are about 1.5% behind Guardians due to the impact of self-healing. Vanguards are a little over 2% behind Guardians for reasons of…their own. It's not clear exactly why this is, especially since Vanguards have the worst defensive cooldowns of the tanking classes. Despite their currently superior survivability, double-Guardian is the worst tanking combination by far, since Guardians do not provide the damage debuff (provided by Shadows in the form of Slow Time, or by Vanguards in the form of Ion Cell). Guardians also suffer from very noticeable weaknesses on very specific bosses in the current HMs, such as Tyrans and Corrupter Zero, owing to their lack of shield bonuses and the specific nature of Saber Reflect. PvP balance is somewhat more varied, but not as disparate as most people believe. Vanguards are the best PvP tanks in terms of mean mitigation, ahead of Guardians by a mere 0.1% and ahead of Shadows by 1.3%. However, this is somewhat balanced seeing as Riot Gas does not work reliably in PvP as a defensive cooldown, meaning that Vanguards need to have a bit more in the way of baseline survivability. Neither Shadows nor Guardians suffer from cooldowns which are less effective in PvP, and thus we see a slight penalization in terms of static survivability. Note that in PvP, Guardians cannot benefit from the debuffs of another tank (assuming 1-tank arena compositions), which means that they must be considered in the full absence of the 5% damage reduction debuff provided by Shadows, Vanguards and Watchman Sentinels. Hybrid Guardians (as well as Shadows, for that matter) are currently omitted from the relative chart owing to the fact that the hybrid Guardian spec falls too far behind in the realm of defensive CDs to be competitive, and thus causes a very deceptive skewing of the relative numbers. A more complete analysis of the standard hybrid tank specs (most notably, Shadows and Guardians) is beyond the scope of this section. It's worth noting that these relative balance figures assume absolutely numerically perfect play on the part of each tank. A single GCD's worth of delay on Blade Storm for a Guardian or a single GCD early on a Kinetic Ward refresh for a Shadow and their survivability plummets. This might explain why Vanguards are somewhat behind the other two tanks, seeing as their active mitigation is far, far less susceptible to diminishment due to player error. Overall, tank balance in 2.5 represents an impressive achievement on the part of Bioware's combat team. I don't agree with every decision they make, but credit where credit is due. All three tanks are viable, and absent the aforementioned double-Guardian composition, any tank may be slotted into any ops group without disadvantage. Projected Survivability One interesting thing to look at is how well the three tanks do not just in content now but in future content. By making the assumption that both armor rating and damage levels are linearly correlated with stat budget, we can examine this question. Note that this is a fairly significant assumption! This graph looks at the projected survivability of each tank plotted over stat budget (shadows = blue; guardians = red; vanguards = yellow). As you can see, the Guardian lead on the other two tanks shrinks consistently as damage levels and gear improves. Shadows don't scale quite as well as Vanguards do (on account of the heavy armor vs light armor distinction), but all three tanks are still within a hair's breath of each other by the time we get to the absolute numerical limit on current stat scaling (a budget of roughly 5k). ------------------------- Disclaimer: this is the work of KeyboardNinja. I merely resurrected it back online. Hope this helps.

Alors d'abord, félicitation pour l'intro qui est assez sympa. J'ai deux remarques à formuler: -sur le fond, tu mets de temps en temps des petites annotations sur des moments clés, mais franchement ça va beaucoup trop vite, il faudrait au minimum repasser la scène au ralentit, ou même mieux, mettre un pavé dans les commentaires pour expliquer ce qui se passe, parce que là on voit juste que tu as roulé sur le mec. -sur la forme, il y a beaucoup de vidéos (non classées) où l'équipe d'en face ne semble pas vous opposer beaucoup de résistance et où vous lui roulez plus ou moins dessus. Ca aurait été plus intéressant a voir contre une équipe qui vous pose problème Votre team de classé est sans beaucoup de surprise au vu des classes FOTM: Deux PT, un sorcier (quoique), un assa. En revanche l'équipe d'en face, deux rava, un TE et un merco, honnêtement je ne m'y attendais pas. Et ils tiennent bien en plus! Par contre on ne comprend absolument pas comment le merco se fait tuer dans le premier match. Un instant il est à 50% de vie et quand la caméra ne le montre pas d'un coup son nom passe en noir. Dans le match suivant votre heal se fait tuer et juste après en un interval de temps très court c'est vous qui défoncez tout le monde c'est un peu magique. Je note au passage que votre assassin n'a pas l'air d'utiliser beaucoup la décharge. Bon a part ça ça fait surper longtemps que je n'ai plus croisé Liny en BG. Il est à la retraite? GG pour la vidéo c'était très intéressant de voir les choses de votre point de vue

Ça me regarde pas à la limite, mais tout de même l'une des plus grosses guildes du serveur qui part en fumée comme ça du jour au lendemain, on s'attendrait à quelques explications...

Tous les consortium sur qui j'aime bien taper d'habitude ont l'air d'avoir déguildé. Qu'est-ce qui se passe?

Voyez les gars je trouve ça dommage qu'on n'ait pas un topic pour déblatérer librement sur le pvp qu'il soit ou pas coté sur notre serveur préféré, alors pas de chichi, pas de tabous, sinon bien évidemment la charte de ce forum qui devra nous épargner les débordements éventuels. Vous savez ya des tonnes de topics qui mettent en valeur les exploits individuels au scores, de classements du meilleur joueur de tel ou tel type, mais ya pas grand chose sur les exploits des équipes. Et pourtant quand les gros premade se relachent et commencent à faire n'importe quoi histoire de faire un ou deux BGs sans pression, il y a parfois moyen pour une team bien coordonnée (ou très chanceuse) de renverser la partie. Alors je vous propose de mettre ici en avant les teams qui fonctionnent bien et de raconter des anecdotes, celles de vos exploits, ou celles de vos ennemis, les tactiques que vous avez mis en place et qui ont foiré, ou bien les monstrueux coups de bols que vous avez eu et les BG que vous avez gagné mais vous savez pas pourquoi. Par exemple je me souviens d'une hyperporte assez sympa, à 5 secondes du premier timer on perd les deux pylones mais on remonte la moitié de l'écart en farmant comme des brutes au mid (j'étais sur un perso coté empire je crois, c'est une strat qui marche bien coté empire^^) et puis à 5 secondes du deuxième timer on reperd les deux pylones, blasé hein? Alors rien à faire on continue de farmer mid, mais à 30 secondes de la fin perso je pars en def histoire de casser du fufu. Et là à 5 secondes du troisième timer l'équipe d'en face se fait tag son pylone et à la fin on gagne la partie. Je me souviens aussi d'une étoile du néant, en attaque je suis en tank je mets ma bulle sur un healer mais à peine on est descendus que 4 mecs d'en face viennent nous chercher jusque chez nous et mettent méga cher à mon healer. Du coup il se soigne et moi je prends tout à sa place. Comme je suis en attaque et que ça se passer à 50m du spawn, je reviens en 10 secondes, je remets ma bulle, je me refais dépop. Le healer tiens toujours. En tout j'aurai fais 5 fois la navette. Pendant ce temps là notre équipe ouvre toutes les portes les doigts dans le nez à 6 contre 4. C'était marrant de jouer les paratonnerres Bon ben là je m'aperçois que j'ai beaucoup parlé de moi mais faut bien raconter le truc de la perspective d'un joueur et comme je passe 90% de mon temps à tag PU... Mais je suis sûr que de nombreuses team ont de bons moment de franche rigolade et d'epic win ou fail à raconter, c'est à vous

J'adore me faire défoncer par des trolls pour une fois que j'essaie de faire dans le constructif. Pour mésinterpréter mon propos à ce point là ça relève du grand art

Hop là petit topic pour intercepter une partie de la grogne: la psychologie du dps est la suivante: * je veux faire le plus gros dps possible => je tape en priorité sur les cibles qui se laissent faire. Par conséquent la réaction de la personne qui se prend le focus devrait etre * Si je suis focus, je met un snare je me déplace, ou je mets un root et je bumpe. Pour les commandos filet+bump est très dissuasif. On distingue deux types de personnes focus: celles qui peuvent se heal et celles qui ne peuvent pas. Celles qui peuvent se heal ont tendance à retarder l'inéluctable en restant sur place et en essayant désespérément de se soigner plus vite qu'ils ne prennent de dégats. Celles qui ne peuvent pas se heal ont tendance à se résigner à aller au respawn et décident de burster sur place en attendant. Erreur: Dans tous les cas il faut bouger. Les classes qui sont souvent focus ont toutes des moyens de controle. Je controle, je me déplace. Si c'est un distant faire le tour du poteau suffit à décourager 80% des gens. Si c'est un cac, un snare à 70% assorti d'un bon bump vous garanti pas mal de répit, et se mettre derrière un poteau ne vous fera pas de mal non plus. 90% des DPS ont la flemme de vous suivre dans un endroit dont ils savent qu'il leur faudra plus de 15 secondes pour revenir taper les autres. Et ne perdes pas de vue que tout le temps que vous perdez à cavaler sans rien faire c'est aussi du temps gagné pour votre équipe, car les dps particulièrement collants ne penseront à rien d'autre et certainement pas à protéger les objectifs tant qu'ils ne vous auront pas dépopé

Nowadays in WZs, one can see that more and more assassins are currently leveling, and it becomes clear that in the players' mind this is the new FOTM class. I think devs should do something about it, because most does not even know how to play the class and keep fooling around, searching to ambush isolated players like operatives love to. Anyway this is leading to a big class disbalance in WZs, and I hope something will be done about it before it is too late

They can just close the X server queuing when their is enough affluence on the server, right? I am sure that passed 2:00 AM you won't bother being with 7 complete strangers rather than doing nothing at all or waiting forever for the next WZ

Alt-f4 is enough. I am on SSD hard drive so the game launch pretty fast. Compared to using the "leave WZ" button, I will maybe loose 20s at most. And as I stated it is highly improbable they will detect the alt+f4 because even if they did that it would not solve the problem. I took the unplug example, but I have many other options like for example tuning my firewall, and interrupt only enough to get disconnected do the servers page, so that I don't even have to close and re-launch the game. Sorry for you, but it is quick and easy to go around your proposition. In the arenas if you are considered a week class as I am and focused from the start by those F*** sneaky assassins, you often get 0 medals during the fight, which gives absolutely no valor or commendations at the end. If you force me to stay, I can maybe do a /stuck and wait for the end of the fight? By doing so I ensure I won't get replaced and my team mates will be 3v4 all along, which is quite deadly most of the time. Only solution is to set up a "no arenas" option, and you will soon see that non ranked arenas will get deserted pretty quickly because most players don't like it

+1 Currently devs are ruining the game, mixing groups totaly on random, you can have 4 heal vs 4 tanks, you can have 4 dps repu vs 2 heal and 2 tanks repu, all this is absolutely awful, and the main reason they are not filtering is to not make the queues longer... But they are just disappointing more and more people. With interserver they could set up intelligent filtering, and also filter on the solo ranked index, that is the only way to make WZ interesting again

Even if it was doable I would not want it because, often the leavers are the worst players and when they are replaced by good players the match can be reverted to a good one. Additionnaly I HATE arenas and will keep leaving if thet put me in. Now, to the point that interest you: If they put a debuff on me when I click "leave the WZ", but not on DC-ed, I will just do alt+F4 instead of leaving, in order to be treated as a DC-ed. Some games can intercept the alt+F4. I imagine they could do it, but don't think they will because it has huge impacts. But even if they did it, I would still have others options to be considered as a DC-ed. For example I can just unplug the ethernet cable until I am actually DC-ed, and they have no possibility of inferring if being DC-ed is volontary or not, or maybe by analyzing the traffic and observing that instead of becoming fairly slow before the DC-ed, there was just an abrupt shut-down... But that is very very complicated and would require a huge effort and probably be buggy as all the features they add, so I don't think they would take the risk. As a conclustion, they have no possibility to find out who was really DC-ed and who faked it, so if they set a "logout debuff" they have to put it also on DC-ed and it is unfair and would be impopular so they won't do it.

Hi! there are lot of info around there with very complicated formulas, that are probably already complicated for native english speakers, but which are even harder to exploit for foreigners, especially with all those abbreviations. I wish everybody could easily understand how mitigation works inside swtor, so that anybody can use the formulas to make his own interpretation. Or to say it the other way, I think the elite club of mathematician tank specialists is too small, and I think a little more vulgarization would help grow its basis. ----------------------- Damage types ------------------------- As far as I know, there are 6 damages types: - force/tech melee damages - force/tech range damage - force/tech damage over time - physical melee damage - pysical ranged damage - internal damage over time And there are about 4 mitigations mechanisms - defense (D) - shield/absorb (S/A) - resistance ® All mitigation mechanisms do not apply to all damage type, for example you can't parry nor defend against DOT. But I don't really know if you can parry a range tech damage. If there is an expert around there that could make an array, with the correspondancies? I am quite sure for not force/tech damage, I think it is: - physical melee damage => D,S/A,R - pysical ranged damage => D,S/A,R - internal damage over time =>R -------------------------- Mitigation Mechanisms ------------------------- When someone, boss or player tries to hit you, there are several tests that are made: X) HIT ROLL: accuracy vs defense This roll determines if you are hit. Actually there could be two rolls: one too see if the enemy misses (applies only when it's accuracy is less than 100%), and one to see if your defense works: Example 1 Attacker => accuracy 90%, Defender => defense 20% - probability to miss: 10% - probability to avoid: 20% => one roll hypothesis : total probability to hit: 90% - 20% = 70% => two rolls hypothesis : total probability to hit: 90% x 80% = 63% Example 2 Attacker => accuracy 110%, Defender => defense 20% - probability to miss: 0% - probability to avoid: 20% - 10 % = 10% => total probability to hit: 90% X) CRITICAL HIT ROLL: crit (and surge) There is no possibility of parrying a critical hit. Once we know the attacker have successfully hit, we have to determine whether it is critical. For example, if the attacker has a 30%/80% crit/surge stats: The probability of a critical hit is 30% and the damage are 1.8 x the normal damages a non mitigated attack would have done Ok so for attacker1 the total probability of a critical hit would be 63% x 30% = 19% and of a normal hit 44%. For attacker 2, the total probability of a critical hit would be 90% x 30% = 27% and of a normal hit 53%. X) PARRY ROLL: shield (and absorption) The attacker has managed to hit, this is not a critical hit, it is time to see if we can parry it. The probability is determined by the shield percentage while the absorption determine the amount of mitigated damages. For example if the incoming attack was of 10.000 dmgs and our tank have 45%/30% shield/absorption, there is a 45% chance that it will absorb 30% of the attack, so over 100 clean non critical hits, 55 will inflict 10.000 dmgs, and 45 will inflict 10.000*70% = 7000 damages, for a total of 865k inflicted damages against 1M dmgs that would have been inflicted on a non tank. As a conclusion, one can consider that on an average base, the tank will absorb 45% x 30% = 13.5% of incoming damages at this stage. X) RESISTANCE: skills, talent, class specifics and armor rating On all incoming damages, resistances ratio are applied depending on the damage type, the armor ratings, and a lot of other things like skills, talents and buffs. X) SPECIAL: pvp protection mechanism When a tank is guarding someone, it is as if every attack someone makes on the guarded target was also made on the tank, but the final damage taken are divided by 50% for both. For example the attacker can miss on the guarded target, but not on the tank, which will be the only one to take damages, or the attacker can crit on the guarded target, but not on the tank, etc. ------------------- Conclusion: ---------------------- In PVE: As far as I know, bosses don't miss nor crit, as a consequence on 100 incoming physical damages for a 20/50/30 tank: 20 are avoided, 40 are mitigated at a 30% ratio, 40 are taken as they are, total damage is 40*1.7=68 and you still have to apply resistance on those. In PVP: Against a physical attack coming from a player with 100% accuracy and 20% crit: On average, considering 100 hits: 20 are avoided, 20 are critical hits, 30 are clean normal hits, 30 are mitigated by 30%. But since most players use a lot of attacks that are not mitigable, or use tricks to have a 100% crit proability on their stronger attacks, the shield importance is way smaller than in PVE. This is why the defense/hp optimization is pertinent in PVP. ------------------------------------ That's all for my current understanding on tanks mechanisms, please correct my mistakes and add the missing details. Once we have all of this clear, the understanding of your formulas will be a lot easier