18 réponses à ce sujet

[Rokzum]

 About to start a warrior, and I want to take towards Reaver or Templar, what stats and skills should i work on? Str, Con, Dex? 2hand weapons? Whats the cap on each stat? How much is enough for each one?

[AXidenT Gamer]

I know for DW warriors you should go 36 in dex, rest in strength...

[Theramond]

For 2-handers, I suggest pumping strength only to increase damage output and accuracy as much as possible (especially accuracy, as it is poor in the initial stages of the game. In such a case, use precise striking to compensate. As you pump strength, there will be less need for such compensation). I should also note that dex does not increase hit rate for the 2-hander. There is no need to pump con, willpower, or any other stat for that matter. Finally, I suggest going berserker-champion or reaver champion.

Templar-champion is good for sword and board tanking. Go 38 strength, 26 dex, and the rest con for tank. I must say that I've experienced incredible survivability when I made my PC a tank. The PC can easily out-tank Shale or Alister.

Go DW if you want to create a sole DPS warrior. Despite the looks of it, 2-handers appear to have a mix of various DPS, tank, and controller abilities. Don't get me wrong though, they hit good. DW warriors deal somewhat less damage in one hit. However, they are quicker and much more efficent. Consequently, those DPS numbers will better substantiate in the long run...

Modifié par Theramond, 18 décembre 2009 - 09:53 .

[Theramond]

Here's an example of 2-hander stats. This was my third 2-hander runthrough. I should mention that the stats on the bottom left were altered by haste

Modifié par Theramond, 19 décembre 2009 - 01:34 .

[atnexus]

My personal favorite is a dual wielding warrior. You need to have Dex of 36, Str of 42 (for the heaviest armor), The rest of the points I tend to put in Con, this way you're walking around with near 500 hps by end game. While you do less damage per hit you will hit a lot more times then either with a two hander or sword and board, especially once you get Momentum sustained ability, coupled with Haste it will be even more insane. Any other build will feel like its fighting in slow motion after that.

Modifié par atnexus, 19 décembre 2009 - 02:53 .

[dkjestrup]

Okay, theramond, sorry, but that really isn't the best way to build a weapon + shield warrior (pumping con). The best stat to pump on a tank is dex.



Talents wise, the most important skills to get are Shield Expertise (and consequently Shield Wall) Shield Mastery (and consequently Shield Tactics) Taunt (most important tank skill there is) and Threaten.



Specialization wise, you'll definitely want a Templar/Champion. Take Champion as your first spec. Templar isn't that great in itself, but it allows you to wear the Knight Commanders Armor (can be brought from that "best previously owned" shop on the Frostback Mountains, which gives 40% magic resist. Champion gives you Warcry + Superiority, for an AoE knockdown attack, great when surrounded and in need of a potion, and Rally+Motivate, for defense.



You should always be running Threaten, Shield Wall and Rally, as sustainables. Save all your stamina for taunting, and for shield bashing enemies that are overwhelming your allies.



Stats wise. You want minimum possible strength. Everything else in dex. For strength, either get 42 strength, so you can wear Wades Dragonbone Plate armor, for 40% fire resistance, and Imperial Legion armor. OR get 39. You need 39 to wear Knight Commander's plate, and one of the best, if not the best, armor in the game is the Juggernaut armor, which requires 38 strength. So by getting 39 strength, you can protect yourself from magic and melee. If you really struggle with the dragons without Wades gear, use Shale or Alistar as your tank, but to be frank, even on Nightmare you don't really need the fire resist gear. However, 42 strength gives you access to the best shield in the game, Howe's Shield. So really, it's just a matter of judgement. I personally go for 42, but you would do fine with 39.



Now keep in mind you can use items to equip your gear. Helm of Honleath, Key to the City, Harvest Festival Ring, Vanguard, Earthheart's Portable Bulwark, Ornate Leather belt, all give boosts to strength. So does Shadow of the Empire, but it doesn't matter, because when you equip your plate, you won't have enough strength to equip your helm, and vice versa, kinda hard to explain. That leaves you with a +11 bonus to strength. So instead of 42 strength, you only need 31 strength (or 28 if going for 39). Take away the +4 bonus you get from the fade. That leaves you with either 24 or 27 strength from leveling.



Required stats (NOT including the Fade)



Strength: 27 (or 24)

Dexterity: Everything else



You may want to get 20 con if you're really struggling, but you shouldn't need it with the fade bonuses.

[Theramond]

dkjestrup wrote...

Okay, theramond, sorry, but that really isn't the best way to build a weapon + shield warrior (pumping con). The best stat to pump on a tank is dex.

Since you seem to be so certain of this, can you substantiate your claim with qauntative evidence that dex is better than con? I would just like to see some numbers for comparison.

Modifié par Theramond, 19 décembre 2009 - 06:13 .

[vicariouscheese]

the numbers are in a different thread

but basically if you pump dex, youre 90% unhittable by white/yellow mid game. early game con is better for obvious reasons, but by mid/end game, it comes down to 300 health and only getting hit 1/10 times, or 500 health and getting hit pretty much every time, and its not that long before enemies do 50 dmg easy and that 500 goes to less than 300 in no time with mobs

[atnexus]

Thats why you get massive armor, when you're shaving off 35 points of damage off every hit, most mobs hit you for very little.

[Theramond]

vicariouscheese wrote...

the numbers are in a different thread
but basically if you pump dex, youre 90% unhittable by white/yellow mid game. early game con is better for obvious reasons, but by mid/end game, it comes down to 300 health and only getting hit 1/10 times, or 500 health and getting hit pretty much every time, and its not that long before enemies do 50 dmg easy and that 500 goes to less than 300 in no time with mobs

Can you post the link? In the meantime I performed my own experiment and statistical analysis. While my data is not based on algorithms, it is based on real-time data. I ran three experimental "groups", to include a control "group." I ran three trials for each group. Each trial (which was video recorded) consisted of attacking the same group of bandits located in the wheat fields at Lothering. The time it took (from when the first attack occured) to when the PC died was approximated. All other possible variables that could be controlled were controlled. That is, the same equipment was used, and the same talents and skills were allocated and activated.

The PC for all groups was leveled to level 20 (I was using the developer console). For the Control trials, all points were allocated to strength. For the Str_Dex trials, 38 points were allocated to strength, while the remaining points were allocated to dexterity. For the Str_Dex_Con trials, 38 points were allocated to strength, 26 points to dexterity, and the remaining to constitution. For the Str_Con trials, 38 points were allocated to strength, while the remaining points were allocated to constitution.

Of course there are some weaknessess in my experiment. I have a small sample size, and the behavior of the AI is also highly variable, which may confound the data. The results were as follows:

[Edit] I brought the estimated times down to a hundredth of a second. The results came out slightly different, but it is much more accurate...

Modifié par Theramond, 20 décembre 2009 - 07:41 .

[Takrandro]

Whoa! if this were about world of warcraft, i'd not be suprised, but this is a single player game,and while it is interesting, u don't need an ultimate character to beat certain bosses with other players at your side... while it can be fun, i rather go with what my gut tells me is the best, and not try to overdo it,since doing so takes away all the challenge there is in this game...\\



Just an opinion.

[Theramond]

Takrandro wrote...

Whoa! if this were about world of warcraft, i'd not be suprised, but this is a single player game,and while it is interesting, u don't need an ultimate character to beat certain bosses with other players at your side... while it can be fun, i rather go with what my gut tells me is the best, and not try to overdo it,since doing so takes away all the challenge there is in this game...

Just an opinion.

I am taking an experimental/statistics course next semester, so maybe I am preparing, lol. Anyways, I've done about 5-6 playthroughs and this renews my interest. Anyways, there are so many people who have questions about what each stat does. This is a way of giving a more concrete answer that is based on raw, real-time data rather than heuristics.

Modifié par Theramond, 20 décembre 2009 - 12:49 .

[Theramond]

From a couple more experiments I ran, it looks like the contribution dex and con attributes give to longevity are rather miniscule over all. Going all dex, or some dex and con is not going to make that much of a difference, as they both seem to be equally beneficial (the differences are within milliseconds). What appears to matter most is the type of equipment that you are wearing (and of course the talents acquired). So, as far as attributes go, the best route is probably to get strength up to 42 somehow, so that you can wear massive dragonbone armor...

Modifié par Theramond, 20 décembre 2009 - 01:07 .

[Nick the Weregoat]

Theramond wrote...

vicariouscheese wrote...

the numbers are in a different thread
but basically if you pump dex, youre 90% unhittable by white/yellow mid game. early game con is better for obvious reasons, but by mid/end game, it comes down to 300 health and only getting hit 1/10 times, or 500 health and getting hit pretty much every time, and its not that long before enemies do 50 dmg easy and that 500 goes to less than 300 in no time with mobs

Can you post the link? In the meantime I performed my own experiment and statistical analysis. While my data is not based on algorithms, it is based on real-time data. I ran three experimental "groups", to include a control "group." I ran three trials for each group. Each trial (which was video recorded) consisted of attacking the same group of bandits located in the wheat fields at Lothering. The time it took (from when the first attack occured) to when the PC died was approximated. All other possible variables that could be controlled were controlled. That is, the same equipment was used, and the same talents and skills were allocated and activated.

The PC for all groups was leveled to level 20 (I was using the developer console). For the Control trials, all points were allocated to strength. For the Str_Dex trials, 38 points were allocated to strength, while the remaining points were allocated to dexterity. For the Str_Dex_Con trials, 38 points were allocated to strength, 26 points to dexterity, and the remaining to constitution. For the Str_Con trials, 38 points were allocated to strength, while the remaining points were allocated to constitution.

Of course there are some weaknessess in my experiment. I have a small sample size, and the behavior of the AI is also highly variable, which may confound the data. The results were as follows:

[Edit] I brought the estimated times down to a hundredth of a second. The results came out slightly different, but it is much more accurate.

I can't get youtube from here, so I'm guessing you explain in further detail the testing process.  All well and good.  1:  Con makes it so you can take more damage, period.  2.  Dex makes you avoid melee/ranged attacks more often.  I will assume for the test you had a full-health warrior getting attacked by something in a certain gear set and the amount of time is how long it took them to die, since I can't watch the video, I am going to sound like an idiot, but I already clicked reply with quote, so I'm doomed to my peril.

1:  If multiple things were hitting the character, how many at once?  Did you start the clock when the first one hit them, or when the last one started hitting them?

2:  If I'm reading the chart right, and those are times (in seconds) the character survived, does this account for special abilities?  What about magic spells?  I'll assume that resistances were the same on all trials.

3.  Physical and Mental resistances play into the role of tanking.  If I'm sleeping, knocked down, stunned, or otherwise comprimised, I'm not at my peak.  I can't chug a potion, interrupt a cast, or taunt when a pack of pissed off gingerbread men join the fight.

4.  I like a little Willpower in my build.  Being able to use all four attacks, a taunt, and a few more attacks without all of my rejuvenation cooldowns used on my tank let's my healer keep my party active.

5.  This is a single-player game, and an easy one at that, where tactics and cone of cold make enough of a difference you don't need perfect stat placement and ideal gear selection in order  to go to TotGC.  Play to what suits you.   Sometimes you're going to get knocked down because you fail a physical resistance check, and have some meanie use his lightning flame-thrower thing like some sorta punk diablo wannabe, and a bunch of rogues are gonna be hitting you while this is happening,and it's gonna drain your stamina so you can't get that last stun off.  If you want to make it through that, plan accordingly.

[Theramond]

Nick the Weregoat wrote...

I can't get youtube from here, so I'm guessing you explain in further detail the testing process.  All well and good.  1:  Con makes it so you can take more damage, period.  2.  Dex makes you avoid melee/ranged attacks more often.  I will assume for the test you had a full-health warrior getting attacked by something in a certain gear set and the amount of time is how long it took them to die, since I can't watch the video, I am going to sound like an idiot, but I already clicked reply with quote, so I'm doomed to my peril.

1:  If multiple things were hitting the character, how many at once?  Did you start the clock when the first one hit them, or when the last one started hitting them?

2:  If I'm reading the chart right, and those are times (in seconds) the character survived, does this account for special abilities?  What about magic spells?  I'll assume that resistances were the same on all trials.

3.  Physical and Mental resistances play into the role of tanking.  If I'm sleeping, knocked down, stunned, or otherwise comprimised, I'm not at my peak.  I can't chug a potion, interrupt a cast, or taunt when a pack of pissed off gingerbread men join the fight.

4.  I like a little Willpower in my build.  Being able to use all four attacks, a taunt, and a few more attacks without all of my rejuvenation cooldowns used on my tank let's my healer keep my party active.

5.  This is a single-player game, and an easy one at that, where tactics and cone of cold make enough of a difference you don't need perfect stat placement and ideal gear selection in order  to go to TotGC.  Play to what suits you.   Sometimes you're going to get knocked down because you fail a physical resistance check, and have some meanie use his lightning flame-thrower thing like some sorta punk diablo wannabe, and a bunch of rogues are gonna be hitting you while this is happening,and it's gonna drain your stamina so you can't get that last stun off.  If you want to make it through that, plan accordingly.

1. I "started the clock" when one of the bandits took a swing, regardless if it dealt damage or not. The other variability that you might notice in the video is closing distance. Sometimes bandits seem to engage and surround the PC quicker. Other times, this occured perhaps milliseconds slower. I attempted to offset such variables by attempting three trials for each group, and calculating the mean.

2. The only changes to the PC across groups were attribute changes. That is the same talents, equipment, and special abilities were used across trials. Shield wall was activated, no other abilities were used except for passive abilites which were the same. There were no companions involved, and subsequently no magic.

3. Right, and certainly the attributes will alter some of the resistences, which may account for some of the difference between experimental groups. The dependent variable is longevity, the independent variable (thing being manipulated) is attributes. Activated talents were not considered, because that is not what is being measured in this experiment. When you design an experiment, you want to isolate all other variables that have a potential causal relationship. If I had activated talents, the data would have certaintly been confounded (not just because there is a causal relationship, but activated talents themselves have considerable variability). Sure, it lacks ecological validity (meaning that it does not replicate actual gameplay circumstances), but I think the data is still valid.

4. I think what I have done so far (excluding subsequent experiments which I leave to myself) indicates that what matters most is getting the right armor on, which means allocating points for strength. The rest doesn't seem to matter as much. Go all dex, some dex and con, or perhaps will, it won't make much of a difference unless if you do something stupid like spend the rest of it on magic. Then we are talking about slight disadvantages here. What I am basically saying is that attributes don't have much of a global effect as one might think, although it does contribute to the overall effectiveness of the PC, and allows the PC to acquire armor and talents which may be more beneficial than just the attributes themselves. This is almost along the lines of what you are saying, is it not?

5. Well, I don't see any harm being done in trying to figure out what makes the best tank, rogue, etc. I have been trying to figure it out for some time by just playing through the game, but attributes still seem ambiguous. I was also wondering if dex is more beneficial than con, or vice versa, which none of the tank-building guides clearly answer. So, I've attempted to eliminate that ambiguity. Also, after about 5 playthroughs, I am trying to find things to renew my interest. So if I can create an uber tank and put the game on nightmare mode, that might keep me going for a couple more playthroughs. Maybe if he's uber enough, I will try playing solo, who knows.

Modifié par Theramond, 20 décembre 2009 - 05:29 .

[atnexus]

A note on Nightmare mode, I am not sure if there anything nightmare about it, you should just try it and see what happens, For the first few hours I played I had to keep checking my options to make sure it was set on nightmare because it was just so easy. Ultimately, if you're good at tactics and micromanage your party during combat, it should be no different than hard difficulty. I can't compare it to easy or normal, never played on those.

[Daniel Thorne]

To theramond, your statistics are a bit wrong. You need at least 30 trys in each category and you have three. You cant do statistics with so few observations.



And remember there are three kinds of lies.

Lies

Damned lies

and statistics

[vicariouscheese]

atnexus wrote...

A note on Nightmare mode, I am not sure if there anything nightmare about it, you should just try it and see what happens, For the first few hours I played I had to keep checking my options to make sure it was set on nightmare because it was just so easy. Ultimately, if you're good at tactics and micromanage your party during combat, it should be no different than hard difficulty. I can't compare it to easy or normal, never played on those.

since youve only played hard and nightmare, nightmare doesnt seem like much of a jump
basically the difference between normal and nightmare isnt that noticable once you understand the game mechanics.  the difference between easy and normal is huge though.

look at this chart
http://dragonage.gul....php/difficulty

[Theramond]

Daniel Thorne wrote...

To theramond, your statistics are a bit wrong. You need at least 30 trys in each category and you have three. You cant do statistics with so few observations.

That is something I have noted previously in this thread, along with other weaknesses. It is what it is...

Modifié par Theramond, 28 décembre 2009 - 08:27 .