Skill census/analysis of the SPF Mat/Pat/Guardian build compilation

Swamigoon

Well-known member
Apr 18, 2020
149
394
63
I tip my hat (and genuflect with deep respect) to the folks that worked so hard (much of the work is still underway!) to migrate and preserve content from the old site. This thread is my attempt to do something that shows gratitude to everyone involved. More than just saying "thanks," I wanted to produce something tangible that would add value to the migrated content.

Regarding Items and Skills

I'm not the first to assert as fact this idea: that the two main factors that contribute to success playing a Diablo II character are, in order of importance: (1) good equipped items, and (2) wise investment of stats and skills.

For items, we're pretty much at the mercy of the fickle minor deity whom we reverently call "PRNG." Notice how the Item Find Thread gets so much attention. In that thread, we celebrate the "blessings" occasionally bestowed by Pseudo-Random-Number-Generator (peace be upon her). I wonder: are there any other threads here that get more contributions, views, and likes?

But skills - those choices are under each player's complete control! Regardless of item wealth or trading opportunities, we are all able to enjoy the awesome consequences of our skill selections. When we consider our characters' skill allocations, a favorite resource here seems to be the most popular of the "pinned" threads: the SPF Mat/Pat/Guardian Build Compliation. I love that thread. I'm so pleased it was migrated successfully. I consult it frequently for ideas and inspiration about how to plan skill point investment choices. I also enjoy exploring it just simply for admiration and entertainment.

When we browse through our SPF P/M/Gs thread, we observe many examples of build ideas that have become widely popular, because of how quickly they enable players to achieve success. In starting a new project a few months ago, I also selected to use such "easy" builds, so that I could quickly advance my first few characters to patriarch/matriarch. But as @broxi_111 recently reminded us with #ThisBuildSUCKS challenge, trying to defeat hell difficulty Baal with skills and builds that are less popular, or even unique, is also very appealing.

What are some of the skills that have had little or no representation among SPF M/P/Gs? Do those merit further exploration? Also, are there interesting qualities yet to be discovered in the commonly used skills and builds?

Data driven skill census and analysis

To study those questions, I spent the last several weeks compiling and summarizing the skill investment levels reported in the SPF Mat/Pat/Guardian build compilation thread. I have attached the results here as seven CSV spreadsheet files, one for each hero class. I used the following methodology to collect these data:

Every skill has a column in a spreadsheet; these are organized by skill tree columns. Every reported Mat/Pat/Guardian (hopefully I haven't accidentally omitted one) in the compilation has a row. Each cell contains a number between 1 and 20, showing how many skill points were invested in that skill for that character. If no points were invested in that skill, the cell is left blank. Sadly, despite my best effort, I am not able to make the data 100% accurate, but I plan to edit and update the attachments, in order to correct errors I find or others point out. Happily, for most characters, exact skill investments were reported in their Mat/Pat/Guardian post. (GoMule text dumps for the win!) Very often, I tried to use this handy check for data validity: I could verify that the total skill points spent were equal to eleven more than the character's achieved level.

Unfortunately, sometimes an M/P/G post only reported some of the skill point counts. A common phrase among these was "...and the usual one-point wonders." If I could find clues as to exactly what those skills were, I recorded them. (Example: if a post about a sorceress includes a description about teleporting past a nasty unbreakable immune, I inferred that a point must have been invested in both telekinesis and teleport, and thus recorded it in the spreadsheet.) But if I could find no such clues, I usually left a spreadsheet cell blank, rather than just guess at what was their actual skill investment. In principle, I wanted to err on the side of conservative accuracy, rather than speculate. (Example: when there was only one skill point unaccounted for, but I can't tell whether it was used for frozen armor, static field, or warmth as the one-point-wonder, therefore all three were left blank. Another example: if a post listed "...all remaining skill points in holy shield," but I couldn't figure out how many points that would have been because the character's Patriarch level was never noted, then I recorded 2 points in holy shield. Reason: there must have been more than one point invested, and the actual hard points in the skill may well have been higher, but the actual investment couldn't be determined.)

Also, I tried to give the benefit of doubt to each contributor, presuming all skill declarations to be accurate as reported. For example: if any of the Pat/Mat/Guardian posts reported more skills point invested than would have been possible for a character's reported level, I just assumed that the reported level was that at which hell Baal was first defeated, and the character must have gained more levels and invested more skill points since achieving Mat/Pat/Guardian.

Sometimes, I could find no skill data at all in a Mat/Pat/Guardian post. The most common reason for this was because the skill data was included as links to skill tree screen capture images that no longer exist; most of these are broken links to sites such as ImageShack. For a few of these, I could infer skill data based on other items in the report, such as skill damages copied from an LCS. But if I could find no skill data at all, I put a row of x's in the spreadsheet for that character, which, unfortunately, effectively excluded that character from the skills census.

After I searched through each Pat/Mat/Guardian post and filled out each spreadsheet, I analyzed overall investment of each skill with histograms. Over the next few days/weeks, I will post those statistics, along with some of my thoughts and observations. Feel free to add to this conversation if you're interested or you have thoughts of your own to share.

Preface: the default skills

Before I discuss invested skill points, please let me quickly mention the three "level 0" skills that every character always gets without investment: attack, throw, and unsummon.

0attack.png Attack (normal attack)

I use normal attack all the time with my favorite build: the druid hunter. If you aren't an Amazon, and you use a bow or crossbow, it may well be the only weapon-based attack skill available. Some other notable builds that might have to rely on this skill are single-tree builds without 0skill items, such as aura-only Paladins, "mojo" (curse tree only) Necromancers, mastery-only Barbarians, and passive Amazons. I think my favorite write-up of a normal-attack matriarch Amazon is this one: Zaira, by @Vildecor .

*Swamigoon stuffs tongue firmly into cheek* - A participation trophy goes to these posts in the M/P/G compilation, who also specialized exclusively in normal attack: Melonfarmer, by AJK and One, by @Drystan .

0throw.png Throw (throw equipped item)

Several of my untwinked characters successfully used this skill with fulminating potions against normal difficulty Andy. I also played a bit with an axe- and kinfe-throwing necromancer once, years ago, but he's still stuck in NM act 5. I believe it's pretty difficult for throw to be effective without some amazing gear and support. My favorite write-up of a thrower: Arien, by @Pb_pal .

0unsummon.png Unsummon

I use unsummon rarely, in very particular situations. One such situation is the Maggot Lair in act 2. I generally want my mercenary to help fight bugs in those tight passageways, but sometimes he can't because he's stuck behind one of my summoned minions. In those situations, it's usually best to re-summon the obstacle to a more useful spot farther down the passageway, but when fighting around tight corners, I often just need to unsummon instead. I also used unsummon recently for some Assassin pit run experiments, in which I needed to do some runs without bringing a supporting shadow minion into the pit.

-

Okay, now I'll look at the rest of the skills, starting with the first column in the Amazon's Javelin and Spear Skills tree, her physical melee skills. If you enjoy this thread and you want to chime in, or even take me to task about my findings or opinions, please do! - Swamigoon
 
Last edited:

Swamigoon

Well-known member
Apr 18, 2020
149
394
63
Oops, I forgot to attach the spreadsheets. Here they are.
 

Attachments

  • Amazon.csv.txt
    30.6 KB · Views: 9
  • Assassin.csv.txt
    21.8 KB · Views: 1
  • Barbarian.csv.txt
    29.5 KB · Views: 2
  • Druid.csv.txt
    19.9 KB · Views: 1
  • Necromancer.csv.txt
    24.1 KB · Views: 1
  • Paladin.csv.txt
    32.9 KB · Views: 2
  • Sorceress.csv.txt
    45.5 KB · Views: 4

Swamigoon

Well-known member
Apr 18, 2020
149
394
63
Jab

Amazon01.jpg

144 M/Gs invested in Jab. It's a common prerequisite skill, but is also frequently used as a one-point wonder to deal physical damage. Fifteen M/G's invested more than one Jab point; seven maxed it. Of those, I think my favorite reads were of @TopHatCat64 's Wahine and @Kitteh 's Hannah.

Impale

Amazon02.jpg

Only twenty M/Gs invested in Impale, and all but one did so merely as a Fend prerequisite. But one intrepid player, NagisaFurukawa, focused a character named Nagi entirely around this skill. The advice in that post is short and to the point: "Listen up, and listen up good, because I'm only saying this once. Don't EVER try to make one of these." Impale isn't Nagi's only unique skill, though; we will see her represented again later.

I should have paid attention! I actually tried one of these myself. She advanced as far as Hell Act II a few years ago, but she's still stuck there at level 76. I'm not very motivated to try to bring her farther along. Her damage rate with Impale is about the same as her damage rate with normal attack.

Fend

Amazon03.jpg

This skill seems to have a few SPF fans. Fifteen M/Gs maxed this skill. Of these, my favorite post was Grape's report about Charlotte. Also, Stony's write-up about Bodiccea made me laugh.

Complaints about the Dodge and Avoid skills causing problems with Fend are commonplace. However, I found one notable matriarch exception, Kimpii's Poke. She enjoyed maximum investment in both Fend and Valkyrie. Because of elite gear and a wildly expensive runeword, she didn't seem too unhappy about occasionally "whiffing" while fending.

Personally, I haven't yet tried a Fend build; I think I'll keep it in mind when I look for a new challenge someday.
 
Mar 16, 2020
4,294
334
83
I love stuff like this! It will be very interesting to see at a glance the skill distribution spreads across the documented M/P/Gs. I look forward to seeing the rest. Good work!
 
  • Like
Reactions: Swamigoon

Swamigoon

Well-known member
Apr 18, 2020
149
394
63
After the physical melee skills, in the next column are the Amazon's lightning melee skills.

Power Strike
Amazon04.jpg

Power Strike is both a prerequisite and a synergy boost to the other lightning skills, so it's not surprising to see it represented in 127 of our M/Gs. However, I only found one player that used it as a primary end-game skill - @Grisu was forced to make the best of it eleven years ago in a quirky tournament, giving us Guardian buraz . As far as I can tell, every other player that invested more than a point did so solely for the synergy bonus.

In the skill description, the only thing that makes Power Strike distinctive over the other lightning skills is its attack rating boost. Unfortunately, it doesn't actually improve attack rating at all. The Amazon's Cold Arrow and Fire Arrow skills also fail to apply a claimed AR boost. As far as I can tell, Blizzard never bothered to acknowledge nor try to fix these flaws; it is probably a 20-year-old bug. It's quite a regrettable shame; I would have very much liked to see players try builds that would have taken advantage of Power Strike's attack rating bonus, if it had just worked as advertised. Only the Amazon class seems to suffer from this bug. The AR boost does work as advertized with the Necromancer's Poison Dagger, the Druid's Fire Claws, and each of the Assassin's charge-up martial arts skills.

But the way things are, there seems to be no reason to build a Power Strike focused Amazon; the only reasons to invest in the skill are early-game for a little lightning damage, or late-game to synergize a better lightning skill.

Charged Strike

Amazon05.jpg

This is a strong contender for "the most loved Javelin/Spear skill."

At level 20, the skill releases seven little floating bits of lightning, each causing between 1 and 322 damage. My current matriarch Ritzy stabs enemies with the unique Titan's Revenge ceremonial javelins, boosting the skill level up to 24, and she also has 22 points invested in synergies, so her charged bolts each do 1-1312 damage. For large monsters (notably act bosses), if she is well positioned, she can get all seven bolts to collide with the target upon striking, which means she inflicts 7-9184 lightning damage per melee attack. I find it so very satisfying to watch Diablo's or Baal's health meter visibly shrink upon each of her attacks. With even more +skills gear, she could get eight or even more bolts per attack. With all possible synergies maxed, the maximum damage per bolt climbs to 2254; it climbs much higher with +skills gear and/or gear that reduces enemy lightning resistances. I found three level 99 matriarchs that each achieved that fully synergized pinnacle: @EatSleepDiablo 's PhysFury, frozzzen's Mio, and zgpmf's PureDamagezon.

As expected I found dozens of standard builds, that combined maxed Charged Strike and Lightning Fury. Some, like my Ritzy and like pharoah's Kalita, supplement those skills with a strong Valkyrie. Many others, like Kartongen's Vahana, add a Freezing Arrow skill to the arsenal.

Delightful variant Charged Strike write-ups that I like and recommend are Grape's Maiden and felixbavaria's Hope. A recent fantastic Charged Strike build that I really enjoyed reading is @Babyhell 's impressively detailed diary, showing how untwinked Frida manage to achieve Guardian.

Lightning Strike

Amazon06.jpg

Not quite as popular as Charged Strike, Lightning Strike still gets a lot of love among our matriarchs and guardians. I dabbled with this skill, but not extensively. It's not a fast boss killer like Charged Strike, because it's designed to be used against groups of monsters rather than single targets. In crowd situations it can be pretty effective. Moreover, watching its lightning arcs bounce around crowds of monsters is sweet "yummy" eye candy.

If you're interested, please read Doctor Clock's Leper write-up, to read how gear with "hit causes monster to flee" mods (or Terror spell charges) can enhance this skill.
 

Swamigoon

Well-known member
Apr 18, 2020
149
394
63
Poison Javelin

Amazon07.jpg


About two dozen characters invested heavily in this skill. According to what I read in the compilation, almost all of them did so mainly for its synergy boost to Plague Javelin. But @Maltatai 's Rapunzel made Poison Javelin a main feature of her build. It is a pretty rewarding read for such a long wall of text.;)

Poison Javelin lasts a long time. Maxed and with enough +skills, a monster might stay poisoned for a full minute or more, much longer than any other poison skill of any class. The poison duration alone makes the skill appealing, even without any synergy investment. Fully synergized, one successful attack delivers about ten thousand poison damage over that long time span. Equip some items with +skills, enhanced poison skill damage, and reduced enemy poison resistance, and that damage could more than double or perhaps even triple! I don't know how many hell difficulty super-uniques could be "one hit killed" with a single poison javelin strike, but I think it would be fun to try it on each of them at various /players settings.

Lightning Bolt

Amazon08.jpg

Most adventurers used this skill only for prerequisite or synergy purposes, but I did find two players that mentioned the skill prominently in their posts.

Here is Neksja's judgment of this skill for Jessica: "a decent one monster killer." She had no skill points at all in any of the melee skills, but maxed both Bolt and Fury.

The other matriarch was Alfonso the Great's Dragoness, who completely skipped the melee lightning skills. She specialized in Fend, with a maxed Lightning Fury to handle physical immunes. But while questing through hell difficulty, she received seventeen points in Lightning Bolt. It's not completely clear from the write-up, but it looks as though LF tended to consume Dragoness' mana too quickly, and LB's lower mana cost made it a better PI solution.

My observations: players that want to supplement Lightning Fury with another lightning skill tend to choose Charged Strike rather than Lightning Bolt. CS requires melee proximity risk, and also requires two more prerequisite skills than LB, but for those relatively small marginal cost it tends to do a lot more lightning damage.

Next up: Plague Javelin and Lightning Fury observations.
 

Pyrotechnician

Active member
May 11, 2020
312
84
28
Pennsylvania
I love this thread, very interesting analysis, I've now clicked on this thread every time you've posted, hahaha.

In regards to your point on how many hell difficulty monsters poison jav can kill in one hit, I think this would be a perfect time to mention SSoGs guide on the ability of poison javelin to OHKO Baal. I've searched and searched for it but I can't for the life of me find it on the old site, I'm just running into a ton of broken links. Anyone else remember that? I feel like it would be a fantastic thing to share again.
 
  • Like
Reactions: Swamigoon
Mar 16, 2020
4,294
334
83
In regards to your point on how many hell difficulty monsters poison jav can kill in one hit, I think this would be a perfect time to mention SSoGs guide on the ability of poison javelin to OHKO Baal. I've searched and searched for it but I can't for the life of me find it on the old site, I'm just running into a ton of broken links. Anyone else remember that? I feel like it would be a fantastic thing to share again.

Found it!

 

Jocular

Member
Jul 17, 2020
27
44
13
Found it!


Hmmm. Reading that thread and the talk regarding Venom+physically hitting baal with the Javelin rather than the cloud resulting in Venom's duration lasting for Poison Javelin's length instead of 0.4 seconds...does that also work for Rabies and Poison Dagger?

edit: Further on in the thread it says it does work with PDagger, but not Rabies. Okay, I really want to make a Poison Dagger build now!
 
  • Like
Reactions: Swamigoon

Pyrotechnician

Active member
May 11, 2020
312
84
28
Pennsylvania
Found it!


Fantastic. Thank you!

Hmmm. Reading that thread and the talk regarding Venom+physically hitting baal with the Javelin rather than the cloud resulting in Venom's duration lasting for Poison Javelin's length instead of 0.4 seconds...does that also work for Rabies and Poison Dagger?

edit: Further on in the thread it says it does work with PDagger, but not Rabies. Okay, I really want to make a Poison Dagger build now!

The interaction with venom and the poison skills is really pretty interesting and commonly overlooked. The reason it works so well with poison javelin is because venoms poison damage is only meant to last for less than half a second, but poison javelins duration is so long and it applies the venom rate across that entire duration. But yes, if you read the post, the javelin has to connect or else it won't work. It doesn't impact the clouds just the poison damage applied to the weapon itself.
 
  • Like
Reactions: Swamigoon

Swamigoon

Well-known member
Apr 18, 2020
149
394
63
Plague Javelin

Amazon09.jpg

Silya by felixbavaria, Chimaera by @TopHatCat64 , Artemis by @crawlingdeadman , Askunken by @Maltatai , and Yue_Ying by Harrid were my favorite exhibits of Plague Javelin.

Compared to Poison Javelin, Plague Javelin's higher damage rate defeats monsters a lot faster, even though its poision duration is significantly shorter.

This is one of the few attack skills that I actually prefer to have on the left mouse button, rather than the right. If I don't aim carefully at a monster and I miss, which often happens when I get click-happy with RMB, the poison cloud erupts somewhere at the edge of the screen, rather than at and around the monsters I want to engage. Before I can try again, I have to make my 'zon run away while I wait for a delay timer to expire. But having Plague Javelin on the left mouse button forces me to take extra care to "name-lock" an intended target (or mash the shift key) before I click, which improves my control over the location where the green billowing mist expands.

Lightning Fury

Amazon10.jpg

The other contender for most popular L&S skill. If you always want to be able to transmute Wirt's Leg again and again, run after run, Lightning Fury works pretty well for that. Long live the Cow King!

I'm finding great difficulty in narrowing the choices down to just a few favorite LF build write-ups. Try this: go to the Amazon M/P/G compilation post, find a hundred of the linked lightning fury posts, and write their names on a dartboard. Then bring a javelin and a blindfold to the zoo, sneak into the primate enclosure, blindfold and give the javelin to one of the monkeys there, and then tickle its ribs until it stabs one of the names on the dartboard. Then read that post. If you still aren't satisfied with what you read, go watch the YouTube highlights video that @Kitteh posted last year in his write-up for Serendipity. If you still aren't satisfied, you can at least be content that you were able to hear the correct pronunciation of "Lah-dee-dah" from one of your fellow SPFers.
 

Maltatai

Active member
Mar 17, 2020
448
222
43
Sweden
@Swamigoon
So you're calling the guardian thread a wall of text, what would you say about the actual story? :) Rapunzel did by the way use Lightning Bolt to great effect as well but it isn't mentioned much in the guardian thread. It is a one-point wonder on par with berserk in my opinion, excellent for taking out individual ghosts for example. Looks great too.

No discussion of poison, or plague for that matter, javelin would be complete without mentioning the awesome effect of venom, as it combines venoms damage per second with poison javelins duration, more on that in for example the beginning of this post:
https://www.purediablo.com/forums/t...of-two-untwinked-javazons.83/page-3#post-1109

Also, it's Askungen, not Askunken, lol :D Askungen is Swedish for Cinderella but Askunken would mean "the A-skunk", which would certainly be a good hint of the skill choices though. I'm not intending to be nitpicking, just to point out a very entertaining, and apt, mistype. Lastly, what an interesting thread.
Prime Minister Jim Hacker: "So your statistics are facts but my facts are merely statistics..."
 

Swamigoon

Well-known member
Apr 18, 2020
149
394
63
In the first column of the P&M Amazon skill tree are four magic spells.

Inner Sight

Amazon11.jpg

Inner sight reduces enemy defense by a fixed amount (not a percentage or ratio). For most, it's just a prerequisite skill. For the rest, I thought it was interesting to see such a wide variety of IS investment levels. Seven M/Gs maxed it.

This skill is also available to all players that choose to hire an Act 1 Rogue. At higher levels, she frequently casts this spell at skill levels somewhere in the teens. This mercenary option can help characters with low attack ratings hit more often and advance more successfully.

In the write-up for Cattleprod, AJK wrote what I think was the most persuasive explanation for maxing this skill and boosting it further with +skills gear: because it "completely negates the defense of the smaller creatures, and seriously cripples the defense of the tougher ones."

Other nifty builds that maxed Inner Sight were Guardian BigSword by @Cyrax , and @Maltatai 's Snövit. (Hopefully I spelled the character name accurately this time. I'm relieved that you didn't feel too insulted by the skunk pun typo, Maltatai. I just hope I didn't offend Askungen herself... 😟 )

Slow Missiles


Amazon12.jpg

The only benefit for spending additional skill points in Slow Missiles is to increase the spell's duration. The only matriarch that maxed SM was the same that maxed Impale: NagisaFurukawa's Nagi. This character had a situation where a long duration for Slow Missiles seemed justified. Impale's attack animation duration is somewhere around a full second (about 25 animation frames, give or take a few IAS breakpoints). With fast enemy missiles and a slow attack speed, the Amazon may remain stuck in an impale attack for so long that the missile is sure to connect and cause harm. Slowed enemy missiles brings the Amazon more opportunity to get out of the way; a long duration means the player shouldn't need to micromanage spells as much during a battle.

Decoy

Amazon13.jpg

Decoy gets another very wide variety of skill levels on our matriarchs and guardians. Each invested skill point increases its life and duration, but also decreases its mana cost. It's pretty much a disposable minion, so each player may have a different opinion for the "sweet spot" at which life/duration/mana cost is most effective. Some players may feel that a disposable minion shouldn't need much life, so a lower skill level is just fine. Other players would rather reposition the disposable decoy many times during battle, or even while scouting new territory. In those cases, higher skill investment would reduce drain on the blue bulb with all of those recasts.

Another reason some players invest more points in decoy is for the synergy bonus it gives for Valkyries. More life lets a Valk withstand more damage.

As I understand it, the code that implements the behavior of most monsters in the game gives a high targeting priority for decoys. In other words, when a monster decides how to behave, if a decoy is planted somewhere nearby, then there's a very high probability that the monster will target the decoy rather than any other nearby hero or minion. I don't think this applies to all monsters, though - I think there may be some super uniques which are more likely to ignore the decoy and choose to attack something else.

I remember laughing after I read a post by someone (sorry, I don't remember the source) who ranked Decoy as the number one very best minion, because unlike other summons or mercenaries, the Decoy is never a "moron;" it always behaves exactly as expected.

In @jiansonz 's report of Guardian Martina, I was pleasantly surprised to read about a novel use of decoy: by casting the decoy into a corner, Martina was able to effectively disarm Talic's most harmful attack.

Valkyrie

Amazon14.jpg

I was also surprised at the wide variety of skill levels I saw for Valkyrie. However, I was not surprised to see its popularity: there are over 190 M/Gs that used this skill. I believe the most impressive write-up featuring Valkyrie is about a hardcore character that used no weapons at all, except for one particulary stubborn superunique monster. For everything else, Wolron's Wolazon relied almost exclusively on this summon and an Iron Wolf mercenary to achieve guardian.

Each time it is summoned, the Valkyrie gets a randomly generated set of either magic or rare gear (the exact gear depends on the Valkyrie's skill level). Many players try to summon at least with skill level 17, so that she can get her best possible rare weapon. Sometimes I get annoyed when the Valkyrie gets "hit causes monster to flee" on her gear; when that happens I usually just resummon her. Other times, I'm quite entertained when she gets "hydra shield" on her gear, and then I'm disappointed after she gets killed and the pretty hydras stop popping up around the monsters.

I remember feeling so very disappointed when I first tried wearing the Peace runeword for its chance-to-cast a level 15 Valkyrie minion. That Valkyrie would just disappear a second or two after she was summoned! It turns out that I also needed to invest at least one hard point in the Valkyrie skill, and only then would the "proc" Valkyrie stay around and fight.