Hypothesis: AGI and DEX matter a _lot_ in open ambush

[Goat]

I don't have super-high quality data on this yet - I'm uncertain about the exact scenario in past tests, etc.

But when I was minimally enhanced (I think it was just +6 dex stat, yielding +38 DEX bonus and +30 AGI bonus), I could aim head with a lance about 50% of the time and with a naginata about 70% of the time.
With 45 dex bonus/41 agi bonus, I recently succeeded 72% of the time with a lance. (82 successes, 32 missed area, 8 ebps)
With 52 dex bonus/48 agi bonus, I recently succeeded 94% of the time with a naginata. (47 successes, 3 missed area, 3 ebps)

So some extra AGIDEX appears to have made a 20% difference in both cases (18 more AGIDEX in the lance case and 32 in the nagi case). I know of no other relevant differences between the two trials.

In all cases, level 100 sorcerer, 151 CM (enhanced), 0 ambush, against a combination of executioners, combatants, and radicals. I leg them with 708 before attacking, and (almost always) let them swing to un-turtle first.

Of course, the "low AGIDEX" trials were like a year ago or something, and I might have been less careful about my recordkeeping then... may have kept swinging after giving a rank 3 head wound, etc.

I hope to get around to a more proper "low AGIDEX" trial with the naginata soon. I may not bother with the lance, because removing those enhancives would leave me with an 8 second lance ambush, which is just annoying. Still, I think my research has shown it plausible that statements like "stacking AGI and DEX enhancives will make a worthwhile difference in aiming" and "AGIDEX races can get away with a little less ambush" are true.

[gilchristr]

Myth appears to be plausible

Seriously though thanks for posting this.

[gilchristr]

Goat, maybe one or both of them matter a lot when you are relying only on CM and have no ambush training...

Since you have no ambush

[Goat]

Yeah, I think I understand the latest working idea, that CM and Ambush impacts are independent and each soft-capped, and that leaves a lot of possibilities for agi and dex play in. I think I'll pick up 24 ranks ambush with the latest fixskill, and maybe I'll manage to do agidex boost/no agidex boost trials with 12 and 24 ranks ambush while I'm at it.

[Goat]

Looks like AGIDEX does matter, at least with decent amount of CM and no ambush - about 20% higher success rate for an extra 26 AGIDEX (+/- random chance due to modestly-sized samples).


Combatants, Executioners, Radicals
Always aim for the head
Swinging a naginata
Target always prone (I 708 first)
Target always in stance off except for some attacks on radicals (haven't the patience)
I try again on the same mob if the first shot misses the head, but stop if the first shot causes a rank 3 head wound without death.
151 ranks of CM
0 ranks of ambush
Some of the higher-AGIDEX trials involved rank 1 hand/arm/neck scars, but nothing more.




Stats|Success|Missed Location|Failed to find opening|E/B/P|location success on hits
52 DEX bonus, 48 AGI bonus | 47 | 3 | 0 | 3 | 94%
40 DEX bonus, 34 AGI bonus | 37 | 13 | 0 | 8 | 74%

N.B. My previous findings with lower DEX/AGI showed 74 hits, 40 missed locations, and 8 EBPs, (65% location success) but that was under less controlled conditions.

I also think it's interesting the E/B/P rate was higher with lower AGIDEX. I have a sense that destroyers E/P me less than they used to, and I wonder if it's because I've upped AGIDEX over time.

[gilchristr]

thanks for posting this!

Are the sizes of the three different types of creatures you used for data the same?

[gilchristr]

Were you wearing any spells that give phantom combat maneuvers, that you werent wearing in the first test? Like mobility.

We don't know that the combat maneuver ranks from mobility work for anything besides CMANs, i.e. we dont know that they help with aiming, but just curious.

[Goat]

thanks for posting this!

Are the sizes of the three different types of creatures you used for data the same?

Yes.

The triton combatant is medium in size and about six feet high in his current state.
The triton combatant is medium in size and about one foot high in his current prone state.

The triton radical is medium in size and about six feet high in her current state.
The triton radical is medium in size and about one foot high in her current prone state.

The triton executioner is medium in size and about six feet high in her current state.
The triton executioner is medium in size and about one foot high in her current prone state.

They are different levels, though (96/98/100), which I imagine matters. I'm relying on that idea that I kill roughly the same proportion of each monster on each hunt, though clearly that won't always be the case, especially for the smaller samples (~50 hits). I do keep the raw data separate per creature, but the obviously the samples are smaller if I break it out that way.


Were you wearing any spells that give phantom combat maneuvers, that you werent wearing in the first test? Like mobility.

We don't know that the combat maneuver ranks from mobility work for anything besides CMANs, i.e. we dont know that they help with aiming, but just curious.

I was wearing mobility in this latest round of tests and possibly not in earlier ones, however, per Coase, "This only applies to CML calculations" (https://gswiki.play.net/mediawiki/index.php/Mobility_(618)_(saved_posts))

[Goat]

10 ranks of ambush good for 13% higher success rate (vs. 0 ranks) in the following trials.
14 more ranks of ambush good for 6% higher success rate after that

Both rows: 151 ranks CM, 45 dex bonus, 41 agi bonus, swinging a lance.

Ranks of ambush|Success|Missed Location|Failed to find opening|E/B/P|location success on hits
0 | 82 | 32 | 0 | 8 | 72%*
10 | 69 | 12 | 0 | 2 | 85%
24 | 91 | 9 | 0 | 9 | 91%**

(third row posted in a follow-up edit on 4/10/2016)

* Data gathered with rank 1 scars on left hand and right arm, and a rank 1 neck scar for the last 20 hits.
* Data gathered with rank 1 limb/neck wounds/scars at various points

[Rhovan]

Have you considered encumbrance in your tests?

[gilchristr]

Ohh good all the same size!

[Goat]

Have you considered encumbrance in your tests?

Not explicitly, but I'd guess than 90%+ of my swings are with no encumbrance, and <=2% with more than light encumbrance. It would be interesting to see if big encumbrance makes a difference, but I'm not sure I've got the patience for all that RT.

[Ardwen]

Your hypothesis makes sense to me, I have zero ambush ranks and somewhere north of 60 combined bonus between the two and miss less then 20 percent of my open aimed attacks, guesstimate there, as I don't track the hits, but many hunts I don't miss at all when I aim.

[Goat]

Your hypothesis makes sense to me, I have zero ambush ranks and somewhere north of 60 combined bonus between the two and miss less then 20 percent of my open aimed attacks, guesstimate there, as I don't track the hits, but many hunts I don't miss at all when I aim.

Weapon size seems to have a lot to do with it, too. I don't have an super-controlled comparison between my naginata and my pike, but the naginata appears to be a fair bit easier. If we're talking bastard axe, that might be in between the 6-sec base naginata and 9-sec base lance in terms of aiming difficulty.

[gilchristr]

I am at 68 combined bonus, which I thought was good until I saw your 100+.

Ardwen's crappy "somewhere north of 60 combined" made me feel better though.

[Goat]

I am at 68 combined bonus, which I thought was good until I saw your 100+.

Ardwen's crappy "somewhere north of 60 combined" made me feel better though.

If it makes you feel any better, I spend like 25m a year on enhancive recharging :-)

[gilchristr]

Yea thanks Goat.

Speaking of enhancives, is your shop link broken? In any case, I saw a +6 blunt weapon pinworn in your shop (was it 2.2m?)... I'll take it, deliver to Witbier in the landing please.

[Wrathbringer]

People like you guys are why we can't have nice things. You'll just buttrape them into oblivion in an effort to maximize instead of using them casually like everyone else. When I hear about people like you guys, I'm forced to admit that fusion/enhancives is indeed broken.

[Goat]

Speaking of enhancives, is your shop link broken? In any case, I saw a +6 blunt weapon pinworn in your shop (was it 2.2m?)... I'll take it, deliver to Witbier in the landing please.

It was, thanks. Playershops changes all the URLs every once in a while.

[Goat]

My quality of data is a lot lower with destroyers because the scatter is not nearly as cooperative with researchers as Nelemar's third floor is. In particular, I had a variety of injuries during these trials. I probably won't bother to try improving the quality of data or number of trials, but I thought some might still be interested in what I've got.

Both rows: Headshot attempts on destroyers (with them almost always in offensive and standing), 151 ranks CM, 45 dex bonus, 41 agi bonus, swinging a lance.

Ranks of ambush|Success|Missed Location|Failed to find opening|E/B/P|location success on hits
0 | 47 | 31 | 0 | 13 | 60%
10 | 25 | 5 | 0 | 9 | 83%


I may never go past 10 ranks ambush, given that it gets me pretty close to what I want, and it costs me 43 PTP per rank. It'll depend on how things go during next Duskruin run, or when sanctum of scales finally opens.

[Goat]

Bump for update to post http://forum.gsplayers.com/showthread.php?100482-Hypothesis-AGI-and-DEX-matter-a-_lot_-in-open-ambush&p=1842283#post1842283

tl;dr - 0 ranks ambush, 72%. 10 ranks, 85%. 24 ranks, 91%.

Oh, also, naginata eye shots.
Both rows: combatants/executioners/radicals who are always prone and mostly in offensive, 151 ranks CM, 52 dex bonus, 48 agi bonus.

Ranks of ambush|Success|Missed Location|Failed to find opening|E/B/P|location success on hits
0 | 16 | 13 | 0 | 2 | 55%
24 | 30 | 10 | 0 | 3 | 75%

Smaller sample size, but still interesting.

[gilchristr]

There is some interesting information on aiming in a saved post on gswiki re: predator's eye.

The GM expressly states that skills AND stats are taken into account in the second stage of an aiming check. If stats are taken into account, it would sure make sense that dex/agi could be the ones.

Another interesting point is the cap of 95% in the second stage of the formula. There are lots of posts mentioning that a certain amount of ambush ranks didnt help them, it could be related.

Another interesting point is the 5% fumble rate in the first stage and the 95% cap in the second stage. Since both checks need to succeed, the odds of them a success on the aim attempt is 95% x 95% = 90%. So I wouldn't expect that anyone could get greater than 90% chance of success for head/eye/neck shots on a standing target (it may be possible to get better odds on those locations on a prone target if the bonus for being prone is greater than the body location penalty, which we dont know)

[gilchristr]

Actually it sounds like 90% is the effective max for legging a standing target, too, by applying what is stated in that GM post and some math.

[gilchristr]

I was wrong the GM post does not say there are two rolls. I think there is one roll that checks for fumble, failure on a 1-5.

Then, there is an amount added to that roll to see if 100 is reached (success). The amount added could be skill bonus and stats added to the roll according to as follows:
CM/4 + ambush/4 + stats, then cap at 95 before going to the modifier stage.

Then there are some modifiers added:
- penalty for hard locations (Unclear where penalties for large weapons go... they probably exacerbate the penalty for hard locations)
- bonus based on sitting/ prone


The reason why goat saw help from stats is because he isnt trained enough in skills (250/4) + (50/4) = 75. His stats still helped him close the gap to 95. I think someone who had a total 360 combined CM and ambush skill would not see any difference from stats. Judging by his posts I think it could be (AGI bonus + Dex bonus)/4 ... increases would help up to a combined bonus of 80 even with his 10 ranks of ambush (if he drops the ambush those stats could further help to even larger values)

[Goat]

Thanks. I had no idea about that saved post. For reference: https://gswiki.play.net/Predator%27s_Eye_(saved_post)

Where did "CM/4 + ambush/4 + stats" come from? Didn't you see some evidence that there were individual caps on the contribution from CM and ambush?

[gilchristr]

Sorry, I meant to clue that it was a guess by adding the word maybe, could have been more explicit. Remember open aiming is an OLD system. I doubt its that complicated. Skill/4 would be typical of an old system, as far as skill contribution. Several of the research posts would also seem to roughly fit (I have no idea on the stats, just informed guessing on the skills contribution)