Tinyheroes.com Forums

SAUVA

Snipe v Fire with bows and crossbows 07/16/2018 01:56 PM CDT

TL,DR: Don't snipe with a crossbow.

Over the past few days I've gathered some combat data. Hits were recorded and put on a 0-23 scale (miss - apocalyptic strike). 775 bow, 777 crossbow, 100 agil, 100 str, lelori forester's crossbow, azurel battle bow, capped basi bolt|arrow, juvie wyverns, with STW, with CS, full aim.

BOW	First Arrow	Second Arrow	Combined
average hit	9.02±1.7 n=139	8.61±3.8 n=139	8.84±2.9 n=262	average Mindstates gained	2.2±0.6 n=139
via snipe	8.94±1.6 n=77	8.84±3.3 n=77	8.89±2.6 n=154	via snipe	2.3±0.6 n=85
via fire	9.07±1.8 n=54	8.44±4.2 n=54	8.76±3.3 n=108	via fire	2.1±0.7 n=54
t=	0.45	0.58	0.34	t=	0.38
p=	0.65	0.56	0.73	p=	0.70

So I think if I interpret that correctly, there's no Statistical difference between the first and second arrow either when using snipe over fire under the above listed conditions right? So why do I get this data with crossbow:

CROSSBOW	Hit
average hit	7.65±2.2 n=225	average Mindstates gained	1.5±0.6 n=225
via snipe	7.08±2.5 n=103	via snipe	1.4±0.7 n=103
via fire	8.15±1.9 n=101	via fire	1.5±0.6 n=101
t=	3.428	t=	0.78
p=	0.00074	p=	0.44

IF I am interpreting that p value correctly that would indicate that there -is- a -Statistical- difference between the use of snipe and fire. And fire does more damage. Does that seem right? Why would snipe make no difference for bows but make crossbows worse?

~Hunter Hanryu
>Everything Rangers have is just a lame version of something cool.~Morkim

DIMINISHEDANGEL

Email Author

Re: Snipe v Fire with bows and crossbows 07/17/2018 12:48 AM CDT

What's puzzling to me isn't the pronounced difference between snipe and fire for xbows, but the relative lack of difference for bows. I was always under the impression snipe's multiplier (or whatever you want to call it) would be significantly weaker than FIRE as compensation for the tactical advantage of remaining in hiding.

Perhaps it's a percentage reduction, and the additional power allotted to crossbow shots for longer load times means a higher loss of result? Sort of like how a 20% debuff means more at 1750 ranks than 1000 ranks.

SAUVA

Email Author

Re: Snipe v Fire with bows and crossbows 07/17/2018 07:58 AM CDT

>I was always under the impression snipe's multiplier (or whatever you want to call it) would be significantly weaker than FIRE as compensation for the tactical advantage of remaining in hiding.

A successful snipe (I want to say it's one of the hardest checks in game) is supposed to impart an accuracy bonus. As I understand it, accuracy can help damage, but only to a point in Combat3.0. I figure I'm past that point on juvies since I miss < 1% of the shots I take at full aim. All things equal, snipe was not intended to impart a penalty on success to "compensate" for staying hidden. The super hard check and the "abuse it and loose it forever" threat are the "balance" to the no visible name "bonus". I use the term bonus quite loosely.

All of the above is "as I understand it based on epedia and memory". If any of that is wrong, I'd love to see the GM reference posts.

~Hunter Hanryu
>Everything Rangers have is just a lame version of something cool.~Morkim

DRSCOTT

Email Author

Re: Snipe v Fire with bows and crossbows 07/19/2018 11:05 AM CDT

Very interesting to see actual statistics in a DR discussion. That said, I don't think you should look at second hits - it just adds so many additional variables. Ideally, each target would be solidly balanced. I'd be interested to see how initial snap shots compare using fire or snipe. Obviously, that's a lot more work to compile. Maybe the accuracy boost from snipe would be evident there?

ANARCHIE

Email Author

Re: Snipe v Fire with bows and crossbows 07/19/2018 01:11 PM CDT

Wouldnt the variance in creature stats cause a drift in the data as well? Would it be worth adding the compared appraisal values stat vs stat during the data collection?

SAUVA

Email Author

Re: Snipe v Fire with bows and crossbows 07/20/2018 12:19 PM CDT

I'm sure there are better ways to gather the data, but I'm just using my normal training script as this started with the question "what's the best way to train ranged". I wouldn't fire flat-footed if I have the balance boost from CS, I wouldn't snap shot b/c of the huge increase in miss rate, and I wouldn't single load b/c the learning from a single shot is bad compared to dual load (no stat data for that yet, I'm sorry). Below is the data as of my last hunt.

Ignoring 2nd arrow and combined arrow data (if you want to), the first arrow from a composite bow is all the same (fire v. snipe v. poach). Interestingly fire v. snipe on a short bow has hit a 10% significance threshold. I like 5% much better as an arbitrary significance alpha so I won't make any definitive statements, but it is interesting. Also look at that number for the mind states/shot on snipe v. fire from a battle bow. Strange right?

For the crossbow data, clearly snipe is worse than fire or poach in damage output. IF that is intended, I think it sucks. On the bright side, the extra 3 seconds of loading the arbalest v. forester gives an approximate 1.6 multiplier to exp per shot. That's well worth it in my opinion!

Shout out to Nasrid for hooking me up with an arbalest for these tests!

BOW	First Arrow	Second Arrow	Combined		average Mindstates gained
average hit (all)	7.98±2.3 n=418	8.14±3 n=418	8.06±2.7 n=836		2±0.6 n=418

fire (battle)	9.12±1.7 n=77	8.78±3.8 n=77	8.95±3 n=154		2.1±0.6 n=77
snipe (battle)	8.96±1.6 n=102	9.06±3.1 n=102	9.01±2.5 n=204		2.3±0.6 n=113
poach (battle)	8.67±2.5 n=76	9.39±2.6 n=76	9.03±2.6 n=152		2.2±0.6 n=76
p(fire v snipe, battle)	0.539	0.602	0.834		0.083
p(fire v poach, battle)	0.201	0.249	0.789		0.336

fire (nisha)	6.74±1.8 n=42	6.83±2 n=42	6.79±1.9 n=84		1.7±0.5 n=42
snipe (nisha)	6.13±1.8 n=71	6.59±1.7 n=71	6.36±1.8 n=142		1.7±0.5 n=74
poach (nisha)	6.47±2 n=36	6.44±2.2 n=36	6.46±2.1 n=72		1.7±0.5 n=36
p(fire v snipe, nisha)	0.09103	0.50856	0.09360		0.87846
p(fire v poach, nisha)	0.54709	0.42173	0.31041		0.94287
				fire v fire	0.000120642
				snipe v snipe	3.5751E-10
				poach v poach	4.24909E-06

CROSSBOW	Average HIT		average Mindstates gained
average hit	7.66±2.2 n=261		1.5±0.6 n=225

fire (forester)	8.02±2.2 n=115		1.5±0.6 n=115
snipe (forester)	7.15±2.4 n=112		1.5±0.6 n=133
poach (forester)	7.55±2.2 n=125		1.4±0.6 n=125
p(fire v snipe, forester)	0.00500016		0.79
p(fire v poach, forester)	0.101562391		0.62

fire (arbalest)	10.56±3.3 n=41		2.3±0.7 n=41
snipe (arbalest)	9.09±3.4 n=47		2.2±0.8 n=50
poach (arbalest)	10.69±2.3 n=26		2.5±0.5 n=26
p(fire v snipe, arbalest)	0.043701084		0.639204501
p(fire v poach, arbalest)	0.850718732		0.348332481
		fire v fire	1.08691E-09
		snipe v snipe	1.27147E-08
		poach v poach	3.13081E-15

~Hunter Hanryu
>Everything Rangers have is just a lame version of something cool.~Morkim

DRSCOTT

Email Author

Re: Snipe v Fire with bows and crossbows 07/21/2018 12:03 PM CDT

Ah, I didn't read carefully - I didn't realize this was dual load.

I'm curious how a sniper's crossbow would compare to the one you used. It does seem odd that snipe would have lower damage - I would expect equal damage.

The 1.6 exp modifier matches my more informal testing - I've been using an arbalest for awhile for that reason. Time to lock is important for me since I'm training all weapons, although the long load time does kinda drive me crazy.

DragonRealms · General Discussions