Help analyzing data: Ninjistu HCI/DCI

[Lord Kotan]

I gathered some data on TC using Darts.

One character had 120 wrestling, the other had 120 ninjistu (and wrestling, if that matters).
===========================================
0 HCI vs 40 DCI:
5/10
7/10
5/10
5/10
5/10
4/10
5/10
2/10
4/10
5/10
(47/100)

0 HCI vs 0 DCI:
3/10
5/10
6/10
5/10
4/10
2/10
4/10
7/10
8/10
6/10
(50/100)
35 HCI vs 0 DCI:
4/10
6/10
5/10
10/10
7/10
7/10
5/10
4/10
6/10
8/10
(62/100)

===================================================

 

[spoonyd]

I gathered some data on TC using Darts.

One character had 120 wrestling, the other had 120 ninjistu (and wrestling, if that matters).
===========================================
0 HCI vs 40 DCI:
5/10
7/10
5/10
5/10
5/10
4/10
5/10
2/10
4/10
5/10
(47/100)

0 HCI vs 0 DCI:
3/10
5/10
6/10
5/10
4/10
2/10
4/10
7/10
8/10
6/10
(50/100)
35 HCI vs 0 DCI:
4/10
6/10
5/10
10/10
7/10
7/10
5/10
4/10
6/10
8/10
(62/100)

===================================================

I don't think you have enough trials yet but it would suggest that HCI does help you.

 

[Old Man of UO]

I don't think you have enough trials yet but it would suggest that HCI does help you.

Ya... the results fall into binary distribution (rather than a normal distribution) and require much higher sample size to get significant results. What you show looks likely there is a difference, but would show no significance in a statistical test.

 

[Harlequin]

These are just the hit chances right? Not successfully poisoned rates?

 

[Chaosy]

Ya... the results fall into binary distribution (rather than a normal distribution) and require much higher sample size to get significant results. What you show looks likely there is a difference, but would show no significance in a statistical test.

LOL!

The first time I've seen someone talk about actual statistics in UO...

 

[spoonyd]

Honestly I thought UO was all about mathematics and statistics. Knowing how those work out better equips you to design a strong template.

 

[Maplestone]

What is the hypothesis that is being tested? That the chance is always 50-50?

E(X)=0.5

so ... s^2=100(0.5)(1-0.5)=25 (so s=5, or 62 is about 2.5 standard deviations away from the expected value ... still somewhere around a 10-15% chance of happening, suggestive, but not good enough for proof ... needs a bigger sample size)

(wow I'm rusty, it's been way too long since I've done this sort of stuff)

 

[Chaosy]

I'd just rather plug it into Prism 5.0 and have it analyze for me.

 

[WarUltima]

What is the hypothesis that is being tested? That the chance is always 50-50?

E(X)=0.5

so ... s^2=100(0.5)(1-0.5)=25 (so s=5, or 62 is about 2.5 standard deviations away from the expected value ... still somewhere around a 10-15% chance of happening, suggestive, but not good enough for proof ... needs a bigger sample size)

(wow I'm rusty, it's been way too long since I've done this sort of stuff)

Omg you reminded me of my university days. Stop it. Electrical Engineering courses STILL haunt my dreams after 4 years. I am glad its over.

 

[Maplestone]

I'd just rather plug it into Prism 5.0 and have it analyze for me.

*looks down at slide rule*

um ... yeah.

*checks for grey hairs*

 

[Maplestone]

Omg you reminded me of my university days. Stop it. Electrical Engineering courses STILL haunt my dreams after 4 years. I am glad its over.

sorry