Menu
What's New
By:
Filters

Merry mayhem scoreboard

falkao said:
Using the Mann-Whitney-U-Test we can test if the median difference is significant and the p-value is .107. Since the p-value is greater than .05, we can't conclude from that, that there is a difference in medians yet.
Click to expand...

I knew it :)
40 would be too little :silly2:

Now more important things come to mind looking at that data:

Solo shooting:
Code: 
1.32	6.97	84.55	6.86	1.32	4.01	4.24	4.44	6.40	6.48

2.07	1.36	4.25	2.96	5.66	7.54	4.36	4.61	6.23	7.02

1.30	5.26	4.56	3.33	4.32	3.03	3.51	4.06	2.18	4.56

3.29	4.81	2.09	3.95	2.20	4.12	5.40	2.10	5.85	6.33
Total loot: 248.90 PED
Loot average: 6.22 PED
Total loot except best and worst result: 163.05
Loot average for the interextreme values: 4.29

With help:
Code: 
2.04	3.17	6.30	1.92	2.15	1.84	3.05	3.19	6.58	1.99

1.95	6.96	3.01	7.45	5.66	7.21	6.09	5.43	4.37	1.99

2.19	2.11	2.36	6.05	6.18	2.13	2.17	2.18	4.35	1.17

2.13	2.86	1.92	2.24	1.20	2.17	4.55	6.02	7.25	3.29

Total loot: 146.87 PED
Loot average: 3.67 PED
Total loot except best and worst result: 138.19
Loot average for the interextreme values: 3.63


And falkao please teach them how to get a 8 PED average on Hogglo they sure need some help because they are way off no matter what they seem to try.
falkao said:
Atm I'll do get a mean of 8.18 PED per Hogg young as loot. My model predicts that I should get something between 8.4 and 8.6 on the long run. Tax is .04 and hence the 8.18 do fit quite well.
Click to expand...

(only 1 value has above 8 PED so either your predictions of loot on Hogglo are wrong
or loot obtained in globals should be considered as one of the most important sources of it
and therefore comparing values of such low ammounts of data will most certainly result in high margins of error)
 
Last edited:
MG Mighty said:
And falkao please teach them how to get a 8 PED average on Hogglo they sure need some help because they are way off no matter what they seem to try.
Click to expand...

Till the last month a hogg dropped in mean normal loot of 3-3.5 ped and globals with a mean per kill of 5 PED. Not sure if there have been some changes over the year, but I'll a get a mean of 4 PED on normal loot atm, like Fluske.

So on a Hogg y, without a global, you'll get only about 30-40% back and you need globals to get a return rate of at least 80%. This is the intention of the mob as a typically event mob.
 
very impressive work falkao,and thx also mighty...it sure does look like average loot is lower indeed with pretagging,

but we will do another 30 hogs solo and with preshooting for the greater good of science ;)

ps;im also interested now how loot compares to solo/pretag when we team in a normal way(dam stackshare)
for normal teamhunting,how many kills would be sufficient to get meaningful data ?
 
fluske said:
ps;im also interested now how loot compares to solo/pretag when we team in a normal way(dam stackshare)
for normal teamhunting,how many kills would be sufficient to get meaningful data ?
Click to expand...

hehe, the investigators effect ;)

The needed sample size depends on the assumed difference and variance of loot. For the variance in terms of standard deviation I would assume 1.9 PED as observed in the 80 kills. For the difference there are several possibilities.

If we assume that the difference is similar to pretagging, and further assume that the looter does the same amount of dmg, then a similar sample size is needed per group. Since we already have solo observations we might incorporate them as well. Hence only the new team group is needed. There is however the possibility of a bias, since hunting time was different.

However, I would expect that team loot is more or less the same as solo loot. Therefore we have the problem to show equivalence. This is much harder as we must assume a relevant equivalence range. If for instance, 1 PED is regarded as equivalent, then we need again 60-70 kills per group. If the equivalence range is smaller, say 10 PEC, then n=5667 kills per group are needed. For 50 PEC it would be n=227 per group.
 
falkao said:
Fig. 2: Histogram of loot with respect to hunting type
[br]Click to enlarge[/br]
Click to expand...

sry when we're going off topic from the original thread intention.

I'm still thinking on this down shift and why it is noticeable only below 5 PED. If pretagging lowers loot, then I would have expected that all loot is lowered and not only a range. Would make more sense for team loot. When part of loot is lowered, then there would be space for globals. Hence when in team you could get a global more often.

Already curious on the team loot data.
 
ok,did another run with same setup as before...i solo first hog,then next hog my wife pretags it with exactly 12 hits with apis/dante..theh i solo..etc..

numbers solo

31.61 4.88 2.34 3.28 5.35 4.97 5.47 7.77 4.67 5.73

125 3.17 4.81 1.94 2.89 2.13 3.32 2.98 2.30 4.64


4.22 1.38 5.79 7.68 6.11 7.38 6.93 1.95 2.08 3.21
 
and numbers with pretagging


3.32 7.42 2.34 2.16 2.03 1.99 6.19 6.52 2.86 3.17


1.88 4.36 54 5.58 3.25 4.48 3 3.97 6.13 7.27



5.73 4.39 6.16 3.41 2.04 1.14 5.82 3.04 4.52

263
 
next we did an normal teamhunt on about 60 hog youngs,these were all in a row,i set teamloot to looter takes all and did the looting myself since i was looter in previous tests


7.25 3.05 6.27 2.22 4.51 3.85 6.25 3.35 6.50 2.02


2.76 1.25 4.58 2.03 4.40 2.92 1.27 2.48 3.51 102


3.07 3.97 4 3.22 6.22 2.94 3.10 4.21 1.99 4.04


5.48 5.78 2.17 7.22 2.24 5.43 3.35 2.05 3.98 5.62


7.50 4.32 6.79 4.59 4.18 5.68 5.52 4.19 4.33 3.18


2.84 3.23 2.20 2.13 2.86 31.26 4.78 5.63 3.95 4.28


3.11 7.21 3.01 2.83
 
fluske said:
ok,did another run with same setup as before...i solo first hog,then next hog my wife pretags it with exactly 12 hits with apis/dante..theh i solo..etc..
Click to expand...


Now all get's very interesting.

Here the boxplots for overall loot and loot below 5 PED (would get similar results for below 6 PED but not 7 PED).

Fig. 1: Overall Loot
[br]Click to enlarge[/br]


Fig. 2: Loot below 5 PED
[br]Click to enlarge[/br]


We do have the same trend as before. Loot below 5 or 6 PED is shifted downwards in the pretagged group, whereas solo and team loot are comparable.

Using Kruskal-Wallis and Mann-Whitney-U-test I'll get the following p-Values. The signifcant p-values are marked.

Tab 1: p-Values
fluske_p_values.jpg


Solo and Team Loot below 5 and below 6 PED are significantly higher than pre tagged and there seems to be no differences between solo and team.


So now we can ask, why part of loot below 5 and 6 PED is diminished when pre tagging?


add:

Fig. 3: Survival function of loot
[br]Click to enlarge[/br]

Especially between 2 and 5 PED loot is quite lower in the pretagged group with respect to the others.


Fig 3: Survival function with respect to pretagged yes/no
[br]Click to enlarge[/br]

It seems as classes are the same but weights are different when pretagging.
 
Last edited:
A bit of wild speculation:

It might be that the base loot is somehow set based on the amount of HP the mob has at the time the looter starts shooting at it.

What if the looter is the tagger and the support only does some extra damage after it has been tagged.

I think that variation should also be tested to make the research more complete.

Cheers
Siam
 
rhogenbe said:
A bit of wild speculation:

It might be that the base loot is somehow set based on the amount of HP the mob has at the time the looter starts shooting at it.

What if the looter is the tagger and the support only does some extra damage after it has been tagged.

I think that variation should also be tested to make the research more complete.

Cheers
Siam
Click to expand...

sure there is now space for the wildest speculations ;).


We know that loot considers hp dmg done to the mob. From this we can conclude that the system stores overall hp dmg done to the mob. Furthermore, we know from team hunting, that also dmg done by an ava is stored.

There are now several possibilities to incorporate this dmg into loot. My assumption till now was, that according to hp dmg done loot is calculated and subdivided afterwards to looters. Since when pretagging, the taggers are not considered as looters, loot might be lower.

The problem is that this lower loot can only be observed till 5 PED. So it could be, that loot classes do behave differently. But who knows.
 
rhogenbe said:
A bit of wild speculation:

It might be that the base loot is somehow set based on the amount of HP the mob has at the time the looter starts shooting at it.

What if the looter is the tagger and the support only does some extra damage after it has been tagged.

I think that variation should also be tested to make the research more complete.

Cheers
Siam
Click to expand...

the term pretagging might have been confusing..in fact..protocol was that i (main shooter and looter) tagged mob with bravo and when it came into 50 meter range both me and my wife started blasting away simultaneously,

my wife stopped after excactly 12 hits,so last 2-3 shots were for me to finish it
 
those boxplots really look interesting falkao,and it confirms what ive always suspected..namely that teamloot is somehow lootpenalized

its very obvious in the pretagginggroup versus solo,
but also slightly in the solo versus team

now i understand why we had such an massive loss at mayhem since we did solid 36 hours of pretagging,and i reckon every team that pretagged must have suffered losses

pretagging in events and killstealing makes ma money(since loot is disproportional to decay/ammo) so i doubt they wil make changes to the mayhem format next year :)
 
fluske said:
those boxplots really look interesting falkao,and it confirms what ive always suspected..namely that teamloot is somehow lootpenalized

its very obvious in the pretagginggroup versus solo,
but also slightly in the solo versus team

now i understand why we had such an massive loss at mayhem since we did solid 36 hours of pretagging,and i reckon every team that pretagged must have suffered losses

pretagging in events and killstealing makes ma money(since loot is disproportional to decay/ammo) so i doubt they wil make changes to the mayhem format next year :)
Click to expand...

Yep, about 16k ped losses. Most of it was in the markup of the Isis BL1450 (L)'s and the Bul Tac70 (L)'s i used and the fact that i paid for decay and ammo of my support team (total approx 6k ped).

Although the pretagging reduces loot i am not sure that all of the reduction goes to MA. I would not be surprised that a major part went into the HoF of Neo.

Cheers
Siam
 
rhogenbe said:
Yep, about 16k ped losses. Most of it was in the markup of the Isis BL1450 (L)'s and the Bul Tac70 (L)'s i used and the fact that i paid for decay and ammo of my support team (total approx 6k ped).

Although the pretagging reduces loot i am not sure that all of the reduction goes to MA. I would not be surprised that a major part went into the HoF of Neo.

Cheers
Siam
Click to expand...

I don't think so because he didn't profit either (with markup).
 
fluske said:
those boxplots really look interesting falkao,and it confirms what ive always suspected..namely that teamloot is somehow lootpenalized

its very obvious in the pretagginggroup versus solo,
but also slightly in the solo versus team
...
Click to expand...


Although the boxplot shows a slightly lower median for team vs solo, from a statical point of view we can only conclude that there is a difference between solo vs pretagg and team vs pretagg.

Please consider also that hp dmg done is higher when solo with respect to team due to the higher kill speed. This might lower loot in team.

There is however one possible effect. For loot below 5 PED I do get a mean of
3.25 PED for solo with s = 1.17. Team has a mean of 3.23 with s = .9. So the standard deviation s seams to be lower in team and this is also confirmed by the Levene's test (p = .03). Hence loot in team is more uniform which might again be a result of a more homogeneous dmg done.

Values for PreTag are a mean of 2.7 and s of .95 and hence quite comparable to team (Levene's p = .68). So also here I do find a more homogeneous loot.


For me the most mysterious part is the reduction of loot below 5 PED in the pre group. Nevertheless, as we know now, thanks to your data, that this effect exists pretagging strategies have to be revised. It could also be that this effect is not intentional and MA might modify loot calc, but as you already mention, it would not make much sense for them.
 
You must log in or register to reply here.
Back
Top