Let's talk about Luck and ARLS Formula

[Sotc]

There might be something to the arena token drop/luck score correlation. My speed rage has +45 luck and he seems to get a noticeably more arena token drops than my other glads. As for gauntlet end prizes/conquest end prizes, not so much.

[Adoede]

Thanks for that bit, SOTC. I'll edit the original post.

Thanks to new info from Prinny, Pit Lord, Kreegan and Apoc, we have to fine-tune the formula to incorporate different "weightier stats". (Now I'm committed and have a brain-worm pushing me to figure this out)

So...can you guys chime in on some details for me?

1) What is the highest bonus you've ever seen for each stat on a piece of gear? Maybe this will help figure the different "tiers" of the stats. So far, I have:
Str +25, Int +25, Agi +27, Sta +22, Size +11, Pres +30, Chi +30, Luck +25

2) It seems the simple version of the formula seems "close" because the average gold cost per stat point is somewhere around 1.75. I'm trying to figure how the different stat tier's pricing works out so it averages around there and gives us a "weighted" version of the formula.

Some versions I tested that seemed to work well were these:
1) Str, Int, Agi, Sta & Luck are 1.75 gp each. Size is 3 gp each. Pres and Chi are 1 gp each. This averages out to 1.72 cost per stat.
2) OR based on the tiers I mentioned above, Str, Int, Luck at 1.75 gp each, Size is 2.5 gp each, Pres and Chi are 1.25 gp each, Sta is 2 gp each, and Agi is 1.5 gp each. This also averages out to 1.72 cost per stat. (and seems to match my glads perceived luck even better than the original formula)

If you'd like your glad to be included in my calculations, send me their stats and gold cost - and tell me if you think they're lucky or not. In fact, send me your luckiest glads so we can test the formula. So far, I only have 12 glads I'm using as my test cases for my spreadsheet.

[Alba Kebab]

Seen size +12 on a piece of leg gear just yesterday.

[Adoede]

I ran the weighted formula (#2 version) with the 12 HOL glads and this is a comparison of the weighted vs. non-weighted formula results:

Non-weighted simple formula | Weighted formula
Glad 01: Non-weighted 80 | Weighted 92
Glad 02: Non-weighted 74 | Weighted 88
Glad 03: Non-weighted 84 | Weighted 79
Glad 04: Non-weighted 83 | Weighted 91
Glad 05: Non-weighted 64 | Weighted 74
Glad 06: Non-weighted 64 | Weighted 78
Glad 07: Non-weighted 81 | Weighted 87
Glad 08: Non-weighted 63 | Weighted 65
Glad 09: Non-weighted 69 | Weighted 77
Glad 10: Non-weighted 77 | Weighted 84
Glad 11: Non-weighted 90 | Weighted 92
Glad 12: Non-weighted 92 | Weighted 92 (this is Cynaidh's new glad...lucky either way!)

As you can see, many had a significant difference. It's even more noticeable with non-HOL glads

Name _____St In Ag Sta Si Pr Ch Luck Cost Non-weighted/Weighted
Non-HOL 01 61 69 65 67 75 73 73 53.7 923 _____44/54
Non-HOL 02 66 74 64 67 60 66 83 55.3 908 _____39/55
Non-HOL 03 67 72 67 60 63 70 63 60.3 893 _____48/60
Non-HOL 04 74 60 78 67 73 56 63 73.9 946 _____70/74
Non-HOL 05 60 63 71 70 76 72 63 62.0 929 _____56/62
Non-HOL 06 51 71 75 58 74 65 52 66.7 890 _____63/67
Non-HOL 07 60 74 59 77 65 86 66 54.6 925 _____42/55
Non-HOL 08 65 57 68 84 51 57 54 74.4 880 _____67/74
Non-HOL 09 65 55 64 77 55 70 65 57.6 867 _____44/58
Non-HOL 10 69 68 75 84 62 67 77 65.6 970 _____52/66
Non-HOL 11 68 72 75 71 57 65 61 72.9 927 _____61/73
Non-HOL 12 56 75 73 61 67 55 63 71.6 901 _____65/72
Non-HOL 13 60 65 66 69 62 66 60 62.1 877 _____53/62
Non-HOL 14 60 53 70 67 64 59 51 65.0 848 _____61/65
Non-HOL 15 61 60 55 61 58 66 54 56.4 810 _____48/56
Non-HOL 16 64 73 68 66 48 77 85 50.7 885 _____25/51
Non-HOL 17 61 66 48 67 69 75 72 42.6 859 _____33/43
Non-HOL 18 73 54 56 75 57 57 45 70.7 850 _____69/71
Non-HOL 19 72 55 74 79 58 73 72 60.3 923 _____44/60
Non-HOL 20 66 63 56 65 85 65 79 46.1 913 _____43/46
Non-HOL 21 59 55 48 52 67 67 60 44.1 779 _____37/44

Some of the non-weighted results are ridiculous (#16 has 25 estimated luck non-weighted)
So the "weighted" formula definitely looks more accurate...

Waiting on some more gear data before I post the revised "weighted" formula for you all to enjoy.

[Pit Lord]

Estimating the exact "weight" of each stats probably won't be as simple as looking to the highest stat presented by the items. The highest stat is just that - highest possible stat in given item - if the item has the highest possible stat but the other stats are lacking it means the stat took a large part of the item budget.

We must first assume the items have "battle rank" budged. What is battle rank? Its the same as the rank (block rank, crit rank etc.). Its the minimum quantity of stat possible (although 1hp would be the minimum this will make the things too complicated).

We can than estimate the weight of each stat in "battle rank" points. As example:
1 rank = 1 battle rank
1 chi = 6-7 battle rank
1 int = 11-12 battle rank
1 deff = 3-4 battle rank
1 size = 20 battle rank
and so on...
This is just example not actual values.

The problem with this approach is the items have different budgets....
There certainly are different quality items of the same color - "common" legendary, "rare" legendary...

What I mean if we can estimate the "battle rank" of each stat including luck by looking the items then we will be able to estimate very close the gold price of each primary stat presented (or hidden) in the slaves. This way we can estimate the luck very close.

This all is valid if the luck is connected with the other stats and not random roll...

Then there is the question how each stat or combination of stats influences the performance of the gladiator.

The best thing about this game is you can play it 5 minutes per day and you can also easily sink hours in exploring it.

[Apoca1ypse]

I bet Nates reaction to this thread is as follows:

[Adoede]

Hahaha - nice one Apoc

[Adoede]

What I mean if we can estimate the "battle rank" of each stat including luck by looking the items then we will be able to estimate very close the gold price of each primary stat presented (or hidden) in the slaves. This way we can estimate the luck very close.

This all is valid if the luck is connected with the other stats and not random roll...

I get what you're saying Pit Lord and I agree with the concept. But the big question is how do we determine the relative weight of each stat in comparison with the other stats? The reason I wanted to see the highest bonus values for each stat is that I think it will give us a clue to the "weightiness" (battle rank) of each stat with the other stats.

So far I have:
Size: +12
Sta: +22
Int: +25
Luck: +25
Str: +27
Agi: +27
Pres: +30
Chi: +30

That shows us 5 tiers of stats...and Size is 2.5x more valuable than Pres/Chi on face value. At least this is a start to trying to figure what the gold cost per stat might be.

As for whether there is a link between Gold cost and luck, my post about my nearly identical glads is supposed to argue that there is truly a seeming correlation.

[Adoede]

As for Ahhnold, ignore him from any studies. TBH I dont actually know how much he cost, I just know that it was 1170+. I spoke with Nate about him and was told that his luck isnt bad. As far as I can tell, his luck is average. I've run a hell of a lot of 2H ragers (7?), and he doesnt feel unlucky compared to the others. In regards to not being a top 10 glad with his stats, that is very much a function of his style, compared to anything else. I just really dont want to run him with a pair of axes.

Hey Apoc, I'd like to pick on Ahhnold because he's the only glad that we know at least a little about his luck (isn't bad...). All my weighted models put his luck pretty high. That's the case for three different formulas I've tried. Now did Nate say his luck isn't bad (as in it's pretty average)...or that his luck isn't bad (as in hey, pretty good!).

And is there anyone who has a glad and knows their cost and luck?

[oedi]

I got a couple of old glads with very similar cost and total of stats that might be interesting to check out.

76 74 67 74 73 68 60 gold cost 981 total stats 492
77 83 73 67 63 65 63 gold cost 980 total stats 491