719 Lore Benefit Research

[crb]

http://www.virilneus.com/blog/2009/06/18/719-lore-benefit-research/

[Xaerve]

Able to copy/paste here? My filter at work blocks your site for some reason.

[Kuyuk]

wtf is that


http://www.virilneus.com/719graph.png

[Kuyuk]

if he wanted to make this more accurate, he'd do 0, 10, 20, 50 ranks, or more, etc

[Danical]

if he wanted to make this more accurate, he'd do 0, 10, 20, 50 ranks, or more, etc

There's supposed to be huge DI after 10 ranks or so I thought some GM said.

[Bhuryn]

I guess you could say 10 ranks definatly increased the min and max potential damage result, I wouldn't call it much of a net gain though. There's clearly no large shift in results in that dataset.

[crb]

if he wanted to make this more accurate, he'd do 0, 10, 20, 50 ranks, or more, etc
except this is DC, not major fire. There is "significant" diminishing returns after 10 ranks, and supposedly no returns (or near enough as makes no matter) after 20.



I guess you could say 10 ranks definatly increased the min and max potential damage result, I wouldn't call it much of a net gain though. There's clearly no large shift in results in that dataset.

Ya, I wouldn't call it worth it. Any recommendations for a sorcerer to train in elemental lore will hinge on the elemental lore review and the 400s. This training for DC is a post-cap indulgence.

You could also figure the TP costs of the lore vs. the TP cost for more spell ranks where the more spell ranks would get you a higher endroll and so more damage.

[Danical]

You might want to try comparing the crit ranks of the crits with similar endrolls.

Mark finished the fire crit table.

[Swami71]

My understanding of DC was mana controls helped damage and lores helped crit. Reading the official docs it does say damage for lores but I don't see any clear damage difference in your results.

[Celephais]

Yeah I don't think you have nearly enough data points to draw any relevant conclusions. That and as Danical said this may be a 4 dimensional problem you're representing in 3 dimensions (Endroll, Damage, Lore Ranks and crit rank).

[Deathravin]

My understanding of DC was mana controls helped damage and lores helped crit. Reading the official docs it does say damage for lores but I don't see any clear damage difference in your results.

That's because you tried to look at the data as he put it in a big graph.

If you look at the data itself, and draw conclusions from that, It's fairly obvious there's a 12-18% damage increase between 0 and 10 ranks among large groups of data points with similar endrolls. I Still don't think it's good enough for the cost by any means.

[Danical]

The problem is confounded by the fact that crit damage is location specific so a rank 9 head crit is nowhere near the same damage as a rank 9 chest/back/abs. Therefore, looking at straight damage increase isn't the right way to do it.

[Fallen]

The problem is confounded by the fact that crit damage is location specific so a rank 9 head crit is nowhere near the same damage as a rank 9 chest/back/abs. Therefore, looking at straight damage increase isn't the right way to do it.

Agreed. DC, aside from using it to put down stuff like Glacei, is a crit killer. I happen to believe the lores were well worth the cost in TPs as a post cap investment. I cast the spell probably 7 out of every 10 casts in the field. At even slight gains, that makes it worth it.

[crb]

My understanding of DC was mana controls helped damage and lores helped crit. Reading the official docs it does say damage for lores but I don't see any clear damage difference in your results.
No, mana controls are supposed to help mana returns.

Sorcerers do not possess a spell where mana controls modify damage. I wish we did, we've asked for it, but we've been Nilvened on that.



Yeah I don't think you have nearly enough data points to draw any relevant conclusions.
966 datapoints. I think that is statistically significant.

You can download the excel sheet and sort it by crit to put all like crits nearby, you'll get data like this:



0 207 70 Head explodes, splattering sizzling bits of flesh and bone everywhere.
0 205 60 Head explodes, splattering sizzling bits of flesh and bone everywhere.
0 208 60 Head explodes, splattering sizzling bits of flesh and bone everywhere.
0 130 50 Head explodes, splattering sizzling bits of flesh and bone everywhere.
0 187 50 Head explodes, splattering sizzling bits of flesh and bone everywhere.
0 213 70 Head explodes, splattering sizzling bits of flesh and bone everywhere.
10 220 55 Head explodes, splattering sizzling bits of flesh and bone everywhere.
10 185 55 Head explodes, splattering sizzling bits of flesh and bone everywhere.
10 121 50 Head explodes, splattering sizzling bits of flesh and bone everywhere.
10 196 50 Head explodes, splattering sizzling bits of flesh and bone everywhere.
10 141 50 Head explodes, splattering sizzling bits of flesh and bone everywhere.
10 206 50 Head explodes, splattering sizzling bits of flesh and bone everywhere.
10 199 70 Head explodes, splattering sizzling bits of flesh and bone everywhere.
10 215 55 Head explodes, splattering sizzling bits of flesh and bone everywhere.
10 231 50 Head explodes, splattering sizzling bits of flesh and bone everywhere.
10 188 60 Head explodes, splattering sizzling bits of flesh and bone everywhere.
10 231 50 Head explodes, splattering sizzling bits of flesh and bone everywhere.


Now that is interesting. Roll a 213, get this head crit, and 70 damage, with 0 lore ranks. Roll a 220, get this head crit, and 55 damage, with 10 lore ranks. That is a significant degree of randomization. Twice I rolled a 231 and only got 50 damage on that crit.

This is why I graphed it. It allows you to visualize the results and see that, 10 lore ranks did produce higher damage some times. But the damage is so randomized, it wasn't all of the time.

The anecdotal evidence from people like Evarin could just be a psychological effect. This is why in medical testing they give people placebos. He remembers his crits more once he starts looking for them to validate his lore training.

For the instant death crit:



0 184 70 Fire completely surrounds an Ithzir seer. Blood boils and heart stops.
0 198 70 Fire completely surrounds an Ithzir seer. Blood boils and heart stops.
0 164 70 Fire completely surrounds an Ithzir seer. Blood boils and heart stops.
0 197 70 Fire completely surrounds an Ithzir seer. Blood boils and heart stops.
0 216 75 Fire completely surrounds an Ithzir seer. Blood boils and heart stops.
0 221 70 Fire completely surrounds an Ithzir seer. Blood boils and heart stops.
0 210 70 Fire completely surrounds an Ithzir seer. Blood boils and heart stops.
0 250 70 Fire completely surrounds an Ithzir seer. Blood boils and heart stops.
0 177 70 Fire completely surrounds an Ithzir seer. Blood boils and heart stops.
0 186 70 Fire completely surrounds an Ithzir seer. Blood boils and heart stops.
0 205 70 Fire completely surrounds an Ithzir seer. Blood boils and heart stops.
0 207 70 Fire completely surrounds an Ithzir seer. Blood boils and heart stops.
0 194 70 Fire completely surrounds an Ithzir seer. Blood boils and heart stops.
0 213 70 Fire completely surrounds an Ithzir seer. Blood boils and heart stops.
0 203 70 Fire completely surrounds an Ithzir seer. Blood boils and heart stops.
0 172 70 Fire completely surrounds an Ithzir seer. Blood boils and heart stops.
0 202 70 Fire completely surrounds an Ithzir seer. Blood boils and heart stops.
0 151 70 Fire completely surrounds an Ithzir seer. Blood boils and heart stops.
10 183 70 Fire completely surrounds an Ithzir seer. Blood boils and heart stops.
10 224 70 Fire completely surrounds an Ithzir seer. Blood boils and heart stops.
10 216 70 Fire completely surrounds an Ithzir seer. Blood boils and heart stops.
10 142 70 Fire completely surrounds an Ithzir seer. Blood boils and heart stops.
10 201 70 Fire completely surrounds an Ithzir seer. Blood boils and heart stops.
10 178 70 Fire completely surrounds an Ithzir seer. Blood boils and heart stops.
10 182 70 Fire completely surrounds an Ithzir seer. Blood boils and heart stops.
10 197 70 Fire completely surrounds an Ithzir seer. Blood boils and heart stops.
10 223 70 Fire completely surrounds an Ithzir seer. Blood boils and heart stops.
10 184 70 Fire completely surrounds an Ithzir seer. Blood boils and heart stops.
10 153 70 Fire completely surrounds an Ithzir seer. Blood boils and heart stops.
10 167 70 Fire completely surrounds an Ithzir seer. Blood boils and heart stops.
10 173 70 Fire completely surrounds an Ithzir seer. Blood boils and heart stops.
10 172 70 Fire completely surrounds an Ithzir seer. Blood boils and heart stops.
10 181 70 Fire completely surrounds an Ithzir seer. Blood boils and heart stops.
10 241 70 Fire completely surrounds an Ithzir seer. Blood boils and heart stops.
10 162 70 Fire completely surrounds an Ithzir seer. Blood boils and heart stops.
10 225 70 Fire completely surrounds an Ithzir seer. Blood boils and heart stops.
10 202 70 Fire completely surrounds an Ithzir seer. Blood boils and heart stops.
10 203 70 Fire completely surrounds an Ithzir seer. Blood boils and heart stops.
10 159 70 Fire completely surrounds an Ithzir seer. Blood boils and heart stops.
10 141 70 Fire completely surrounds an Ithzir seer. Blood boils and heart stops.
10 158 70 Fire completely surrounds an Ithzir seer. Blood boils and heart stops.
10 192 70 Fire completely surrounds an Ithzir seer. Blood boils and heart stops.
10 152 70 Fire completely surrounds an Ithzir seer. Blood boils and heart stops.
10 182 70 Fire completely surrounds an Ithzir seer. Blood boils and heart stops.
10 198 70 Fire completely surrounds an Ithzir seer. Blood boils and heart stops.
10 179 70 Fire completely surrounds an Ithzir seer. Blood boils and heart stops.
10 206 70 Fire completely surrounds an Ithzir seer. Blood boils and heart stops.
10 184 70 Fire completely surrounds an Ithzir seer. Blood boils and heart stops.
10 176 70 Fire completely surrounds an Ithzir seer. Blood boils and heart stops.


10 ranks got it as low as 141, 151 for 0 ranks. That may indicate something. But 0 ranks got 75 damage once, 10 ranks always was capped at 70 it would seem.

On the back death crit:



0 150 70 Back burnt to the bone. Smoke curls up from what's left of an Ithzir seer.
0 200 70 Back burnt to the bone. Smoke curls up from what's left of an Ithzir seer.
0 197 70 Back burnt to the bone. Smoke curls up from what's left of an Ithzir seer.
0 226 70 Back burnt to the bone. Smoke curls up from what's left of an Ithzir seer.
0 179 70 Back burnt to the bone. Smoke curls up from what's left of an Ithzir seer.
0 222 70 Back burnt to the bone. Smoke curls up from what's left of an Ithzir seer.
0 188 70 Back burnt to the bone. Smoke curls up from what's left of an Ithzir seer.
0 157 70 Back burnt to the bone. Smoke curls up from what's left of an Ithzir seer.
0 219 70 Back burnt to the bone. Smoke curls up from what's left of an Ithzir seer.
0 179 70 Back burnt to the bone. Smoke curls up from what's left of an Ithzir seer.
0 190 70 Back burnt to the bone. Smoke curls up from what's left of an Ithzir seer.
0 210 70 Back burnt to the bone. Smoke curls up from what's left of an Ithzir seer.
0 218 70 Back burnt to the bone. Smoke curls up from what's left of an Ithzir seer.
0 196 70 Back burnt to the bone. Smoke curls up from what's left of an Ithzir seer.
0 188 70 Back burnt to the bone. Smoke curls up from what's left of an Ithzir seer.
0 218 70 Back burnt to the bone. Smoke curls up from what's left of an Ithzir seer.
0 215 70 Back burnt to the bone. Smoke curls up from what's left of an Ithzir seer.
0 207 70 Back burnt to the bone. Smoke curls up from what's left of an Ithzir seer.
10 213 70 Back burnt to the bone. Smoke curls up from what's left of an Ithzir seer.
10 194 70 Back burnt to the bone. Smoke curls up from what's left of an Ithzir seer.
10 191 70 Back burnt to the bone. Smoke curls up from what's left of an Ithzir seer.
10 201 70 Back burnt to the bone. Smoke curls up from what's left of an Ithzir seer.
10 200 70 Back burnt to the bone. Smoke curls up from what's left of an Ithzir seer.
10 187 70 Back burnt to the bone. Smoke curls up from what's left of an Ithzir seer.
10 187 70 Back burnt to the bone. Smoke curls up from what's left of an Ithzir seer.
10 206 70 Back burnt to the bone. Smoke curls up from what's left of an Ithzir seer.
10 181 70 Back burnt to the bone. Smoke curls up from what's left of an Ithzir seer.
10 167 70 Back burnt to the bone. Smoke curls up from what's left of an Ithzir seer.
10 156 70 Back burnt to the bone. Smoke curls up from what's left of an Ithzir seer.
10 230 80 Back burnt to the bone. Smoke curls up from what's left of an Ithzir seer.
10 197 70 Back burnt to the bone. Smoke curls up from what's left of an Ithzir seer.
10 215 70 Back burnt to the bone. Smoke curls up from what's left of an Ithzir seer.
10 237 85 Back burnt to the bone. Smoke curls up from what's left of an Ithzir seer.
10 142 70 Back burnt to the bone. Smoke curls up from what's left of an Ithzir seer.
10 202 70 Back burnt to the bone. Smoke curls up from what's left of an Ithzir seer.
10 199 70 Back burnt to the bone. Smoke curls up from what's left of an Ithzir seer.
10 190 70 Back burnt to the bone. Smoke curls up from what's left of an Ithzir seer.
10 230 70 Back burnt to the bone. Smoke curls up from what's left of an Ithzir seer.
10 162 70 Back burnt to the bone. Smoke curls up from what's left of an Ithzir seer.
10 160 70 Back burnt to the bone. Smoke curls up from what's left of an Ithzir seer.
10 173 70 Back burnt to the bone. Smoke curls up from what's left of an Ithzir seer.


0 lore ranks got it on a roll of 150 at the lowest, 10 lore ranks 142. 0 lore ranks was always capped at 70 damage, 10 lore ranks once got 80, once 85, but otherwise was mostly capped at 70 as well. The two times 10 lore ranks got more damage it had a higher end roll than I ever achieved with my 0 lore ranks tests. Which could also be why I'm not seeing a low under 150 for the 0 lore ranks either.

If someone went into the data and filled in a new column for crit rank and just copied in the crit rank for the crit, it'd be easy to do a new graph of endroll to crit rank. I'd be willing to bet it looks like the endroll to damage graph though, with mostly overlaps, and just a few instances where a benefit is seen.

[yesicj]

966 datapoints. I think that is statistically significant.

It could be, but it depends on what you're sampling (assuming you're using "statistically significant" as a descriptor of quantity, rather than as its used in analyzing statistics).

I don't have your actual spreadsheet in front of me right now, so I'm going off of what you just posted: 966 datapoints, but only 17 involving, for example, the exploding head death crit (the specific crit I chose is irrelevant, it was just the first listed in that post). 13 injury locations, each with 10 different crit levels produces 130 different possible results. Over 966 points, that's roughly 7 shots per location on average, 4 with lore and 3 without (380/966 ≈ .4). 7 data points is small enough in number that you could essentially be just as likely to over- or under-estimate the difference lore makes as you would be to getting it right.

Don't get me wrong, the data could be useful for some broad comparisons (and I know I sure as hell don't have the patience to put that kind of list together). But for the type of stuff it seems like you're interested in, you'd need quite a bit more of it (which, again, would take the kind of patience that I know I don't have).

[Fallen]

If I ever come back I will cough up some data. I was on my way to 10 ranks of lore in each element. It could be the placebo effect, as Viril mentioned, but I doubt it. As I worked towards 10 ranks in an element, the crits in that location got better. I had the same CS, the same targets.

[crb]

denil posted crit analysis on the officials

if someone has logs w/ 20 ranks we can analyze those. I have a script I wrote to do it. it is mostly automasted

[TheLastShamurai]

My Sorc has 15 in each element, I could go DC some stuff if you want.

[crb]

to limit extraneous data from uncontrolled variables I'd like to keep all data to hits on ithzir seers. So if you could DC a few hundred ithzir seers, ya, have at it.

[TheLastShamurai]

to limit extraneous data from uncontrolled variables I'd like to keep all data to hits on ithzir seers. So if you could DC a few hundred ithzir seers, ya, have at it.

Ah, no dice then. Dudes only like 72.

[crb]

Two more bits of data.

Avg Endroll::Crit Rank at 0 Lore Ranks



avg(endroll) critrank
149.1667 1
167.0000 2
152.9200 3
149.1282 4
154.6522 5
170.5882 6
171.0339 7
172.0000 8
193.9450 9


Avg Endroll::Crit Rank at 10 Lore Ranks



121.5000 1
126.0000 2
146.1176 3
152.6765 4
157.7544 5
163.0933 6
168.9870 7
172.8169 8
187.1186 9

[Fallen]

Is this what is being discussed on the sorcery boards, Viril? Any GM comments at all (on anything) of late? Any word on Strath? What about the E-lore review?

[crb]

Yes.
No.
No.
No.

You know that warm fuzzy feeling you sometimes get? Ya, not caused by Simu.

[Fallen]

Yes.
No.
No.
No.

You know that warm fuzzy feeling you sometimes get? Ya, not caused by Simu.

It is from that thing in my pants, isn't it?

[crb]

here is the graph denil made with my data after plugging in the value for crits, I've now updated the xls download on my blog post to contain the crit values.

http://www.virilneus.com/719results.jpg

[Celephais]

I wouldn't put much stock in the whole "With lore gets worse > 225" area... your data points start becoming very scarce around then, especially for no lore.

[Rolton-Sammich]

I wouldn't put much stock in the whole "With lore gets worse > 225" area... your data points start becoming very scarce around then, especially for no lore.

It's actually not a data point problem, it's that the lines are only estimated by a linear and a squared term (OLS regression). I doubt that the Real Game Formula curves quite like that; it's probably just a function of the crit rank cap. Once you get up to 220+, you're getting so many crit ranks of 9 without lore that you no longer see much of a benefit. Because the line was estimated to have only one curve in it, this is the curve that best fits the effect, but it does not take into account that the outcome can't get past a crit rank of 9. (For anyone statistically minded, this data really calls for a tobit regression analysis due to the right-censoring after the outcome = 8.)

But (ignoring the bit at the end where the lines actually cross) this graph does capture something of what you will experience as a benefit. When you hit 220+, you're most likely to be obliterating each body part the crits land on, with or without elemental lores.

[Rolton-Sammich]

The more I look at my previous graph on 719 lore benefits, the less content I am with the regression approach. Here's a table and graph based on a more raw approach to the data. I think this gives a more accurate impression of the lore effect on crit rank:


Endroll With Lore Without Lore Gain
110 4.7 4.3 0.4
130 5.7 5.1 0.6
150 7.1 6.2 0.9
170 7.2 6.5 0.7
190 7.9 7.2 0.6
210 8.0 7.9 0.2
230+ 8.5 8.2 0.4


http://i560.photobucket.com/albums/ss47/denilisshort/719again.jpg

My interpretation remains essentially the same as what I posted before. 10 lore ranks appear to give an effect equivalent to roughly a +20 warding margin for the aligned crit. If one was using 719 heavily, elemental lore training might be a more efficient training plan than other methods (such as overtraining sorcerer spells) of boosting the spell's effectiveness.

~Denil

[Fallen]

I think if you're capable of regularly warding your targets by margains as high as 200 with DC, lores aren't likely going to make much of a difference. Less than that, though, and you see the effects mentioned by Denil.

[TheWitch]

Oye.

All this math at which I fail.
I've read this thread a couple times, and my takeaway is don't bother with lore till post cap. That accurate?

Viril, I had data for you at 24 ranks of fire, but when I got pissed off and quit, I mass deleted pretty much everything. I'll dig around and see if that escaped my trashcan rage by some chance.

Evarin, you can read the officials as a guest, just FYI.

[Xaerve]

Can I dig around with you? :) I'll wear my construction worker hat you bought me!