This report was completed as part of a task for Ashlynn Sedai of the Yellow Ajah. I realized today that I have not yet published my findings for the White Tower as a whole, so I wanted to do so to raise awareness about this weave and raise awareness of some factors that might (or might not) influence healing. Please feel free to contact me individually with any questions or concerns.
In the Light,
Amarea Helsen
Tower Accepted
Brown Ajah Apprentice
___________________________________________________________________________________________________
INTRODUCTION
For this project, the assignment I received was straightforward and open-ended: to research the Cure Serious Wounds weave. Before discussing how I approached this task and my findings, I will first briefly describe its basic properties.
The Cure Serious Wounds weave (henceforth abbreviated as CSW for brevity) requires 1 flow of Earth, 4 flows of Water, and 2 flows of Spirit. Due to its relative complexity, wilders are unable to learn it, and it is only available to women trained at the White Tower and some rare other clans as an exception, such as the Wisdoms. The weave requires 10 SP and 15 pulses of time to form the flows. The healer can both initiate and complete the CSW weave while she is engaged in combat as long as she is not interrupted. She can initiate the weave on a subject who is also engaged in combat, but will not be able to complete the weave if the subject is still in combat when it finishes, as she cannot get close enough to apply the flows.
When considering CSW and potential research questions, an obvious target of study would be basic information on how much Healing the weave applies. However, rather than make assumptions that the weave works equally effectively in all situations, and to provide the most objective review of the weave, I expanded my view to include a list of all the factors that might possibly affect Healing, with the plan to empirically test whether or not they have an impact.
I believe the following factors merit further study to conclusively rule in or out whether they have a significant impact on amount of Healing:
(1) Percentage of weave training
(2) Healer's level
(3) Healer's clan rank
(4) Healer's Intelligence score
(5) Healer's Willpower score
(6) Subject's total hitpoints
Previous researchers have studied what factors affect other skills, such as landing rate and damage of backstabbing (e.g., https://wotmudarchives.org/forum/viewtopic.php?t=2442), which have later been replicated and confirmed by others (e.g., https://wot.writtenrealms.com/playground/stab). It is a logical assumption that many of the same factors could feasibly apply to channeling effects. As such, I hypothesized that one's Intelligence and Willpower would have no direct impact on Healing amount; that level likely has a significant impact on Healing up until level 30, but likely no further differences for anyone between level 30 and level 51; that clan rank might apply a modest bonus parallel to rank damage applied to damaging weaves; and that percentage of weave training obviously affects weave success rate but may also directly impact final amount Healed. Anecdotal evidence I have collected over time suggests that the channeler's skill dictates the amount Healed, not the subject's hitpoints, but I do not know if this has ever been formally studied.
A secondary goal of this research project, once completed, would be to ascertain a reasonable formula to determine the amount of Healing. This might be in the form of a simple XdY dice roll, such as with weapon damage, but there may be additive or multiplicative factors at play.
I cannot independently test each of these factors, because many of these factors are outside of my personal control, namely my rank, Intelligence, and Willpower. I can change my level, but dropping it significantly for the purpose of this study is inconvenient, so I have deferred that study for now. This report thus consists of two separate studies. In Study 1, my aim was to collecte baseline data about the weave under ideal circumstances (i.e., using myself as the Healer with 99% training in CSW). In Study 2, my aim was to test the effect of weave training percentage as an independent variable on amount Healed. The general methods for how I approached this was essentially identical across studies.
METHODS
In both Study 1 and Study 2, I was the sole Healer involved in data collection. This was important to keep all other individual difference factors (i.e., level, rank, stats) constant across studies. As much as possible, I also tried to minimize the number of subjects who participated, hopefully to minimize the potential impact of target factors on the outcome, such as their total hit points (although I will speak on this briefly in the conclusion). Without any specific rationale for determining target sample size, I arbitrarily set a goal of 100 samples for each study. I ultimately had 3 brave volunteers: Reyl (43 samples), Accepted Perenelle (113 samples), and Elder Frey (48 samples). Data were collected between 1/12/25 - 1/18/25.
In Study 1, I trained in the CSW weave to 99% with Sheriam Sedai. In Study 2, I chose to set my training at a level that was significantly lower, but not so low as to significantly impede the data collection process so as not to waste too much time of my subjects; I settled on 78% weave training. A future study could use an even lower number (e.g., 50%) to make the analysis more robust.
Data collection consisted of two phases: the damage phase, and the Heaing phase. In the damage phase, Reyl and Frey chose to seek out hostile enemies and allowed them to whittle their health down to Beaten or Critical injury status. Perenelle used a simpler approach and wove Ice Spikes upon herself repeatedly until she was Battered or lower. In the Healing phase, I wove CSW repeatedly, while the subject constantly checked their exact health amount and told me out loud their new exact hit points any time this changed, whether from recovery on the hour (tick), half-hour (half tick), or actual Healing. This allowed me to ensure that the Healing amount was measured independent on one's natural recovery. While the subject and I both rested, I then copied results into a spreadsheet for tracking purposes.
STUDY 1 RESULTS
In total, 104 samples were collected at 99% weave training. The following sample statistics were calculated:
Mean = 31.88
Standard Deviation = 8.35
Minimum = 13
Maximum = 46
Median = 33.5
Mode = 40
An immediate intriguing finding was the actual distribution of the data. In a typical dice roll consisting of an XdY format, we would expect a normal distribution; for example, in a 10d6 roll, we would expect the highest frequency of results near the statistical average of the roll (here, 35) with the frequencies of results decreasing in size as they move to the low and high extremes (10 and 60, respectively). For an example, see here: https://anydice.com/.
Instead, the Heaing amount of 40 was disproportionately represented (n = 22), such that the median and mode of the sample were both above the mean. The frequency of results is represented visually below in a simple histogram plot with "X" representing a result in the sample and the number of X's indicating the number of results at that amount:
13: XX
14: --
15: XXX
16: X
17: XX
18: --
19: X
20: XX
21: XXX
22: X
23: XXX
24: XXXX
25: XXXXX
26: XXXXX
27: XXX
28; X
29: XXX
30: XXX
31: XXX
32: XXX
33: XXXX
34: XXX
35: XXXX
36: XXXXX
37: XXXXX
38: XXX
39: XXX
40: XXXXXXXXXXXXXXXXXXXXXX
41: X
42: X
43: XX
44: --
45: X
46: XX
This distribution of results suggests that the actual formula underlying the CSW weave is not a simple XdY format. The closest approximation of 10d5 would generally result in a similar mean, minimum, and maximum as the obtained sample, but it does not explain why the distribution was so significantly skewed to the positive end.
STUDY 2 RESULTS
In total, 100 samples were collected at 78% weave training. The following sample statistics were calculated:
Mean = 32.11
Standard Deviation = 8.37
Minimum = 10
Maximum = 50
Median = 33
Mode = 40
Compared to Study 1, the distribution statistics were largely identical. Of note, the mode was still 40 and the sample was still significantly skewed with an unusually higher number of results of 40 HP healed (n = 13). However, the skew was slightly less pronounced, as shown in the following histogram:
10: X
11: --
12: --
13: X
14: --
15: --
16: --
17: X
18: XX
19: --
20: XXX
21: XX
22: XXX
23: XXXXX
24: XXXXXXX
25: XXXX
26: X
27: XXX
28; XXXX
29: XX
30: --
31: XX
32: XXXXX
33: XXXXXXX
34: XXX
35: XXXX
36: XXX
37: XXXXXX
38: XX
39: XXXX
40: XXXXXXXXXXXXX
41: XXXXX
42: X
43: --
44: X
45: X
46: --
47: X
48: X
49: X
50: X
Importantly, based on a 2-sample t-test with independent means (https://www.graphpad.com/quickcalcs/ttest2/), the amount of Healing provided at 99% weave training (M = 31.88, SD = 8.35) was not significantly different than the amount of Heaing provided at 78% weave training (M = 32.11, SD = 8.37), t(202) = 0.196, p = 0.845.
Although it was not a primary focus of this study, I incidentally collected a nearly equal amount of data from Reyl and Frey with ~400 maximum hitpoints (n = 91) and Perenelle with 280 maximum hitpoints (n = 113). A 2-sample t-test similarly revealed that the amount of Healing for the non-channeler subjects (M = 32.55, SD = 8.45) was not significantly different from the channeler subject (M = 31.55, SD = 8.26), t(202) = 0.851, p = 0.396.
DISCUSSION
The first and most important finding of this preliminary study was that the amount of Healing provided by Cure Serious Wounds is about 32 HP; anecdotally, this is almost identical to the average damage of a single Flame Strike weave. This means that if a channeler has a Willpower of 18, maximum energy in saidar, and does not fail any weaves, she can provide an average of 448 HP of Healing before she expends her energy. This is a sizable amount, and its potential impact on the flow of battle cannot be understated, especially because a channeler's SPs recover at a fairly quick rate, especially if she has higher Intelligence and/or Willpower.
Secondly, this study suggests that percentage of weave training in CSW has no direct impact on the amount of Healing it provides. Collecting more data at an even lower weave training level (e.g., 50%), while tedious, would help to confirm this finding.
Although I did not explicitly plan to statistically evaluate the effect of subjects' maximum hitpoints on amount Healed by CSW, my sample allowed me to respond to this question and reasonably conclude that subjects' hitpoints have no meaningful impact on Healing. As suspected, this confirms that the weave does not merely recover a certain proportion of the subjects' health, and instead provides a raw amount that is independent of the subject. This is anecdotally supported by offering Healing to weaker individuals who have a significantly lower level or hitpoint maximum.
The slight differences in the distribution curve for Study 1 and Study 2 samples was interesting, but not significantly different enough to actually affect Healing in a measurable way. In both studies, especially Study 1, it was surprising and intriguing to find that 40 HP was the modal response. If this pattern is replicated by future studies, it would strongly suggest that a simple XdY formula is not sufficient to model the amount of Healing, and instead there are likely multiple factors in place that influence the amount of Healing. Because of its skew, this most likely includes at least one multiplicative factor, as additive factors alone (e.g., adding a flat amount based on rank) would result in a normal distribution curve but centered around a different mean.
Of the factors remaining that I have identified as plausible targets but have not yet tested, the Healer's level would be the most likely culprit; if level was at least partially correlated with the amount Healed, that could explain why the distribution is centered around a modal result significantly higher than its average, while still allowing for a broad range of potential outcomes.
One potential limitation is that some confounding variable related to the subjects could have had an impact on the results. However, I do not believe this is likely. All three subjects were of equal levels of worldly experience (level 51) at the time of data collection. Even if the subjects' mental statistics were possibly a factor, the statistical test comparing Frey + Reyl's data to Perenelle's data indicated no significant differences, and given that Perenelle's Intelligence and Willpower (19 and 19) are assumedly much higher than the others', this seems unlikely.
Overall, I am eager to continue this study and will share my results at a later time once I complete further research into the potential impact of other factors on amount Healed. For now, I thank you for the challenge to delve further into the Cure Serious Wounds weave, and I hope that you find these data informative and useful.