Data Analysis
Quick Links:
- Objective
- Legend
- Summary
- Procedure
- Results for Rohane
- Results for Mipsy
- Results for Talinia
- Results for Velm
Objective:
|
Legend:
ET - elapsed time
TT - time interval for character to make a move
mt - multiple target
ch - Celestial hammer
dd - direct damage
gd - group direct damage
gh - group heal
gs - group shield
h - heal
mes - mesmerization
The default skill used for Rohane: melee attack
The default skill used for Mipsy: direct damage
The default skill used for Talinia: melee attack
When innate haste increase, time taken per turn decrease.
Rohane's 0 haste time for Melee Attack is 5.6 seconds.
Mipsy's 0 haste time for Direct Damage is 6.5 seconds.
Mipsy's maxed haste time for Direct Damage at level 48 is 3.4 seconds.
Talinia's 0 haste time for Melee Attack is 5.6 seconds.
Velm's 0 haste time for Celestial hammer is 6.9 seconds.
Velm's 0 haste time for Group Heal is approximately 10.5 seconds.
Velm's 0 haste time for Mesmerization is 9.5 seconds.
In general, Velm is the slowest. Rohane and Talinia are the fastest.
Mipsy's direct damage is approximately twice as fast as group direct on 4 enemies. Group haste is the slower of the three spells.
Time/ turn is influenced by these factors: innate haste, overall level, type of skill used and for some skills like group direct and group heal, the number of monsters/characters targeted. (i.e when more team members need to be group healed, the slower Velm is for the next turn)
External factors affecting time/turn excluded from this report: mesmerization, slowing strike, stunning strike executed by enemies and the use of potions .(i.e dataset is removed when our team members are mesmerized, slowed, stunned, healed by potions or throwing potions)
Other skills like slowing strike, stunning strike, battle taunt and combat focus of our characters are not analysed in this report.
Just simply record down of the elapsed time for each character's turn for different
levels. Subtracted to get the time interval between each turn. The tables below
are only a small portion of the the data collected. As there are too much data,
it isn't feasible to throw them all in here. The tables were chosen as samples
because they reflect the results most accurately.
Results
Test Levels: 1 to 5
Test Venue: At the start of Chapter 1. Monsters do not possess any abilities.
Remarks: Heal by potions
not taken into account.
Results:
Table 1a Rohane's 0 haste timing is 5.6 seconds.
Recommendation: It is efficient to spend at least 3 haste near the initial appearance of the character. The first 3 innate haste skill will haste your character significantly. The remainder of the innate haste levels can be built up gradually as the increment is lesser.
Hypothesis 1: Character's timing is not affected by overall level.
Table 1b Average timing when haste is kept constant while overall level increase
Table 1c Average timing when haste is kept constant while overall level increase
Hypothesis is wrong. Time taken by character to make their move is affected by an increase in character's overall level. However, the increment is too small, as such we should not rely on this fact. Increasing innate haste is still vital.
Record down the elapsed time for each turn for 3 levels of haste.
> Get the time interval between each turn
Record down the elapsed time for Mips's turn using only direct damage.
Record down the elapsed time for Mips's turn using only group direct damage.
> Get the time interval between each turn for comparison (to see which skill is
faster)
Remarks: At 0 haste, Mipsy takes an average of 6.5sec till her next turn when direct damage is used. (See Table 2a below)
Table 2a. Mipsy's 0 haste timing for direct damage at Level 9
At maximum innate casting haste, her timing was 3.5 sec at level 44 and 3.4 sec at Level 45 for direct damage (see Table 2b and 2c), once again proving again that the overall level would affect TT to a minor extent, as we have found in earlier in Rohane's case.
Direct Damage (DD) vs Group direct damage (GD)
Test Level: At level 44, maxed innate haste
Test Venue: Happy Fun, Non-haunted house
As I did not spend on group direct, I waited till a wand with +4 GD was picked up to test it out. Earlier, I have confirmed that TT is not influenced by the level of a particular skill. Therefore, I could use a weapon boost to test out TTs for group direct without worry. The conclusion was that direct damage is roughly twice as fast as group damage when innate haste is maxed (see Table 2b below). Like Velm's GH, Mipsy's GD timings fluctuate with the number of enemies and the damage doled out. When Mipsy is faced with less enemies, TT for GD is shorter. For 1 enemy shots, GD's timings are similar to DD's.
Based on the info we obtained, lets compare DD and GD to see which is better. First, we assume that both Obliterate and Astral Maelstrom are equally good to facilitate analysis. Assume that none of the attacks are resisted by enemies: For a 4 enemy situation, it is a 100 versus 256 (64*4) damage for the first turn. But as the fight continues, you will find that DD has an advantage.
The higher attack power of Obliterate would mean that the monster is killed off much quicker than that of Astral Maelstrom. So as the fight proceeds, you have fewer monsters up against you. However, if you use group direct, you will need a few turns before you kill off a monster, leaving them more turns to attack your party, higher chances of sustaining more damage on your part. This is good for fighting bosses. In a fight with the Four Faeries, you will want to finish them off one by one rather than leave them standing while you slowly shave off their health using GD. This leaves them more chances of killing your team.
Recommendations: Max Direct Damage rather than Group Damage.
Table 2b Comparison of Mipsy's Group Direct and Direct Damage at level 44
Table 2c Comparison of Mipsy's Group Direct and Direct Damage at level 45
Group Haste
Test Level : At level 47 & 48, some characters had their innate haste maxed.
Test Venue: before Hubrid Nox Fortress where the monsters have no special abilities
A comparison of the TT of everyone before and after group haste was casted indicated that they were faster after haste. (Duh! obviously...) The casting of group haste would slow Mipsy for the next round as she uses more energy than when she is attacking but she shows visible improvement in attack speed after group haste (See Table 2d). Going along this line of logic, we would know that using haste potions, which doesn't require any use of Mipsy's haste spells, would be faster. So, perhaps, if you like, haste potions can be used instead if you want to spend Mipsy's skill points on other skills. But I'll still would spend some skill points on Mipsy's group haste because 1) haste potions are individually used meaning each character would have to take a turn hasting themselves, 2)Mipsy's DD may not be effective all the time, especially when facing bosses like Terask. I would rather let her do some group haste than to idle away when the throwing of slow potions are not required.
However, the 30% group haste casted wasn't clearly reflected in the results. Dividing the average of the TT after group haste over the TT before group haste did not show a 30% increase in speed. When it came to Mipsy's turn (before the group haste was casted) the ET for the others were recorded. After group haste was casted, there isn't a change in ET. It was the next turn, after all the characters have made their move that the group haste takes effect. Thats a food for thought...
Table 2d
You might have notice that sometimes bosses, like Terask, cast group haste even though they were group hasted just 2 rounds back. I often wondered about the motive of doing so considering their spell should still be "fresh". Out of curiosity, I tried that out. The results indicated in Table 2e were disappointing. There wasn't any visible improvement when group haste was casted for the 2nd or 3rd time. Perhaps it was because 30% increase in speed for a shorter TT in less significant (example: 30% haste of 5 sec is 3.5 sec but 30% haste for 3.5 sec is 2.45sec). This suggests that you should only renew your group haste only when it is weakened but you shouldn't dismiss the possibility that when high-level haste potions like the 65% ones are used, the effect of the 2nd hasting would be more significant.
Test level: 19
Test Venue: Chapter 2, Outside Mountainside Inn
Data doesn't include turns that involve healing or being stunned.
Melee Attack vs Multiple Target (MT)
The TT examined for multiple target is not the time taken for the MT to reappear (i.e MT2-MT1) but the time taken for the next turn to appear when MT is used.
Table 3a 0 haste timings at level 19
In Tables 3b to 3d, where the timing for MT is compared against the melee attack for the particular level, not much difference is observed, in some cases, MT is faster. In others, MT is slower. As such, I could't say conclusively which skill is faster and could only assume that they are equally fast.
Table 3b
Table 3c
Table 3d
Table 3e
By observation, the time taken for Velm to make a move is significantly longer than the rest of his team. In an unsorted table for 0 haste, the timing varies by a fair bit suggesting that time per turn (TT) is affected by other factors besides haste. Based on the assumption that TT is affected by skill casted, the results were sorted according to skill used, as shown in Table 4a.
Table 4a A sorted table of Velm's timings
In Table 4a, there seem to be less variation, which implies that the character's time between turns is dependent on the type of skill casted. 0 haste timings for ch=6.9 seconds, gh= 10.5 seconds, mes = 9.6 seconds. Using Celestial Hammer gives the fastest TT, followed by Heal and Group Shield, Mesmerization then Group Heal. However, there are still some unexplained variations in timing when other records of different haste levels are examined (see Table 4b and 4c). As expected, when Velm's casting haste level is increased, the time interval between moves shortened. In Table 4b and 4c, time between turns (TT) for the use of Celestial Hammer (ch) are stable whereas group heal (gh) and mesmerization show significantly large fluctuations. This puzzling development stumped me for a while but it revealed the possibility that other unknown factors were involved. It was not until later when I was examining the records for Mipsy's group haste when the answer became clear.
Table 4b
Table 4c
When Velm's data was extracted from each fight record, at level 48 with innate casting haste of 11+4 (see Table 4d below for a sample of the records) where the location of testing is in Haunted Woods after the 4 faeries and before Hubrid Nox's Fort where:
You'll observe that TT is shorter when less team members need to be group healed. Thats it? No, the problem arrives when it comes to comparing the TT where 2 members were healed. The varying TT's seem to indicate that the group heal isn't as clear cut as it should be.
Table 4d
I asked myself these two question:
- Is TT affected by the sum of the amount healed?
- Is TT affected by the individual amount healed?
The answer is NO! In the last section of Table 4d, TT for amt healed of
(14+21) is 6.7seconds vs TT for amt healed of (71+18) is 6.4 seconds. While
there isn't much
variation in TT but there's a huge difference for amt healed.
Further areas of interest for analysis would be on throwing of damage potions. However, as potions are seldom used except for fighting bosses with many abilities, data collected are removed because the data wouldn't be as accurate when typical bosses can stun, slow or mesmerize. By personal observation, potion throwing seem to be faster than melee attacks
My unsolved queries for throwing of potions:
Does throwing potions affect the time interval between turns (i.e faster than using skills)?
How haste potions affect the time interval between turns:
1)If faster, by how much?
2)If they are thrown repeatedly, would the characters become faster and faster & if so, is there a limit to it?
created: 20th June 2004. last modified: 2nd July 2004