Has anyone ever wondered if daytime/nighttime conditions affect the accuracy of sight-based mobs? I've noticed subtle differences in accuracy from day to night while soloing lesser colibri using a parser. My evasion seemed to cap around 88% during night-time, and then fall down to around 77% during the day-time as a level 72 Thief with no evasion bonus gear. Can anyone confirm if daytime/nighttime conditions have an effect on accuracy of sight-based mobs? It would be interesting to see if someone else comes up with similar results.

I suspect that DEX and AGI play a double role in hit rate calculation similar to how STR and VIT play a double role in damage calculation. With damage calculation, the attacker's STR and the opponent's VIT determine the range of possible damage values that can come out. You could use a Gaussian curve to represent the distribution of damage values that can result. Then, attack and defense come into play to determine where within the determined range the damage turns out the most. My understanding is that higher Attack skews the curve left (i.e. lower damage values become less frequent, higher damage values become more frequent), and higher Defense skews the curve right (higher damage values become less frequent).

It is probably true that 2 DEX = 1 accuracy, but it might also have its own direct effect on hit rate as well. I imagine that the attacker's DEX and opponent's AGI determine a range of potential hit rates, then Accuracy and Evasion are checked to determine where within that range the actual hit rate lies. In effect, this would imply that when an opponent's AGI severely outweighs an attacker's DEX, regardless of how much Accuracy+ the attacker has, he or she may never achieve a good hit rate. If you add 10 or 20 accuracy and you don't see hardly any difference in hit rate, you may need more DEX.

I haven't used careful observation and analysis to determine if this truely the case, but one way you might check this is by seeing if hit rates for mobs that just barely reach the threshold of "low evasion" (or "high evasion") differ, or if the hit rate at this threshold is always the same.

I'm skeptical of the claim that "low evasion" is always +20 accuracy. I believe that it has to do with a ratio between accuracy and evasion. I seem to remember that when exping a new job from 1-10 a very few points of acc make the difference.

Accuracy seems to scale in a more more linear way than attack; adding the same amount of accuracy seems to increase your hit rate by the same amount regardless of your initial hit rate. I notice this with Madrigal; singing Madrigal seems to increase the Paladin's hit rate by the same amount as the Thief's, even though their initial hit rates differ by at least 20%.

From my experience with the log parser, "low evasion" translates to at least 75% hit rate, and hit rate is capped at 95%. I haven't tested "high evasion" in the same way, but I think it means 50% or less. --Valyana 09:07, 23 Jan 2006 (PST)

I've just been doing a bunch of accuracy testing that started with testing Madrigal effect, and I verified that the gap between high and low evasion is 40 points wide, at least for a BST75 with a club on a DC Date Eruca. I assume that the earlier testing was not done with a D weapon, so it doesn't seem to vary with the user's skill. And 40 points seems a rather nice, round number to be a percentage. Perhaps it's a set number that rises every few levels? --Valyana 09:59, 2 June 2006 (PDT)

A few months ago I did a bunch of testing with Accuracy, /check results, and hit rates. One thing I found interesting is that for /check purposes, 2 DEX = 1 accuracy, 1 combat skill = 1 accuracy, and of course "+1 accuracy" = 1 accuracy. It works just like Attack. The 20 above/below rule worked like magic as I tested /check results using different weapon skill levels, different amounts of DEX, and different amounts of "+ accuracy."

Another interesting fact is that food that raises Accuracy by a certain percentage, adds the percentage after calculating TOTAL accuracy. For example, if you use Squid Sushi (+15% accuracy), and then equip a Scorpion Harness and a Life Belt (for Accuracy +20), you will actually gain 23 accuracy. Using this method, I discovered that a player's total accuracy is calculated as:

Total Accuracy = 8 + (Combat Skill level) + integer(DEX/2) + (any bonuses).

I did quite a bit of hit rate testing against Tainted Flesh mobs as NIN72 in Labyrinth of Onzozo using different weapons. I didn't do quite enough to make a lot of solid conclusions, but here are some of my hypotheses:

- Adding accuracy in the form of "+ Accuracy" (through equipment, food, job abilities, etc.) suffers from diminishing returns.

- Adding accuracy in the form of Combat Skill points does not seem to suffer from diminishing returns.

- I believe that the attacker's DEX and the opponent's AGI determine a maximum and minimum hit rate range.

(By the way, I am interested to hear what this claim that 1 skill = 0.92 accuracy at later levels is all about. Does that mean that, for example, a difference in 12 skill points results in a difference of only 11 accuracy? Or is this observation based on an observation of the effects of combat skill vs accuracy on hit rates?)--Cryconius 12:53, 8 June 2006 (PDT)

I believe that the 0.92 number came out of testing done by Apple Pie on ffxionline forums, but I haven't been able to find the relevant thread.

In my recent testing, I saw that my Beastly Earring (+5 Axe) only gave me +4 accuracy (probably 4.something). I can't confirm the exact ratio, but I can confirm that it's less than 1:1. --Valyana 05:58, 9 June 2006 (PDT)

I'm looking for the data and process that contibuted to the accuracy formula currently posted. Thanks to the mention of Apple Pie and the FFXIOnline forums, I was able to find this thread, containing this post. Apple Pie referenced the formula (which is not quite the same as the one posted on the wiki) from a JP forum which seems to be focused on gathering data for stats. I did a little looking around on the site, but I couldn't find the 0.92 modifier.

This is the best that I could decipher...

when you weapon's skill is <= 200:

 weapon's skill + [DEX / 2] + accuracy boosts + 10

when your weapon's skill is > 200:

 200 + [(weapon's skill - 200) x 0.9] + [DEX / 2] + accuracy boosts + 10

The equations within brackets are to be turned into integers within the equation, rounding down. E.g., if your Dexterity is 55, the part of the equation that reads "[DEX / 2]" is [55 / 2] which equals 27.5, and rounding down equals 27.

I don't know if the JP forum is the original site for this formula, nor if the formula has been refined since it was originally posted. Thus, I'm hesitant to modify the wiki with the equations I just referenced. Discuss :IEscape 02:47, 11 October 2006 (EDT)

Here is the latest documentation from Accuracy testing from Studio Gobli. It seems they put it at 0.9, not 0.92 for every combat skill above 200--VZX 12:18, 11 October 2006 (EDT)

My mistake for not including that link from Apple Pie's post within my previous 'talk'. Three reservations I have about the formula posted on this section of this JP forum:

1. Is this the work of Studio Gobli?
2. Where is the data and process used to support this formula?
3. This formula was posted on December 19, 2003. It's not exactly what I would consider the "latest documentation." With all of the major updates that have happened within the past thirty-four months, has this formula remained true?

--IEscape 17:25, 11 October 2006 (EDT)

1. Although I think the people are from different board(I checked if there's similar message on certain day... there isn't), the people from Studio Gobli BBS like to refer about the documentation on that website and discuss it on the board.
2. I believe the number shown on that page is the data
3. People who did the test are quite sensitive to the change. It used to be Studio Gobli who first notice there was no critical modifier on Arching Arrow and Heavy Shot until the recent patch changed it. So I believe, if there was change, it should be noted on the thread (which I've been monitoring occasionally in recent months) or it will be put on Studio Gobli main page.--VZX 22:37, 11 October 2006 (EDT)

I parsed against a DC mob for about 1000 swings with a weapon my job has 0 skill in with 0 accuracy gear on, and got very close to 5% hit rate, so I'm putting that as the minimum hit rate.--Gerion 20:50, 16 April 2007 (EDT)

Were you using the latest version of DvsParse? Were you hitting for 0 dmg? It seems you were using only 1 mob. --JKL 05:08, 17 April 2007 (EDT)

DvsParse has a bug that makes it not count some hits, so using it might not produce accurate results.

I parsed this again on an EM Steelshell in Boyahda Tree for a total of 1006 swings as a LV75 NIN/THF with a staff (nin/thf doesn't have staff skill), no food, DEX69+20 and +13 accuracy, which shouldn't matter at all. I stopped when the mob was at about ~50% HP and the resultant accuracy was of 21.372%. The margin of error was around ~2.5%.

I think 20% minimum accuracy for players makes sense since monsters also have the minimum hit rate of 20% and the maximum of 95%. --JKL 07:36, 21 April 2007 (EDT)