Sentinel Spreadsheet data validation (running tally)

[Underfyre]

I meant specifically on the damage procs like Molten Embrace, Offensive Adaptation and Assault Radial. Those are still using (1 + R x (11 x A + 540) / 30,000)

[Bopper]

1 hour ago, underfyre said:

I meant specifically on the damage procs like Molten Embrace, Offensive Adaptation and Assault Radial. Those are still using (1 + R x (11 x A + 540) / 30,000)

The damage will use that formula, but the proc chance should use the */40,000.

[DarknessEternal]

I'm gonna get back this sometime, I swear.

[Underfyre]

I dunno, I've combed over a ton of stuff and reconfirmed a bunch of formulas. I really wish I would have done runs that don't involve any/few variables, and I really wish those variables didn't increase damage so much. I guess I'm gunna have to put together a build that is based off of raw damage and move from there.

 

Damage has fallen away from the observed across the board it seems.

[Bopper]

2 hours ago, underfyre said:

I dunno, I've combed over a ton of stuff and reconfirmed a bunch of formulas. I really wish I would have done runs that don't involve any/few variables, and I really wish those variables didn't increase damage so much. I guess I'm gunna have to put together a build that is based off of raw damage and move from there.

 

Damage has fallen away from the observed across the board it seems.

I'll take a look at some of it tonight, but in the mean time can you tell me how you are factoring in resistance debuffs? I'd like to know your steps that way I can trace it in your spreadsheet. Plus, it will allow me to see right away if you're using the right formulas. Right now, that is the only thing I can think of that could have a major impact other than arcanatime being a liar.

 

Edit: I guess I'll have to wait. I thought procs were fixed, but looks like the OP is still outdated on both its links. I'll say this though, something does seem off with resistance calculations. It's slightly underperforming what I would expect, but it's not too drastic yet. One thing I am noticing, something about the resistance cells (G42, G43, and G44) seem to having multiple contributions. They feed into C11, HOWEVER, I overwrote your formula and set C11 to an actual value. When I did that, I noticed the DPS would still change if I edit G42-44. That does not seem good. 

 

Edit 2: Looks like Opportunity is not using the correct column for resistance debuffs. I think it's only looking at the target's resistance as opposed to what their resistance is after you apply resistance debuffs. This might be the small thing I was seeing in my previous edit. Overall, double check this Opportunity cell. It refers to D2, but I see nothing in D2, which *might* be due to it checking B2 which is actually a merged cell. I dunno, but something seems odd.

Edited June 17, 2021 by Bopper

[Underfyre]

The primary resistance data chunk is just: (-Res_All * (1-Target_Res)) - Target_Res, which is broken down into each ResType. This was mostly so I could give the option of different target resistance options and see how each sets damage type compared relatively. 

 

-Res_All is affected by Purple Patch. Not particularly relevant to Pylon testing, and I'm sure you've seen it, but I figured I'd mention it.

 

In abilities it's Base_Damage * (1 + %Damage) * (1 + ResType)

 

 

Annihilation and Fury work the same, they only fit on multi-target abilities, so I just pull the data if it exists. The data exists if the ability is in rotation.

Achilles' is a headache of a cell to read because it may or may not come into play in multi-target rotations.

Resistance Reductions cell is specifically for Sonic Blast. Works like Achilles' with the multi-target stuff.

 

 

And D2. That's a hidden cell (white out data). There was no easy way to say what resist type Opportunity should be using. Water uses Cold. Steam Spray is a Fire attack, there's no Cold cell in the ability chunk to pull from. So I had to make a static cell that flagged what particular resistance Opportunity was using. So D2.

[Underfyre]

I put a link to the sheet I'm currently working with on the main post.

[Underfyre]

9 hours ago, Bopper said:

Edit: I guess I'll have to wait. I thought procs were fixed, but looks like the OP is still outdated on both its links. I'll say this though, something does seem off with resistance calculations. It's slightly underperforming what I would expect, but it's not too drastic yet. One thing I am noticing, something about the resistance cells (G42, G43, and G44) seem to having multiple contributions. They feed into C11, HOWEVER, I overwrote your formula and set C11 to an actual value. When I did that, I noticed the DPS would still change if I edit G42-44. That does not seem good. 

 

Checked this out. This was a call I made, may or may not be bad decision making. So my thought process was that if you have Opportunity active, the target has the full 25% debuff. So I cut out the chunk where it pulls from the main Opportunity -Res cell at the top and do a new calculation in cell, which pulls data for each -Res cell. So the (small) damage fluctuation when you manually change one of those 3 cells is the Opportunity cells reacting.

 

It's not pulling for the sonic debuffs tho, that does need to be fixed.

Edited June 17, 2021 by underfyre

[DarknessEternal]

Any tedious in game work you need done for testing purposes?  

[Underfyre]

5 hours ago, DarknessEternal said:

Any tedious in game work you need done for testing purposes?  

 

So I have a plan to do runs with a build that has as few an amount of variables as possible. Obviously the highest baseline damage would be best, but Fire in itself has variables with its burn lengths. They're cancel on miss. So the next best option is Electric. Bio as a secondary is still viable because the bonus damage is 100% chance with no ticks or anything cancel on miss. So that will thankfully move things along. I'm on the fence about using offense or defense Opportunity. Opportunity itself isn't a variable as long as you aren't using the Opportunity Strikes proc. The offense proc doesn't have any cancel on miss shenanigans, it's just bonus damage, so it should be okay. The way it's coded it just looks to see how long it takes to clear 90 rage, which isn't likely to change aside from healing yourself.

 

I have this build. It has no damaging procs in the abilities I plan to use. Nothing that will alter combat beyond raw damage. Hybrid is only taken for the 10% damage boost. No Interface. I'm going bare bones. If raw damage doesn't match up, there's something fundamentally flawed with the sheet.

Spoiler

| Copy & Paste this data into Mids Reborn : Hero Designer to view the build |
|-------------------------------------------------------------------|
|MxDz;1406;675;1350;HEX;|
|78DA6593594F135114C7EFD02950CA229B1468A12D658761715F1223B244964824F|
|1B569CA58C6D496B435D147BF81EB77B06E9FC4F5CD35C627D9C15D9FEA99F9FF2D|
|244CD2FECE3D73CFF9DF73E6DCB9EBE3954FA76E9E515AF5B9642C9B8D2E98A99C9|
|53293E573D772B19C954E29FB29935FF0FFABE8C5582A612E1A1349339ECB58F158|
|323A26B1B97071C3B879D94C654D63CC4A472F6412B194158F9ECD5C4D6754D57C3|
|A9D3426ADC492EC4C78B14ADACB0AC75E5836CDC51AC79C35638B6626BB642DB74C|
|2C5B71A398DDD68DCE89A099B9E1938345E4F7A55BFE34FBA805B77AA02B35A2AB9|
|23CF910D41F918FC1D22760536F31D6A57EB8C4E751A53FC1F23FE46F70457434EA|
|68A3258EAFE2107918AC3C421E05AB8F813ED1712156B92E698EF6AAE4732BE4730|
|F70FF2078C02087C1BAA11227A62EA09CF59AC496219F56D60F5F431E7B9A45CB83|
|77BAA7115A0D93A8E1E014390DFACE83EB92CFCBDABCCFD01BDF73B0E505F912F4B|
|F225F8379D1AB426CA145EC1AD4A4D7F891BB2D40B691ED6030086E88762D636AB7|
|516F7087FC0A86BE91DFE9179D7A9EB77E057942AB60C71AB94E6E80914D1766422|
|6BA117AAAB1208F7289A789DD6CB28D80AE22A3D8ADCBBB6676B399DD8CB09B5DEC|
|6617BBD93583188FC4B4F26BB732C62BBE0055038EEA96D4DDCEEFDFFE06BDEC7E4|
|BBE23DF83BD1FC88F6087D41FE6DC8639B77D9CDB3ECE6BDF5F9C26247B3BA1A375|
|725606C8B09CAA8795F7D0D74B7E92B87EF6B8FF1674076E9377C0C1BBE43DF23EF|
|859620DC61AC7F1CD864E9027C953E0C8693204DD3A39D330BB374CDF15E79EBA9D|
|19F3EB8A4FC15E4A2767F779E6755E6BF1688E67B3A2787BE9D9DAE7D9DEEB9916A|
|35B573BBB3E4D0BE23423BF50E54CBD80939F0FEFDAADF68D807AC1BFC76EDB63FF|
|03E6FBECF5|
|-------------------------------------------------------------------|

 

Anyway, if you want to also build this up and run some tests, it'll speed things along and give a wider sample pool.

 

Planned rotation is (22)-(7)-1-4-20-6-1-4-1-6, but use whatever you want.

[Underfyre]

Well shit, I forgot to take off the interface. Anyway, here's the 1 with it.

Expected: 232.51

Observed: 266.76

 

Okay, with that out of the way...

Expected: 200.18

 

Observed:

241.95

237.39

241.28

 

 

These 3 are high enough from the get go that I'm just going to say something's fucky.

[Bopper]

I recommend something other than Bio if you're looking for simplicity. Adding a resistance debuff and added adaptation damage are variables I wouldn't want to mess with when trying to identify where the spreadsheet goes wrong. 

[Underfyre]

Okay, so bear with me here. I'm looking at the recording from one of these last runs. I'm thinking to myself, "Maybe it's not my sheet that's the problem. Maybe it's the timing. Arcanatime is from back in the days of dial-up. Maybe that just doesn't really exist now."

 

So abilities are largely just queued up one after another. I made a cut in the video at the exact moment the ability is activated, and another cut when the next ability is activated. Like zoomed in and finding the exact frame after stretching the video out to 25% playback speed. Then I shrunk it back down to 100% playback speed and lined it up with a second marker. This shot is lined up on 24 seconds.

 

 

So my 1.19 second (arcanatime) ability went off, in under 1.19 seconds, and the next ability is firing. I'm going to compare more abilities, but you know where this is going.

Edited June 18, 2021 by underfyre

[Underfyre]

5 minutes ago, Bopper said:

I recommend something other than Bio if you're looking for simplicity. Adding a resistance debuff and added adaptation damage are variables I wouldn't want to mess with when trying to identify where the spreadsheet goes wrong. 

 

Adaptation damage doesn't vary, and it's following the exact formula from the code. The only resistance debuff at play is from Opportunity, which I can prune out. But I don't think it will help. I think the cast time formula is antiquated.

[Bopper]

3 minutes ago, underfyre said:

 

Adaptation damage doesn't vary, and it's following the exact formula from the code. The only resistance debuff at play is from Opportunity, which I can prune out. But I don't think it will help. I think the cast time formula is antiquated.

It's simple enough to factor in, but it's still a variable you don't need to incorporate. You're right, the resistance debuff is to self so I misspoke there. I was thinking of the -res aura, but I think sentinels don't get that power.

[Underfyre]

2 minutes ago, Bopper said:

It's simple enough to factor in, but it's still a variable you don't need to incorporate. You're right, the resistance debuff is to self so I misspoke there. I was thinking of the -res aura, but I think sentinels don't get that power.

 

No fun stuff like that for Sentinels, just the damage boost. I'm just trying to keep these 5 minute runs not be 7 minutes runs. It's a constant, and the Pylon has 20% resists across the board, so that isn't a factor either.

 

But onto the new factor I'm blaming. At what point does one ability stop and the next one consider to have started. Cuz I'm looking at the animations, and when the damage is caused in chat, and when the next ability triggers on the bar. Everything is out of sync with each other.

[Bopper]

12 minutes ago, underfyre said:

 

No fun stuff like that for Sentinels, just the damage boost. I'm just trying to keep these 5 minute runs not be 7 minutes runs. It's a constant, and the Pylon has 20% resists across the board, so that isn't a factor either.

 

But onto the new factor I'm blaming. At what point does one ability stop and the next one consider to have started. Cuz I'm looking at the animations, and when the damage is caused in chat, and when the next ability triggers on the bar. Everything is out of sync with each other.

I think I brought up that option previously where arcanatime might not really be arcanatime. I believe Number Six ran a test awhile back and they were experiencing something like 0.138s instead of 0.132s. So if you want to just try that, go for it. Change your formula to use that value and see what it does.

 

 

Edited June 18, 2021 by Bopper
Edited value. 0.138s was the tested result, not 0.139s

[Underfyre]

Havoc Punch:

Cast time: 1.5

arcanatime 1.72

Animation time: 1.5

Anim before hit: .83

 

Observed:

Cast time: 1.483

Animation time: ???

Anim before hit: 1

 

[Underfyre]

Zapping Bolt

Cast time: 1

arcanatime: 1.19

Anim before hit: .73

 

Cast time: 1.033

Anim before hit: .75

 

[Underfyre]

19 minutes ago, Bopper said:

I think I brought up that option previously were arcanatime might not really be arcanatime. I believe Number Six ran a test awhile back and they were experiencing something like 0.139s instead of 0.132s. So if you want to just try that, go for it. Change your formula to use that value and see what it does.

 

Well the problem is my observed damage is 20% higher than my predicted damage, so adding more cast time isn't going to help much. Just putting in the numbers for plain old cast time gets things closer, but it' still off. Now I'm looking at 225.31 DPS.

 

So just looking at Sparky's damage. This is with just the 5% resistance debuff from Opportunity. so 16% resistance.

Predicted is 68.80 per attack + 8.43 from Bio.

Observed is 68.74 + 8.87

 

I mean, it's "just" off by 0.39, but it shouldn't be off by anything at all.

 

Lightning Field ticks:

Predicted: 22.34 + 0.98

Observed: 22.33 + 0.97

 

That's what I want to see.

 

Charged Bolts:

Predicted: 128.74 + 11.10

Observed: 128.73 + 11.08

 

Zapping Bolt:

Predicted: 264.45 + 21.75

Observed: 264.43 + 21.73

 

Havoc Punch:

Predicted: 216.15 + 47.03

Observed: 334.72 + 39.49

 

I think I found the culprit.

[Underfyre]

Ok.... Sparky is apparently fucking with me....

 

 

I wasn't watching the whole time, and I picked up the Analyzer after I did my runs, but he's applying his own 5% debuff.

[Underfyre]

After fixing the Havoc Punch slot and adding 5% more -Res for Sparky's contributions, and changing cast times to their base values (no arcana time), we have..

 

Expected DPS of 234.24

 

Observed:

241.95

237.39

241.28

 

For gits and shiggles, the expected DPS with Reactive Radial is now 270.08 versus the observed 266.76

Edited June 18, 2021 by underfyre

[Bopper]

1 hour ago, underfyre said:

Well the problem is my observed damage is 20% higher than my predicted damage, so adding more cast time isn't going to help much

It won't. But tested gameplay shows this to be in game performance so it is something you'll want to keep in mind when attempting to validate a tool that predicts what that in game performance will be.

 

If you're seeing your in game performance to be better than arcanatime, that would be worthy of a new thread that investigates that claim. You'll probably want to do long duration tests and look for what your performance truly looks like. If you are seeing something different than what other users see, then you may want to get other users to provide you with their own performance results to see if there is a discrepancy.

[DarknessEternal]

Well this escalated in complexity.

[Bopper]

6 minutes ago, DarknessEternal said:

Well this escalated in complexity.

Lol, I might have said something similar a year+ ago when I was working with a project team for building a simulator for predicting pylon time performance. Getting simple powersets to work was no problem, but then combo systems and other special mechanics just made it a never ending process.