New Alacrity data for Combat Medics

RuQu · February 8, 2012

*edit* That data was old. Replaced it with the new.

My latest alacrity chart can be seen here.

There are separate scales for the right and left. The right is for the total number of casts, and is on a different scale so it doesn't wash out the detail of the other abilities.

They can be seen with the same scale here.

This was done over a fixed time interval, which should show the benefit of Alacrity for getting in extra casts in the same time period.

We do see this, but what is clear, see the second chart, is that all of the increase in Cast Counts come from increased HS casts (red line), and, in fact, the total number of AP and MP casts drops continually from a starting value of 122 casts to 106 at 39% alacrity. This makes sense considering the fact that Alacrity increases the Effective Cost of abilities.

Looking at the first chart, we can examine the details. We see the decrease in AP and MP usage and increase in HS usage as described above. We also see that, over a 300 second (5 minute) fight, the alacrity increases only earn you 1 extra SCC at 10.5% alacrity, and 2 extra at a very high 30% alacrity. Any claims that alacrity is beneficial because of increased SCC usage are clearly inaccurate.

We also see that our resource mechanics again put a limit on the benefit of Alacrity. This simulation was run with the restriction that AP would not drop us below 4 Ammo during SCC usage. With that restriction in place, we see that there is a limit on the number of AP and MP casts we can get in during a SCC phase, no matter how fast those two spells cast, after which the SCC phase has to be filled with HS casts.

We start out being able to cast AP and MP 66 times during SCC, over a 5 minute fight. This holds steady until 5% Alacrity, where we first manage to fit in extra usages. It eventually comes back down, and from 18.5-31% it is again in the 66-70 range, with some slight fluctuations.

Where do we see the most benefit? From 5-7.5%, with a slightly lower usage in 7.5-14%, but that is still more than you get in at 0-4.5%.

Where do we start seeing Hammer Shot being used during SCC? 8.5%, although then only rarely, likely when entering SCC at a lower initial ammo level. By 15%, we are seeing usage with every SCC, and at 33% we are seeing two Hammer Shot casts per SCC.

So where does this leave us? I would say we want to aim for about 5.5-7% cast time reduction. This should give us enough extra time in SCC to toss out that final Kolto Bomb without clipping our AP/MP cycle, and doesn't waste any itemization sneaking HS casts into SCC.

Once I convert the Combat Medic Alacrity tool over to work with Bodyguards, I will toss the latest version of the software up on Google Code so people can produce similar results with different assumptions. That will be version 1.0.6b, so if you still see 1.0.5, I haven't posted it yet.

Edited February 16, 2012 by RuQu

dotriche · February 12, 2012

How does this not have any replies?

Bump

Aritok · February 13, 2012

Excel charts are cool. Here is mine:

HPS vs Spell haste

And this one shows that your HPS will go up as higher your spell haste is.

How is it calculated? Well:

It is checked over a give time. In this case 10 min. Ammo is always at max regen.

- Now we check how many ammo you gain in this time from normal regen and spells like recharge cells.

- After this we check how many AMP we can cast with the cooldown. Same number for the MP for the ammo reduce gain.

- Check the max number of BI casts, depend on the cooldown.

- See how many SCC you can get from the MP (only MP, ignoring HS it make it easier).

- lets say we cast 3x AMP (and therfore same number of MP) and one KB during each SCC.

- Now we check how many time and ammo is left, and use this for the max MP.

- the remaining time will be spend in HS.

And so easy you get a nice chart with numbers that show something total different .

God, i love Excel and statistic. You can make everthing come true as you want

Within these HPS are no crits, no surge. As they would not change the trend.

If i have time, i could do a cross link for these, to compare the Alacrity vs surge and what is better.

Btw, this chart does not guarantee any accuracy of statement. There can be still errors in the basic idea of the formular. But i can give the datastring to everyone who want to have a look on it. Maybe someone find a major error.

Ellie_ · February 13, 2012

Excel charts are cool. Here is mine:

HPS vs Spell haste

And this one shows that your HPS will go up as higher your spell haste is.

How is it calculated? Well:
It is checked over a give time. In this case 10 min. Ammo is always at max regen.

- Now we check how many ammo you gain in this time from normal regen and spells like recharge cells.
- After this we check how many AMP we can cast with the cooldown. Same number for the MP for the ammo reduce gain.
- Check the max number of BI casts, depend on the cooldown.
- See how many SCC you can get from the MP (only MP, ignoring HS it make it easier).
- lets say we cast 3x AMP (and therfore same number of MP) and one KB during each SCC.
- Now we check how many time and ammo is left, and use this for the max MP.
- the remaining time will be spend in HS.

And so easy you get a nice chart with numbers that show something total different .

God, i love Excel and statistic. You can make everthing come true as you want

Within these HPS are no crits, no surge. As they would not change the trend.

If i have time, i could do a cross link for these, to compare the Alacrity vs surge and what is better.

Btw, this chart does not guarantee any accuracy of statement. There can be still errors in the basic idea of the formular. But i can give the datastring to everyone who want to have a look on it. Maybe someone find a major error.

The only flaw with that logic is that every point spent in alacrity is one point less in other stats, so if you dont even calculate the value of other stats, obviously alacrity will reign superior, there being no competition...

And in a static rotation alacrity starts to lose out anyway because you will complete the large casts faster, yet healing for less than if you stacked other stats instead, and being forced to fill in with hammershot sooner to avoid running out of ammo.

Then you also have to take in account things like critical hit sizes scaling with both aim/power and surge, but not alacrity. It's just not that simple.

Edited February 13, 2012 by Ellie_

Aritok · February 13, 2012

As i said, i didnt take surge in for this. But OT didnt say anything how his numbers come together as well .

And i didnt add logarithmic scale of the alacrity as well. If you do this, the graph will look total different again.

Maybe i can add the two diagramms together that i have to get something that even add surge.

But this will be more than difficult, becauce there are as well skills that give spell haste. With only spell haste on the x-axis, it dont matter how much you get from skills. And you should not forget, that your max. Spell haste is currently 39%, not 50% (for Commando).

If i cant to make it accurate, i need 4 curves (with 0%, 4%, 5%, 9% Spellhaste) combined with correct alacrity value in comparison to surge on a logarithmic scale.

And this way i didnt even consider the Augments, that could be replaced with every stat.

If you want such a chart, with all options, give me 2 years, and i will make one for you

Ellie_ · February 13, 2012

Excuse me Sir, but the OP has a colossal thread at the top of this forum section called "Combat Medic Guide" with the full spreadsheet and explanation. You would find your answers there. Edited February 13, 2012 by Ellie_

Aritok · February 13, 2012

I saw the sheet. And i see the work in it. But as i said already on another thread, there is something strange in it. A change of alacrity of +50 suddenly lower the HPS by alot, another 50 and the HPS is going up again. Without changing anyhting. This should not be this way.

And this curve is telling me the same again. This "break" in the trend of the line, look strange. Like an error in the data. All but the 3 or 4 values with the peak look good.

And a curve, with only one var (Alacrity) can look total different if you use another base to calculate it. The one RuQu is showing, and the one i am showing, have the same curve (HPS). Just calculated on different basics. He is using a base of 300 casts, i am using a base of 10 min. Both are correct, within their own parameters.

BTW, my rotaion over a fixed time, has only one part that is changing. And thats the number of Hammershots you have to fire. ALL comes down to a Ammo problem. You just cant cast more than a specific number of ammo spells in the given time. Becauce your ammo is going out. And this is why my curve is showing a upwward trend. More HS in a fixed time where everything else stay the same = more HPS.

As i said. It all comes down to the basics of the calculation.

But here is another one:

HPS in relation of Alacrity, Surge, Crit

Alacrity and Surge are in sum always 600 in this one. Becauce these two stats are sharing the Enhancement slot, it comes down to these two to compare.

It really is important what your crit value is. Low Crit means alacrity is better, high crit makes surge alot better.

And dont get me wrong. I am not trying to make Alacrity good, or better than all other or something. Alone from the formular of the secondary stats, its the best to mix all.

But i prefer a heal that hit in time, becauce of the faster cast time, instead of a big crit heal that his hitting 0.1 sec to late on the dead tank .

But i thought as well the best peak value would be higher. But from what i saw on my sheet, its really lower than i was thinking. and it comes down to an easy answer if i have to think about alacrity or surge. And it will be surge.

RuQu · February 13, 2012

I mostly hang out on the healer forum, so actually missed this topic coming back up.

I understand what you are trying to say, which is that based on different assumptions, you can get different results.

In fact, I changed my assumptions shortly after the post with those images up there. Some people, like yourself, argued that alacrity should never decrease your HPS (ignoring the opportunity cost). Of course, this ignores the interaction with our resource mechanics, and the increased Effective Cost. Those sharp lines are where increased EC caused us to drop into lower regen after SCC after fitting in extra AP and MP casts.

Revising the logic to never cast AP or MP in SCC if it would put you into low regen, flattens it out a bit. The assumption you run on that we would never leave high is a false assumption, you drop into medium during SCC all the time.

The newer, flatter plot can seen here. The peaks get shifted and flattened, but they are still there. What is really going on can be seen by the yellow line at the bottom. That is the number of times you have to fill time during an SCC with a HS cast because you can't afford to cast AP or MP.

Note that there is a rise at around 5.5%, which is where HPS dropped with the more aggressive assumptions. So previously casting AP then was dropping us below 3. Now that doesn't happen, the peak is hidden, but the problem of wasted alacrity is still there.

Any time you are casting HS during SCC because you can't afford something better is wasted points in Alacrity. HS scales terribly, at about 1/3 the rate of MP, so any points spent to increase HS usage would be better spent on anything else.

Note that the HS usage during SCC rises again at around 16% and grows from there. For no reason should you have more Alacrity than that.

My latest version of my code can be found here. My assumptions for Combat Medic rotations are stated on this wiki page.

I'm more than happy to continue to discuss it, but all assumptions need to be laid out explicitly. If you take the time to model it properly, I think you will come to realize that anyone who thinks, ignoring opportunity cost, that Alacrity will always result in positive HPS has either an overly simplistic model, or unrealistic ideas about how a person plays the game (ie stacking alacrity for a 1.5 second MP, then waiting 0.5 seconds after it before casting another ability to allow the EC to stay the same. Note that you do not get the same result by casting 3 times and then "waiting" with a HS).

Currently I do not have my program written to export HPS curves, but you are welcome to manually iterate it and make a plot. You can change the thresholds where 3 Ammo MP is deemed affordable outside SCC, AP is affordable during SCC, when you use RC, and if RC restores 6 or 8 Ammo.

TempestasSilva · February 15, 2012

Subscribed

-sasori · February 15, 2012

I'm more than happy to continue to discuss it, but all assumptions need to be laid out explicitly. If you take the time to model it properly, I think you will come to realize that anyone who thinks, ignoring opportunity cost, that Alacrity will always result in positive HPS has either an overly simplistic model, or unrealistic ideas about how a person plays the game (ie stacking alacrity for a 1.5 second MP, then waiting 0.5 seconds after it before casting another ability to allow the EC to stay the same. Note that you do not get the same result by casting 3 times and then "waiting" with a HS).

Currently I do not have my program written to export HPS curves, but you are welcome to manually iterate it and make a plot. You can change the thresholds where 3 Ammo MP is deemed affordable outside SCC, AP is affordable during SCC, when you use RC, and if RC restores 6 or 8 Ammo.

It's sadly the other way around. Your base model is you just stand there and heal. You target can't die because it does not matter when the healing is applied as long as the HPS is high enough. Like as long as i heal 50k after 10 seconds i get a hps of 5K right?

This is not the case not for every pve encounter and least of all in PVP. The reasons people stack alacrity is not because they want to maximize amo / HPS. They want to counter Bursts with upfront HPS.

Alacrity is about reducing setback / possibility and even neutralizing it if you can cast fast enough. (1 sec AMP).

It's about utilization not raw power. Your data is valid but not your assumptions about its "use", rather your model abuses alacrity.

Using a real world example i don't need to heal enough in Eternity Vault for alacrity becoming crippling to me due to increased ammo costs. But i can spot heal where i need faster. I can charge bolt spam better with SCC active during vulnerably phases to help raid Damage on the Boss. (Currently 1.1 sec)

The time windows are too small for me to worry about whenever i could have used two more Charged Bolts if SCC could somehow run 10 seconds more (which it never can). Neither if i could deal more damage to the boss IF he would be vulnerable for additional 10 seconds, because i had two ammo more.

The time phase where Surge / Power catch up simply never comes. During short time windows every skill who successfully fires of is more valuable than a raw increase.

I'm not saying alacrity is the best, all i'm saying alacrity has his own field of expertize and shouldn't contest in a raw HPS to AMMO competition for evaluation.

Edited February 15, 2012 by -sasori

maxbaby · February 15, 2012

It's sadly the other way around. Your base model is you just stand there and heal. You target can't die because it does not matter when the healing is applied as long as the HPS is high enough. Like as long as i heal 50k after 10 seconds i get a hps of 5K right?

This is not the case not for every pve encounter and least of all in PVP. The reasons people stack alacrity is not because they want to maximize amo / HPS. They want to counter Bursts with upfront HPS.

Alacrity is about reducing setback / possibility and even neutralizing it if you can cast fast enough. (1 sec AMP).

It's about utilization not raw power. Your data is valid but not your assumptions about its "use", rather your model abuses alacrity.

Using a real world example i don't need to heal enough in Eternity Vault for alacrity becoming crippling to me due to increased ammo costs. But i can spot heal where i need faster. I can charge bolt spam better with SCC active during vulnerably phases to help raid Damage on the Boss. (Currently 1.1 sec)

The time windows are too small for me to worry about whenever i could have used two more Charged Bolts if SCC could somehow run 10 seconds more (which it never can). Neither if i could deal more damage to the boss IF he would be vulnerable for additional 10 seconds.

The time phase where Surge / Power catch up simply never comes.

I have another real world example:

Fabricator where raid gets damage from boss, stun droids, explosion droids, tank from fire. The damage is constant. Healing is constant. If you stack alacrity - you just use your ammo very quickly. Since you heal 0.4 seconds faster but your heals are weaker you have to spam heal until you run dry.

My guild is playing on nightmare the damage is higher than hard mode.

Managing SCC while relying on crits is more beneficial than heal faster and less.

PvP is another matter. I don't care about PvP. Maybe alacrity works better there.

germil · February 15, 2012

Yeah I agree, the intention is good, but there is not so much reason to do model alacrity, at least with taking into account ammo cost.

In you standard non-pressure rotation (keeping high ammo all the time), the healing ouptut is limited by ammo regen (and other stats). Only use of alacrity is to be able to cast more hammer shot, which translates theorically into more heal, and more frequent supercharges, but these advantages are very small. Alacrity does not bring anything significant here, compared to other stats (well it would be interesting to model that (for example a model which keep always ammo >60%), but the result seems already obvious)

It is when you are pressured to produced high HPS (even at the cost of being low ammo ) than the row HPS gain of alacrity really means something. It would then make sense that alacrity is more usefull than other stats in this situation.

-sasori · February 15, 2012

I have another real world example:

Fabricator where raid gets damage from boss, stun droids, explosion droids, tank from fire. The damage is constant. Healing is constant. If you stack alacrity - you just use your ammo very quickly. Since you heal 0.4 seconds faster but your heals are weaker you have to spam heal until you run dry.
My guild is playing on nightmare the damage is higher than hard mode.
Managing SCC while relying on crits is more beneficial than heal faster and less.

PvP is another matter. I don't care about PvP. Maybe alacrity works better there.

Then you would need to prove that my 0.4 faster and lower heal was the issue, because as far as i see it it's not always necessary to heal more than 1 time just because i heal (random number value) 250 less upfront with an MP.

constant healing is a difficult term because the question remains if it is constant enough that:

1. my ammo regen does really becomes crippling when i CHAIN heal enough (12 casts + (no brakes).

2. If you really need to immediate CHAIN and not just simply overheal by doing so.

3. The upfront heal is significant higher than with the alacrity you are comparing too. Best case 1.5 Heal stronger or higher.

If one and two really are the case, ye then power / surge get's better slightly better and with three as a case there is no competition and alacrity straight loses.

But if i need to counter just 6k in roughly 3 seconds i might as well take whatever i like.

2 Casts > 5.5k - 10k (more or less depending on crit luck and gear, just a random value here too)

Now if i use

1x AMP (1.0sec) + 1x MP(1.4sec) + hammershot (1.5 gcd)

Or

1x AMP (1.5) + 1x MP (1.8Sec) (and roughly 800 more heal with AMP+MP)

= result is the same. Then i can still switch targets and heal there.

The used values are just random, but you see it's hard to find a sweet spot to clearly say, alacrity is at a disadvantage, or power / surge at an advantage.

And numbers are not that extreme far off to each other either. Most people would have still a mix of alacrity / power / surge and neither extreme.

Edited February 15, 2012 by -sasori

Aritok · February 15, 2012

@sasori

I tried as well to get faster heals over bigger heals. On one side for the faster patch heals. Just becauce it is better the heal hits, but for less, over a heal on a dead tank.

The problem is, you need an alacrity value of 1350 for your 1 sec AMP. And this already count in the 9% you get from skills.

But you can only get somewhere around 800 from your normal items (without augments).

The other downside for this, Alacrity has a way worse factor than Surge has. A value of 600 Alacrity give you a 52% of max 30% (=15,68% Haste).

But 600 Surge gives you 91,15% of max 50% (=45,57% crit DMG).

Your surge will be long time in overkill range, while Alacrity is not even at the 63% mark.

These was with old surge formular. No idea how much it moved now.

A mix of both will be the best. As well becauce of the DR you get from to high values.

RuQu · February 15, 2012

It's sadly the other way around. Your base model is you just stand there and heal. You target can't die because it does not matter when the healing is applied as long as the HPS is high enough. Like as long as i heal 50k after 10 seconds i get a hps of 5K right?

Yes, the HPS is calculated at the end. Of course, you could never do 50k in the last 2 seconds of the 10s window, so I'm not sure what the point is here. Alacrity increases your HPS over the long term by allowing you to cast HS, our weakest and worst scaling ability, more often. Every other stat increases HPS by improving all of our abilities. The effect on AP and MP usage in the long term is marginal because they are limited by Ammo and cooldowns, not cast time. This isn't a case of infinite resources and no cooldowns where if you could just get MP down to 1 second you would be a crazy HPS generating machine.

This is not the case not for every pve encounter and least of all in PVP. The reasons people stack alacrity is not because they want to maximize amo / HPS. They want to counter Bursts with upfront HPS.

Alacrity is about reducing setback / possibility and even neutralizing it if you can cast fast enough. (1 sec AMP).

It's about utilization not raw power. Your data is valid but not your assumptions about its "use", rather your model abuses alacrity.

I've made it abundantly clear in most of my posts that I don't care about PvP. It is a mini-game where healers are killed first and their resources reset on death, so the resource mechanic is essentially neutralized. Yes, alacrity is probably of great value in PvP. Burn through the ammo, die, repeat. That is also completely irrelevant to any situation where Ammo matters.

Using a real world example i don't need to heal enough in Eternity Vault for alacrity becoming crippling to me due to increased ammo costs. But i can spot heal where i need faster. I can charge bolt spam better with SCC active during vulnerably phases to help raid Damage on the Boss. (Currently 1.1 sec)

The time windows are too small for me to worry about whenever i could have used two more Charged Bolts if SCC could somehow run 10 seconds more (which it never can). Neither if i could deal more damage to the boss IF he would be vulnerable for additional 10 seconds, because i had two ammo more.

The time phase where Surge / Power catch up simply never comes. During short time windows every skill who successfully fires of is more valuable than a raw increase.

I'm not saying alacrity is the best, all i'm saying alacrity has his own field of expertize and shouldn't contest in a raw HPS to AMMO competition for evaluation.

Obviously one of the assumptions is a low mobility fight, where you are only moving to dodge some fire and you could fit that in during a couple of Hammer Shots.

High mobility fights, if you would prefer to model your stat weights like that, can be viewed as a number of small fights. Coming from full ammo and all cooldowns up, you are at your moment of max healing capability. High mobility fights tend to put out less damage during the mobility phases because designers know healers can't heal much on the run. There is usually time for a burst of healing following it.

So high mobility fights tend to take the form: sustained, running, burst heal, sustained, running, burst, repeat.

Running is a resource regen period. If you want to approach it, try setting the fixed time on the calculator to something low, like 30s.

Then there are fights like Fabricator, which maxbaby described.

You can say "all fights are different, models are useless" and just guess at what stats are best based on gut feeling, or you can pick a fight that is easy to model, ie a low mobility fight where damage is consistently coming in, and use that as the basis for doing some comparison.

It's also worth pointing out that, without a model we know these things:

1) Power/Crit/Surge benefit all of our abilities.

2) Alacrity benefits AP and MP on the very short term, and in the long term only means extra HS casts.

Juts looking at those two statements it is clear that alacrity is weaker. All of the other stats have either no DR (power) or DR that is clear and easy to model (a log curve). Only alacrity has an additional DR that is tied to its relation to Effective Cost (EC). The equation for the log part of alacrity's DR is straightforward, and the same as the other ratings. It is this second diminishing return, the one tied to EC, that we are trying to model.

So far all you have done is say "sometimes the fight doesn't match your assumptions. This is true, but sometimes it does and that gives us a place to assess stat weights. On other fight assumptions they might differ slightly, but it is unlikely they will differ much. Some fights might allow alacrity to produce the same result, by providing rest times while running, and a high alacrity build might perform equally in that situation. However, that high alacrity build will not perform the equally in the low mobility.

So if you had the choice between a build that does well in both a high and low mobility situation or one that only works when the fight allows plenty of downtime, which do you choose? I choose the one that works in both, which, lucky me, also happens to be the easier one to model. Constant healing with burst whenever you can (every time SCC is up) sets a sort of upper limit. Any mechanic that allows for rest just makes it easier

Stating up front that I have no intention of coding 15 different rotations, I am open to suggestions on how you think the rotation might be better. But if you agree with the rotation logic, this model puts the most constraints on it, and, I think, best provides us the ability to look at how to change stats to improve performance.

RuQu · February 16, 2012