Quote:
No.|
Quote:
Quote:
|
Logically speaking the only thing that statistics showing that an AT earns rewards at the same general rate as all ATs tells you is that any problems with the AT are not affecting its earning rate. You keep speaking of reward rates as if they in themselves were significant. If this were the case the game could just give people rewards for staying logged in all ATs gaining the same fixed rate. Problem solved.
As to their not being an Aristotelian ideal for performance, in the context of a game barring software errors, the performance for any entity is very much an ideal. The players can and will pick up on what is better and worse and act accordingly. Not working from the model just means you are guessing when looking for problems or when correcting them.
You have to ignore so much and make so many assumptions for the idea of sampling the system and applying corrections based on a particular snapshot in time without a model of performance to guide the changes it boggles the mind.
Quote:
You do, I don't. I also have no specific issue with boggling you while I go about actually getting things done.
|
Originally Posted by Another_Fan
You have to ignore so much and make so many assumptions for the idea of sampling the system and applying corrections based on a particular snapshot in time without a model of performance to guide the changes it boggles the mind.
|
[Guide to Defense] [Scrapper Secondaries Comparison] [Archetype Popularity Analysis]
In one little corner of the universe, there's nothing more irritating than a misfile...
(Please support the best webcomic about a cosmic universal realignment by impaired angelic interference resulting in identity crisis angst. Or I release the pigmy water thieves.)
Quote:
Getting things done. Interesting choice of words.
|
Originally Posted by Arcanaville
You do, I don't. I also have no specific issue with boggling you while I go about actually getting things done.
|
Quote:
I had to get rid of a lot of T's, N's, and G's from my Scrabble tray.
|
Originally Posted by Another_Fan
Getting things done. Interesting choice of words.
|
[Guide to Defense] [Scrapper Secondaries Comparison] [Archetype Popularity Analysis]
In one little corner of the universe, there's nothing more irritating than a misfile...
(Please support the best webcomic about a cosmic universal realignment by impaired angelic interference resulting in identity crisis angst. Or I release the pigmy water thieves.)
Quote:
You should analyze how often you have them and see if there is a correlation with how many times you call Iceland.
|
Originally Posted by Arcanaville
I had to get rid of a lot of T's, N's, and G's from my Scrabble tray.
|
Quote:
Too much work. I'm going to play a game of Scrabble and wait for someone from Iceland to call.
|
Originally Posted by Another_Fan
You should analyze how often you have them and see if there is a correlation with how many times you call Iceland.
|
[Guide to Defense] [Scrapper Secondaries Comparison] [Archetype Popularity Analysis]
In one little corner of the universe, there's nothing more irritating than a misfile...
(Please support the best webcomic about a cosmic universal realignment by impaired angelic interference resulting in identity crisis angst. Or I release the pigmy water thieves.)
This little project intrigues me, and I have no vested interest in either side of the argument, I will join in when I get the chance. I am not terribly familiar with blasters, so anyone want to throw out some suggestions for powersets on a blaster and stalker? Something mid-range...if we were talking scrappers, brutes, or controllers I'd have a much better idea of what to use...
Currently Playing:
Rage King - SS/Regen Brute (50+3)
Soulfire Darkness - Dark/Fire Tank (50+2)
Deaths Final Embrace - Kat/Dark Brute (50+3)
ULTIMATE REGEN GUIDE I22
Quote:
I'm relatively curious about how it'll turn out myself and I'm pretty solidly in the 'liking to see blasters improved' camp.|
Originally Posted by planet_J
This little project intrigues me, and I have no vested interest in either side of the argument, I will join in when I get the chance. I am not terribly familiar with blasters, so anyone want to throw out some suggestions for powersets on a blaster and stalker? Something mid-range...if we were talking scrappers, brutes, or controllers I'd have a much better idea of what to use...
|
The general idea (I think) is more comparing a blaster to another damage AT since that's what people are doing in general; not to stalkers specifically. So I'd say feel free to pick a scrapper or brute too. As far as the blaster goes maybe energy/electric? Energy blast is pretty dang middle of the road when it comes to damage and the knockback really helps with survivability when you're solo.
MA Arcs: Yarmouth 1509 and 58812
I have been following all these threads off an on and have to ask a few questions.
If the experiments show blasters can be played at the same level as other ATs and the data mining from the game show that they aren't, what does that mean ?
Is there room in the game for an AT that just takes more effort to play ?
If the AT can be made to perform well with just the tools the game provides but there are observed performance problems are the problems a player issue or a developer problem.
If the game is about telling stories of heroes and their struggles shouldn't the story of the hero that does overcome obstacles and suffers to do so have a place ?
Ok, a little late, but here are the results for today.. I just finished up. The Stalker got wrecked today. Both characters died once on the way out of the bank after defeating the hero, they just got rushed by Longbow. The Stalker struggled against Goldbrickers though... I knew my endurance couldn't sustain the ranged defense toggle at this point so I went with Caltrops at level 10 (endurance is already rough on the Stalker.) This lead to getting hit with ranged energy (I think it was energy) damage and dying twice. The Blaster was able to immobilize them and debuff their tohit, and that combined with hover meant that they didn't present as much of a challenge to him.
Elec/Ninj Stalker
Total levels: 1
Total defeats: 3
7:22 Atlas Park bank (1 defeats)
7:29 mission complete
7:30 defeated (hospital on map)
7:33 exit
Mr. Bocor
7:36 Retrieve the Loa Bone (0 defeats)
7:49 Steal the book for Bocor (0 defeats)
Auto-complete after glowie. Defeat all, couldn't find last enemies
7:59 lvl 10
8:08 Goldbrickers glowie Newspaper (2 defeats)
8:10 defeated (hospital, 8:11 returned)
8:14 defeated (awaken)
8:21 complete
DP/Dark Blaster
Total levels: 1
Total defeats: 1
12:50 Atlas Park bank (1 defeats)
12:57 Mission Complete
12:57 defeated (hospital on map)
1:00 exit
Mr. Bocor
1:03 Retrieve the Loa Bone (0 defeats)
1:15 Steal the book for Bocor (0 defeats)
1:22 Goldbrickers defeat boss Newspaper (0 defeats)
1:31 lvl 10
1:35 Trolls defeat boss Newspaper (0 defeats)
1:42 Skulls glowie Newspaper (0 defeats)
1:49 complete
Quote:
Absolutely, as an Epic AT. Those should be the Advanced Classes, the ones that are harder to play but allow the experienced player to reap more dividends. It certainly should not be an AT that's available out of the box to every player immediately and theoretically has its own niche that makes players want to play it over any other AT.|
Originally Posted by B_L_Angel
Is there room in the game for an AT that just takes more effort to play ?
If the AT can be made to perform well with just the tools the game provides but there are observed performance problems are the problems a player issue or a developer problem. If the game is about telling stories of heroes and their struggles shouldn't the story of the hero that does overcome obstacles and suffers to do so have a place ? |
Blasters are popular. Newbies like them. That's not a good position to put a difficult-but-rewarding AT.
De minimis non curat Lex Luthor.
Quote:
And that was their design decision and, quite possibly, their mistake. That wasn't the fault of statistics. This is the fault of the Devs constructing their Theory of TF Merit Rewards, then using it without performing any of the iterative processes we've discussed. In general, the Devs are very reluctant to do any kind of iterative tweaking, as they like to "get it right the first time" (even if there's no way to know for sure) then only return to it a few years later.|
Originally Posted by Another_Fan
The devs specifically used median run time and the following formula.
(MedianTime / MPM) * TaskModifier * TimesRunModifier * TimeModifier + ArtificialModifiers http://paragonwiki.com/wiki/Merit_Rewards In other words they took a couple of statistics for each TF and allowed it to set the run times. |
As ArcanaVille said above, one of the primary skills behind statistical analysis is knowing how to read it properly, and knowing what it does and doesn't tell you. It could tell the Devs that the Rewards are imbalanced between two TFs, but it doesn't necessarily tell the Devs how to fix it: it can help, by giving clues as to why something is faster or slower, but it doesn't tell the Devs "construct a formula and use that, don't tweak the content of the TFs".
Quote:
| See that is a really interesting idea lets see what the problem is with it. You are using a sample set of the people currently playing blasters. We know that blasters start as the most popular archetype and end as the third most popular archetype. Intrinsically sampling the statistics of people playing blasters to search for signals of a problem understates the players that found the problem and just decided not to play blasters anymore, while at the same time overstating the effects of people that never felt there was a problem. |
Quote:
| Now if you go about things by modeling blaster performance you are in a much better situation. In terms of the game I don't have to experiment to validate my model the way I would in the real world. Why ? Because barring bugs I know what the maximum damage output can be, I Know how many hitpoints they can have and I know how rapidly and under what conditions they can recover them. For years I have been saying these come up short relative to the other ATs. You don't need to sample the data of player base to know that blasters will underperform. It is written on their very D.N.A.. What looking at the data of how a particular set of players is doing with them in the context of the wider game just processes their inherent strengths and weaknesses through a series of funhouse mirrors. |
Quote:
| Anyway on to the history of science portion. |
Quote:
| Neither Kepler's laws of planetary motion, nor the Heliocentric system were developed with use of the telescope. When you say people used it to predict things you mean they used it to predict things incorrectly. The fact that the tables derived from ptolemy's epicycles were predicting things incorrectly helped prompt the development of the heliocentric model. The telescope was what allowed Galileo to demolish any argument that the universe was geocentric in nature because you could look through it and see that parts clearly were not. |
Kepler's Laws of Motions were created in response to Tycho Brahe's observations, which were in direct contradiction to the predictions made by the Ptolemaic system (e.g. that all stars would be close enough to have an observable parallax effect). Making predictions, finding those predictions to be incorrect, then correcting your model (going from geocentric to heliocentric) is exactly the way the scientific process works.
Quote:
| The world view was that waves need a medium to propagate. When you say magically invent more accurate models, in this case it would mean logically exploring the consequences of the idea that light waves did not need a medium to propagate. This is very similar to epicycle being invented to preserve the concept that the earth was at the center of the universe, |
Quote:
| Regarding Test Driven Development, we used to call that trial and error. It may work but its a sure sign that the people using it just don't have a better method available. |
Global @Diellan - 5M2M
Mids' Hero/Villain Designer Lead
Virtue Server
Redside: Lorenzo Mondavi
Blueside: Alex Rabinovich
Got a Mids suggestion? Want to report a Mids bug?
http://boards.cityofheroes.com/showthread.php?t=285444
just so everyone knows, I'm already doing the same on the above thread, but with an ice/ice blaster. I'm going a bit more slowly than THB (not playing every day, etc), so patrol xp should be a factor (if I start dying after lvl10, I'll lose the debt). I'm using HeroStats to track my progression.
I may just join this experiment by creating (yet) another MM (would've done Stalker, but THB beat me to it) to play, though I already know the results: my MM, whatever build, will level faster and never die. Whether that is because I'm better at MMs, I enjoy them, or there aren't really any enemy groups out there that kill all my pets in under a minute, who knows?
I'm tempted to try something similar to this experiment (regardless of its utility AS an experiment), just for my personal satisfaction. I'd try it with an Eng/Eng blaster (a combo which on my main has so far netted over 7,000 hours on patrol) versus a Scrapper or Controller. If I get the time, I'll give it a go, with some rather specific rules in place to minimise preference bias.
The wisdom of Shadowe: Ghostraptor: The Shadowe is wise ...; FFM: Shadowe is no longer wise.
; Techbot_Alpha: Also, what Shadowe said. It seems he is still somewhat wise 
; Bull Throttle: Shadowe was unwise in this instance...; Rock_Powerfist: in this instance Shadowe is wise.; Techbot_Alpha: Shadowe is very wise *nods*; Zortel: *Quotable line about Shadowe being wise goes here.*
Quote:
I'd suggest Mercs/Pain. It's a low synergy build that doesn't have much standalone capability in either powerset. A good candidate for this kind of test.
|
Originally Posted by Zeh_Masteh
I may just join this experiment by creating (yet) another MM (would've done Stalker, but THB beat me to it) to play, though I already know the results: my MM, whatever build, will level faster and never die. Whether that is because I'm better at MMs, I enjoy them, or there aren't really any enemy groups out there that kill all my pets in under a minute, who knows?
|
Quote:
|
Originally Posted by Arcanaville
Softcapping an Invuln is fantastic. Softcapping a Willpower is amazing. Softcapping SR is kissing your sister.
|
Just to put my 2 cents worth in,If you have the perfect build thats you are very good at playing then yes you have the ultimate build in your opinion,but if you give that build to someone who does not know how to play it well is it still an ultimate build,that person would most likly say OMG this build is pants.My point being that any build can be great providing the person playing it know,s what they are doing and how to play the said power set.For myself I read about how good Kelds can be but have never been able to play them that well,so if asked for my opinion on that set it would be a case of it,s ok I suppose.My first toon was a blaster took her to 50 and still play it now and then.But I,m also into brute,s some AT,s I play are very good some not so good but I put that down to the fact I possibly do not play them right or have the slotting wrong .What the OP is doing is something he wants to prove to himself he is not trying to convince anyone else that blaster,s are great.right that was my 2 cents worth will crawl back into my box till I get up the nerve to come out again 
Prof Radburn controller,Celtic Ice Maiden,blaster,Miss Knockout scrapper,Mistress Davina controller,Stone Hart,tank Split Personality PB.Queen Lostris controller,Fridgid Mary blaster,Shocking Fire blaster Future Elfling defender, Little Weed controller,Capo Angelo MM, Commander Buzzsaw MM, Justice Tank tank all 50,s
Quote:
Translation: You don't think this test matters because the actual experience will vary from person to person?
|
Originally Posted by Prof_Radburn
Just to put my 2 cents worth in,If you have the perfect build thats you are very good at playing then yes you have the ultimate build in your opinion,but if you give that build to someone who does not know how to play it well is it still an ultimate build,that person would most likly say OMG this build is pants.My point being that any build can be great providing the person playing it know,s what they are doing and how to play the said power set.For myself I read about how good Kelds can be but have never been able to play them that well,so if asked for my opinion on that set it would be a case of it,s ok I suppose.My first toon was a blaster took her to 50 and still play it now and then.But I,m also into brute,s some AT,s I play are very good some not so good but I put that down to the fact I possibly do not play them right or have the slotting wrong .What the OP is doing is something he wants to prove to himself he is not trying to convince anyone else that blaster,s are great.right that was my 2 cents worth will crawl back into my box till I get up the nerve to come out again
|
So how about energy/fire blaster and say dark/fire Scrapper? That sound fair?
Currently Playing:
Rage King - SS/Regen Brute (50+3)
Soulfire Darkness - Dark/Fire Tank (50+2)
Deaths Final Embrace - Kat/Dark Brute (50+3)
ULTIMATE REGEN GUIDE I22
Quote:
For those not keeping score at home, here's the estimated time to level for the Blaster and the Stalker for levels 4 through 10 (the data was a little ambiguous for level 3, and it didn't seem that important anyway):|
Originally Posted by TwoHeadedBoy
Ok, a little late, but here are the results for today.. I just finished up. The Stalker got wrecked today. Both characters died once on the way out of the bank after defeating the hero, they just got rushed by Longbow. The Stalker struggled against Goldbrickers though... I knew my endurance couldn't sustain the ranged defense toggle at this point so I went with Caltrops at level 10 (endurance is already rough on the Stalker.) This lead to getting hit with ranged energy (I think it was energy) damage and dying twice. The Blaster was able to immobilize them and debuff their tohit, and that combined with hover meant that they didn't present as much of a challenge to him.
|
Code:
Lvl Bla Sta B dth S dth 4 7 5 0 0 5 14 11 1 0 6 13 15 0 0 7 11 13 0 0 8 27 27 1 0 9 27 26 0 0 10 46 42 1 1 Tot 145 139 3 1
Deaths seem to be costing about one minute of travel per. The two Stalker deaths after level 10 aren't represented yet: they should have a material impact on relative leveling speed for level 11, depending on the survival rate of the blaster.
Working backwards, assuming the average player might not have ninja run, and might zone somewhat slower (a one-ish minute travel time implies fast zoning) lets say the average travel penalty for dying at these levels is closer to two minutes total. How many times would a blaster have to die to make the performance gap at these levels, where there is no debt, equal to about a 20% performance gap, something that would be unambiguously noteworthy? About 14 deaths, or about two per level. Actually slightly less, because at the moment the Blaster is leveling slightly slower even factoring deaths out. Is that an unlikely outcome for average players? Hard to say for certain, but its certainly not a ridiculous level of dying: one bad situation such as that the stalker ran into with the Goldbrickers can generate that sort of thing. Hover blasting with a ranged mitigation set can avoid them (hover + tohit debuff, or hover + knockback, say) but not all blasters will have those tools. Its certainly possible for less experienced players to be in that situation. How often is the question.
Actual debt is going to change things significantly past level 10, and in moderately predictable ways. My guess is that a death will raise in effective penalty from about a minute or so to something closer to 3-4 minutes, at least at these levels, due to debt repayment.
Although patrol XP will complicate things a little. Its important past level 10 that each character is played for about the same amount of time, and always in the same order. The difference is not high, but there's no reason to create a variable skew that doesn't need to exist. Flipping back and forth will create a slight wobble in the amount of patrol XP earned, amounting to a an hour or two which is worth a percent or so of oscillation.
[Guide to Defense] [Scrapper Secondaries Comparison] [Archetype Popularity Analysis]
In one little corner of the universe, there's nothing more irritating than a misfile...
(Please support the best webcomic about a cosmic universal realignment by impaired angelic interference resulting in identity crisis angst. Or I release the pigmy water thieves.)
Quote:
To be more specific, Kepler discovered his laws of planetary motion by in effect data mining them from observations, particularly Tycho Brahe's after he died. It was the accurate data of Tycho that allowed him to determine that planetary motion was elliptical, and not circular. Kepler's biggest indictment of Ptolemy was not just on geocentricity, but on the conclusion that planetary bodies moved in ellipses. This was a serious blow to the notion that heavenly motion was "perfect" and "perfectly architected" by God. Instead, it was imperfect, and not designed to be beautifully arranged.|
Originally Posted by Diellan_
Kepler's Laws of Motions were created in response to Tycho Brahe's observations, which were in direct contradiction to the predictions made by the Ptolemaic system (e.g. that all stars would be close enough to have an observable parallax effect). Making predictions, finding those predictions to be incorrect, then correcting your model (going from geocentric to heliocentric) is exactly the way the scientific process works.
|
What's most important to take away is that while Copernicus was technically correct in designing a heliocentric model that contradicted the geocentric Ptolemaic model, without accurate data his idea of what the simplest and best model would be turned out to be wrong in important ways, and attempts to salvage it with epicycles actually made it worse. Data saved the Copernican model, because without it the model itself had no rudder to steer towards the truth.
[Guide to Defense] [Scrapper Secondaries Comparison] [Archetype Popularity Analysis]
In one little corner of the universe, there's nothing more irritating than a misfile...
(Please support the best webcomic about a cosmic universal realignment by impaired angelic interference resulting in identity crisis angst. Or I release the pigmy water thieves.)
Quote:
If I was going to run a test like this pacing a Blaster against a Scrapper, I would probably pick AR/Energy vs BattleAxe/Invuln. They are probably about as close to average performance in each archetype and comparable to each other in damage types as you could get. Its a bit too much AoE on the blaster side, but the better equivalent from that perspective and damage mitigation perspective would probably be radiation blast, but that's a completely different damage type with no easy comparable melee side set. Electric melee has drain which complicates the comparison: that drain is almost useless against minions and Lts which would dominate the early parts of the test.|
Originally Posted by planet_J
So how about energy/fire blaster and say dark/fire Scrapper? That sound fair?
|
I don't have the time to dedicate to this type of testing at the moment, but that's probably the combination I would use if I did. No combination would be perfect, though, so a better test would probably pick a few combinations to test against each other collectively.
[Guide to Defense] [Scrapper Secondaries Comparison] [Archetype Popularity Analysis]
In one little corner of the universe, there's nothing more irritating than a misfile...
(Please support the best webcomic about a cosmic universal realignment by impaired angelic interference resulting in identity crisis angst. Or I release the pigmy water thieves.)
Just finished up for today, and the Stalker was back with a vengeance. There are a few things I noticed that I find are worth mentioning.
I noticed that for some reason, I'm terrible at using AS while in hide. For some reason, about every third time I go to use it I'll que it in perfect position, go into the little AS stance, but then the power just won't work and I'll still have to pay the endurance cost. I'm not sure if this is WAI and I'm just bad at Stalking or if it's buggy.
The second thing I noticed was that I seriously need to set up binds on the Blaster to switch between Hover and Ninja Run before I start tomorrow. It's incredibly tedious not having one, and one doesn't detoggle the other like with CJ and NR so I have to manually hit two buttons and I feel like it slows me down (not to mention drives me crazy) every time I need to move faster.
Anyways here's the info for today, and thanks Arcana. The debt was totally erased on the Stalker via patrol XP so they were really on pretty even ground. I'll make a point to always play the Blaster first from this point forward (yesterday was the only day I didn't.)
DP/Dark Blaster
Total levels: 1
Total defeats: 2
Mr. Bocor
3:36 Defeat 10 Hellions (0 defeats)
3:41 Defeat all Hellions in hideout (0 defeats)
3:55 Defeat all foes in caves (0 defeats)
4:02 lvl 11
4:10 Mr. Bocor arc complete
Dr. Graves pt. 2
4:11 Survive Zephyr's plan (1 defeats)
4:21 defeated (hospital, 4:23 returned)
4:31 Talk to Dollface
4:33 Defeat 3 Arachnos agents
4:35 defeated (Jumped by a bunch of family toolbags while Arachnos hunting)
4:35 Complete
Elec/Ninj Stalker
Total levels: 1
Total defeats: 1
Mr. Bocor (0 defeats)
4:55 Defeat 10 Hellions
4:59 Defeat all Hellions in hideout
5:10 Defeat all foes in caves
5:16 lvl 11
5:21 Mr. Bocor arc complete
Dr. Graves pt. 2
5:22 Survive Zephyr's plan (1 defeats)
5:28 defeated (hospital,5:31 returned)
5:37 Talk to Dollface
5:38 Defeat 3 Arachnos agents (0 defeats)
5:40 Go to the University in Cap au Diable
5:41 Defeat 10 Luddites (0 defeats)
5:44 complete
Level 11 seems to have been about a tie (44 minutes Blaster 43 minutes Stalker) but the data suggests that the Stalker for some reason accelerated significantly heading towards 12. Won't know for sure until both hit 12.
[Guide to Defense] [Scrapper Secondaries Comparison] [Archetype Popularity Analysis]
In one little corner of the universe, there's nothing more irritating than a misfile...
(Please support the best webcomic about a cosmic universal realignment by impaired angelic interference resulting in identity crisis angst. Or I release the pigmy water thieves.)
Quote:
It's lag; if you hit the AS button before the server registers you've stopped moving, you'll interrupt yourself and still pay the endurance cost.|
Originally Posted by TwoHeadedBoy
I noticed that for some reason, I'm terrible at using AS while in hide. For some reason, about every third time I go to use it I'll que it in perfect position, go into the little AS stance, but then the power just won't work and I'll still have to pay the endurance cost. I'm not sure if this is WAI and I'm just bad at Stalking or if it's buggy.
|
I do it all the time too.
Just starting with the most relevant points first
(MedianTime / MPM) * TaskModifier * TimesRunModifier * TimeModifier + ArtificialModifiers
http://paragonwiki.com/wiki/Merit_Rewards
In other words they took a couple of statistics for each TF and allowed it to set the run times.
Now if you go about things by modeling blaster performance you are in a much better situation. In terms of the game I don't have to experiment to validate my model the way I would in the real world. Why ? Because barring bugs I know what the maximum damage output can be, I Know how many hitpoints they can have and I know how rapidly and under what conditions they can recover them. For years I have been saying these come up short relative to the other ATs. You don't need to sample the data of player base to know that blasters will underperform. It is written on their very D.N.A.. What looking at the data of how a particular set of players is doing with them in the context of the wider game just processes their inherent strengths and weaknesses through a series of funhouse mirrors.
Anyway on to the history of science portion.
Regarding Test Driven Development, we used to call that trial and error. It may work but its a sure sign that the people using it just don't have a better method available.