4 replies to this topic

[Phaesphoros]

To respond to Koniving, who posted some suggestions for DHS, I went on testing grounds and in online matches to find out how big the heat threshold is and how it can be calculated.

This is not a rant and not much of a complaint/suggestion, but rather it turned out that the heat meter behaves objectively not in a consistent way, regarding grouped weapon fire, heat build-up and precision.

Test setup

AS7-RS, 4 ERPPCs, STD325 engine, 11 external DHS, mastered (2x elite heat efficiencies).


In all tests conducted my 'Mech was standing still if not explicitly mentioned otherwise.

The high heat readings were done using the highest displayed heat before dissipation, several tests were conducted to get a solid approximation of this maximum. The testing ground results show that this method is ok.

On testing grounds

(where efficiencies don't work)

Dropped on Frozen City. First alpha: 57 % heat, second alpha: shutdown. An ERPPC currently produces 11 heat, therefore this test suggests the heat threshold is (w/o elite skills) at about 77.
Read on why this test was crap.

Then I stripped all external heat sinks and dropped again.

Dropped on Caustic. Two (2 only) ERPPCs raised heat from 6% to 49%.
I already guessed there might be issues with the environmental temperature, therefore I dropped again. Little did I know...

Dropped on Alpine. First 2 ERPPC salvo raised heat from 0% to 33%. That didn't match with the numbers from Caustic, so I waited for the heat to return to 0 % and the moment it hit 0% I fired again (2 ERPPCs). Heat went up to 43%. (Several tests show the same number.)
Then let the heat return to 0% and wait some longer. Two ERPPCs now raise the heat to 33%, which suggests there is a NEGATIVE HEAT % that is not displayed.

Redid the test with 11 external heat sinks, same methodology as above.
Result is: 33% heat increase from 2 ERPPCs.

Absolute heat threshold

This suggests a heat threshold of 51 w/o external DHS, and 67 w/ 11 external DHS. With these two data points and assuming a constant increase of the heat threshold for every external DHS added, this suggests an increase of the heat threshold of 1.45 per external DHS. The absolute heat threshold w/o external heat sinks matches the formula 30 basic + 10 internal * 2.0 within a slight error margin.

Online matches

In online matches, the heat indicator seems to be highly inaccurate. I have mastered the AS7-RS, therefore I have the elite heat efficiencies.

Dropped on Canyon.
When firing 2 ERPPCs, the heat goes from 0 % (with the methodology described above) to something between 20 % and 24 %, sometimes spiking to 28 % or higher for a fraction of a second. Similar things happen when starting to walk, where the heat is at 2 % from walking.

All 4 ERPPCs raise the heat from 2 % to something between 52 % and 54 %, sometimes only 50 %, sometimes with a 1 second buildup of 2-4 % after the heat has already been increased by the shots. In every case, the dissipation doesn't start immediately but with a 1-2 s delay, unlike on testing grounds. My ping is about 120-130 average.

This suggests a heat threshold between 92-110 (2 ERPPC tests) or 81-85 (4 ERPPC tests). As I can definitely NOT fire two salvos of 4 ERPPCs, the threshold must be lower than 88.


I removed external heat sinks, swapped out the STD325 for a STD300 and put in a gauss with 3 tons of ammo -- lol MORE CHEESE.
Dropped on Frozen City, same methodology.
2 ERPPCs: 0 % -> 31 % heat
4 ERPPCs: 0 % -> 64 % heat
(with the same weird effects like spiking and heat buildup)

This suggests a heat threshold of something between 71 and 69.

Absolute heat threshold

If we assume that the 2-ERPPC-tests w/ 11 external DHS are somehow buggy, and use an average of about 83 for this case, subtract a heat threshold of 70 for 10 internal DHS, we end up with 1.18 per external DHS.



If we assume that the 4-ERPPC-tests w/ 11 external DHS are somehow buggy, and use an average of about 100 for this case, subtract a heat threshold of 70 for 10 internal DHS, we end up with 2.72 per external DHS.

Let's use the formula (30 basic + 10 internal * 2.0 + 11 external * 1.4) * 1.2 elite efficiency then we end up with a heat threshold of 60 (no external DHS) or 78 (11 external DHS), respectively. In this formula an external DHS increases the threshold by 1.68 (= 1.4*1.2).

Conclusion

My ping might be an issue that could explain heat spikes of a fraction of a second (although the client could know that I have unlocked efficiencies, this still is a server-authoritative game) but not the 1-2 s heat buildup / delayed dissipation.



The other thing that stands out is the discrepancy between the 2-ERPPC and the 4-ERPPC tests with external DHS. It might be the case that the heat is higher than displayed, and the delayed dissipation compensates that (during the delay the heat would drop to the displayed value).

No matter what the reason is for these effects, the % heat display is confusing; it also is inconsistent: it IS inconsistent between online and testing grounds, it is inconsistent between shots in online matches, it might be inconsistent regarding the 2/4 ERPPC tests.

As there seem to be negative heat percentages, the heat display is utterly useless for low heat on cold maps. I suppose 0% means that the 'Mech's temperature is lower than some fixed temperature (say 0°C) that can be higher than the environmental temperature.

A more useful 0% value would be the environmental temperature itself, as this immediately removes the negative value problem. On hot maps, it wouldn't be obvious from this display that the temperature is high, but the indicator would be more focused to what is done by the 'Mech and what is under the pilot's control. Disadvantage: Increases (e.g. firing a weapon) and decreases (e.g. dissipation, cool shot) have different magnitudes on different maps (different temperature).

Another possibility is to display a temperature-based value, this would have the advantage that increases and decreases should have the same magnitude on all maps. The problem is of course the 0 value; but one could simply add negative values (possibly w/o bar) like in the Celsius scale (as opposed to the Kelvin scale).

Call for support from other players

Please try it yourself and post your observations including your ping. It would be nice to have some videos as well; my FPS is utterly low and screen capturing doesn't produce useful results in online matches (it's like a slide show and cannot display any spikes).


I'd appreciate any further information or observation.

Edited by Phaesphoros, 01 July 2013 - 06:10 PM.

[Koniving]

Phaesphoros:

Some of this, mainly the difference between training ground and an actual match, and difference between one environment and another is expected.

Frozen City is extremely cold which accelerates your cooling rate. Caustic is hot. Tourmaline is in excess of 200 degrees fahrenheit. Walking generates heat as per tabletop. Jumpjets do as well.

Cooling always accelerates when you power off as the base heat you generate by simply being powered is removed.

Go with the first alpha strike and see if you can replicate it, when you get a consistent number compare that ratio and percentage here.

For example if I fired 66 heat and it spikes to 57%, that suggests the heat threshold is as follows:
66 : 57% = 115.8 : 100%.
That actually sounds awfully high, as my tests on forest colony suggest a threshold in the range of 95 give or take a few while my equations come out with 92.4 for 22 "1.4" heatsinks, but "110.4" for 10 doubles (internal 6 [standard being 3]) and 12 externals (1.4*3 = 4.2 per external 1.4 sink).

It's impossible to determine thresholds in extreme temperatures as they actually add the heat or cooling of the environment.

The best place to test heat for threshold is in Forest Colony (not the snow one) as that's where PGI originally did their heat calculations and testing in terms of temperature. It's the "truest" threshold you can get based on testing alone, preferably in testing ground as the "threshold by 10%" increase times 2 through basic/elite efficiencies only throw the estimates off by that much more.

----------------------------------

I would be interested in official statements on heat threshold and the rates at which they rise as I'm very interested in "why" thresholds raise rather than being purely dissipation rates. For example and these numbers are purely simplified babble: If the standard heatsink cools by 1 and the double heatsink cools by 2 but both have a heat threshold of x amount regardless of the heatsinks, would that not be enough?

Instead we have standard heatsink cools by 1, but raises by 3 (again simplified numbers to demonstrate how absurd the concept feels) while the double heatsink cools by 1.4 and raises by 7, and before you know it we need 30 different charts for every map temperature, every weapon, every heatsink arrangement, whether it's in the engine or placed outside, whether it's this or that.

Why is it not just cools by 1 or by 2? That is to say dissipation (heat sinking) only without threshold changes? But that's just me.

Edited by Koniving, 27 June 2013 - 05:31 PM.

[Phaesphoros]

Koniving, on 27 June 2013 - 04:59 PM, said:

Some of this, mainly the difference between training ground and an actual match, and difference between one environment and another is expected.

Frozen City is extremely cold which accelerates your cooling rate. Caustic is hot. Tourmaline is in excess of 200 degrees fahrenheit. Walking generates heat as per tabletop. Jumpjets do as well.

Cooling always accelerates when you power off as the base heat you generate by simply being powered is removed.

I'm aware of all of this (I think):

• Difference between testing grounds and online matches should be the efficiencies (heat threshold + 20%, cooling efficiency), maybe ping, and things not yet implemented in testing grounds. Environmental temperature probably isn't one of that, because on testing grounds' Caustic, I got 6% heat standing still.
  
• Difference between environments should affect heat dissipation rate, which doesn't matter for the tests I conducted: I raised the heat from a known value and read the maximum from the heat display. This isn't affected by the dissipation rate (or only slightly and shouldn't matter). Hot maps seem to indicate that the heat threshold isn't altered by the environment: If we have a non-0% standby heat, the scale most probably isn't adjusted at the lower end, and it's very improbable it's adjusted at the higher end. It would also be rather strange, as the heat probably indicates the 'Mech's temperature, which physically should have a fixed maximum before systems begin to fail and take damage.
  
• Walking and JJ generate heat, therefore I stood still for most of the tests. In all tests I subtracted the heat % before shooting from the heat % after shooting.
  
• I didn't use any numbers from tests where my 'Mech shut down. Therefore, the increased cooling rate during power-off shouldn't be a factor in my observations.

Koniving, on 27 June 2013 - 04:59 PM, said:

Go with the first alpha strike and see if you can replicate it, when you get a consistent number compare that ratio and percentage here.

I already used that formula to estimate the heat threshold values in the first post. The first alpha strike isn't reliable, as cold maps show that there are sub-0% heat values. Therefore the method of shooting something to get positive, displayed heat percentages, letting it cool down to 0% and in the exact moment it hit 0%, I did the alpha/group firing. This should be accurate within 1%.

Koniving, on 27 June 2013 - 04:59 PM, said:

It's impossible to determine thresholds in extreme temperatures as they actually add the heat or cooling of the environment.

The testing ground results seem to refute this (tests on Caustic suggest the same heat threshold than tests on Frozen City); in online matches I cannot exactly tell because, as this is the actual intent of this thread, the heat meter is utterly unreliable and inconsistent in online matches. I currently do not care as much as what the actual heat threshold is, but rather why there are these issues with the heat meter.

Koniving, on 27 June 2013 - 04:59 PM, said:

[...]
Why is it not just cools by 1 or by 2? That is to say dissipation (heat sinking) only without threshold changes? But that's just me.

I'd appreciate if we could keep suggestions for the heat threshold / heat sinks in the proper sub-forums (like game balance, feature suggestions, ..) and use this thread for observations and feedback on the heat display in the HUD.

[HRR Mary]

Bumping that thread in hopes of some Dev clarification...

[Tarl Cabot]

Actual Heat Scale And Cap - Tested Via Testing Grounds?


I had also done some tests and made corrections as they came through. I used both Standing and Running stances, as well as the caldera and outside the caldera on the Caustic map.

For mine, it is to find the actual heat cap based on the number of SHS, with some tests with DHS in the engine.

As for any radical jump in the heat scale, I contributed that to packet corruption since it was inconsistent and irregular, nor is it the first time a UI piece did not update properly, as I have seen it in other games.

Edited by Tarl Cabot, 01 July 2013 - 07:14 AM.