Upgrades don't apply in some places (Q is it design)

Discussions about AI types, units, tactics & strategy.
Post Reply
nick87720z
Trained
Trained
Posts: 71
Joined: 25 Feb 2014, 16:37

Upgrades don't apply in some places (Q is it design)

Post by nick87720z » 17 Oct 2019, 09:38

I recently looked to betaguide after first tries of skirmish with T4 startup. Some things look strange to me. I stuck at two things - armor upgrades for transport and repair upgrades (despite order, this is more primary for me).

For begning, I assumed by essence that all repair techs get all repair upgrades - all mobile repair turrets, facility, vtol pad and auto-repair. But according to betaguide - repair facility upgrades also apply to vtol pads, but don't to mobile turrets - is it intentional?
I also noticed big difference in game, with situation mentioned above (recent T4 no-base attempt). I always assumed, that basic repair turret, landing pad and auto-repair have same repair speed (lol). Of course, it is perceptive. It was prety hard to find stats about repair turret power (there is no repair "weapons" type in section, where all info about available upgrades can be seen for weapons, structures, bodies, etc). Instead I found it only in database part, where only own stats are showed in table. According to it, repair facility has repair points 40 - heavy repair turrent is only by 1/4 weaker, thus looking like strong addon to sparce ground installation (to not get stranded after chasing retreating units).

Before looking to betaguide, I tried to measure repair speed diff between ground facility, heavy mobile turret and auto-repair. For test I used dragon with two gauss cannons (HP - 10360 or so), with same one lowering HP to 500 or so, before sending target to repair (I used max possible HP to minimize randomness in HP when target is contacted by repair). With this (hopely negligible) error I got following result solving equation in qalculate: auto-repair is ~2.2x slower than mobile turret, and mobile turret power is ~3.6x of facility. Full difference (multiplying both) between AR and facility is 7.5. Would be great to look exactly possible upgrades for mobile turrets and auto-repair. Btw, according to betaguide database, AR is even more powerful than facility (2.5x more) - with upgrades it should make all other repair means obsolete :?

Transport armor upgrades. I initially though it is intentionally made to be weaker than everything else, but HP 500 seem to be basic stat, and it is still higher than for drarong body (350). But when I tested heavy transport and rouble-railgun dragon vtol durability by shooting with vindicator+stormbringer combo unit, dragon vtol survived unless vindicator second salvo, while transport failed in 2-3 seconds (or should I better try with single least armored weapon to keep close to pure body HP?)

Edit:
I made more careful comparison for durability between dragon and transport, this time making vtol with single lancer, so that it doesn't differs by HP from weaponless body (~2600 HP). Weapons - single vindicator, than whirlwind. For dragon it took 6 missiles or 7 sec whirlwind, while for heavy transport it is only 2 missiles (1st taking more than half) and 2 sec whirlwind (did not cound shots for such rapid fire weapon, but could be less for transport if some last shots did not hang spinning around transport hit center, as it happens sometimes with homing projectiles in game).
Last edited by nick87720z on 18 Oct 2019, 09:54, edited 1 time in total.

MIH-XTC
Trained
Trained
Posts: 358
Joined: 31 Jan 2014, 07:06

Re: Upgrades don't apply in some place (Q is it design)

Post by MIH-XTC » 17 Oct 2019, 23:35

Your observations are correct.

In research.json the armor upgrades only apply to “Droids” and transporters are classified as “Transports”. See “filterValue” in the armor upgrade below.
Spoiler:
The “filterValue” refers to the “class” attribute in body.json. Dragon body in body.json has class “Droids” but transporters have class “Transports”. You can try to add another upgrade to the armor upgrades in research.json with "filterValue": "Transports" to see if it’ll upgrade transporters too.

You’re also right about the repair upgrades. In research.json the repair upgrades only upgrade class “Building”.
Spoiler:
You can try adding another upgrade of class “Repair” to see if it upgrades repair droids. I tried adding upgrades for the repair droids once before but now that I’m looking at it again I think I used the wrong class value.

nick87720z
Trained
Trained
Posts: 71
Joined: 25 Feb 2014, 16:37

Re: Upgrades don't apply in some place (Q is it design)

Post by nick87720z » 18 Oct 2019, 06:36

Thanks :)

Update: If it is not possible to specify multiple classes for some upgrade - is it ok adding duplicate group in research.json, differing only by target class? Would not it be more correct (and possible) to make Transport class itself derived from Droid, so that it gets all other upgrades besides kinetick armor (thermal, engine, whatever)?

(looks like this thread is more Development - related than about gameplay)

Update 1: I made experiment with Heavy transport, setting droidType from Transports to Droids. It is not only one property pointing to its transport nature, and at least it did not lose its transport functionality. Units still may be loaded/unloaded, and there is no orders menu (like patrol), but now it has armor upgrades. It now takes ~11,3 whirlwind or 8 vindicator missiles.

nick87720z
Trained
Trained
Posts: 71
Joined: 25 Feb 2014, 16:37

Re: Upgrades don't apply in some place (Q is it design)

Post by nick87720z » 18 Oct 2019, 09:53

Tried to add two more blocks (for "Droids" and "Cyborgs") into "results" section for all repair upgrades in research.json. Although it did not work, there is another strange thing I noticed. If I noticed it right, that repairPoints means points per second, then I got familiar values:
~~ 29.987079, 99.907435, 13.330998 - for heavy turret, facility and auto-repair respectively.

it is clearly seen, that heavy turret is at basic performance, and for facility - 40*(1+0.5*3) => 100 ...
however, same value (100) is expected to be for auto-repair! I.e., when available, it should make facility useless, not even mentioning turrets (may be only to strengthen units, taking concentrated fire, as e.g. case for mobile tracked arty, acquired by CB).
Yet it is on one level with basic repair turret, cyborg engineer and vtol repair.

MIH-XTC
Trained
Trained
Posts: 358
Joined: 31 Jan 2014, 07:06

Re: Upgrades don't apply in some place (Q is it design)

Post by MIH-XTC » 20 Oct 2019, 15:54

nick87720z wrote:
18 Oct 2019, 09:53
Tried to add two more blocks (for "Droids" and "Cyborgs") into "results" section for all repair upgrades in research.json. Although it did not work, there is another strange thing I noticed. If I noticed it right, that repairPoints means points per second, then I got familiar values:
~~ 29.987079, 99.907435, 13.330998 - for heavy turret, facility and auto-repair respectively.

it is clearly seen, that heavy turret is at basic performance, and for facility - 40*(1+0.5*3) => 100 ...
however, same value (100) is expected to be for auto-repair! I.e., when available, it should make facility useless, not even mentioning turrets (may be only to strengthen units, taking concentrated fire, as e.g. case for mobile tracked arty, acquired by CB).
Yet it is on one level with basic repair turret, cyborg engineer and vtol repair.
You're not taking into account the "time" attribute in repair.json. I think it specifies the "repairPoints" over a time interval and auto-repair has a value of "75" whereas the turrets are only "7".

It's a lot easier to view the stats in tabular form (a table). You can copy and paste the textual .json file into this website https://json-csv.com/ and it'll export/download you a copy in a table format.

Here's a screenshot of the original repair.json from a stats editor I made.

Image

jsoncsv
New user
Posts: 1
Joined: 20 Oct 2019, 16:08

Re: Upgrades don't apply in some places (Q is it design)

Post by jsoncsv » 20 Oct 2019, 16:19

Thanks for mentioning (and using) json-csv.com - I am the guy who created the tool.

Just adding some extra info - the converter will show all the nested data in a kind of header / detail report format. Once in a spreadsheet you can actually amend any of the data (add new nested data etc.) You can then convert it back to JSON again using https://json-csv.com/reverse (and import it back into Warzone 2100 again?).

So it's effectively an easy way to edit JSON data (and hence WZ2100 data I guess).

Anyway, if you guys need some customization (or different output format etc), just contact me through the site and I'll be happy to help out.

MIH-XTC
Trained
Trained
Posts: 358
Joined: 31 Jan 2014, 07:06

Re: Upgrades don't apply in some places (Q is it design)

Post by MIH-XTC » 21 Oct 2019, 00:03

jsoncsv wrote:
20 Oct 2019, 16:19
Thanks for mentioning (and using) json-csv.com - I am the guy who created the tool.
Wow that’s very cool of you to show up here on the forums. I used your site to manipulate all of the warzone2100 stat files in Excel years ago. I just recently made reference to it in this video We’re specifically talking about the set of files here. Those are all of the moddable aspects of the game stored in .json format with research.json being the largest at 9,200 lines. https://json-csv.com/ was tremendously helpful in being able to analyze the warzone2100 stat files.
jsoncsv wrote:
20 Oct 2019, 16:19
You can then convert it back to JSON again using https://json-csv.com/reverse (and import it back into Warzone 2100 again?).

So it's effectively an easy way to edit JSON data (and hence WZ2100 data I guess).

Anyway, if you guys need some customization (or different output format etc), just contact me through the site and I'll be happy to help out.
I wish I knew there was export functionality too and not just import. I spent 2 months making a .json serializer in VBA to do the export aspect and even then the code doesn’t work for some edge cases.

Well, there you have it folks, a method of importing AND exporting the .json files back and forth into tabular form. Making adjustments in tables is way easier than a text editor.

On a site note, I've always wondered what you're using on the back end for that site. Are you by chance using Pandas?

nick87720z
Trained
Trained
Posts: 71
Joined: 25 Feb 2014, 16:37

Re: Upgrades don't apply in some places (Q is it design)

Post by nick87720z » 21 Oct 2019, 05:50

You're not taking into account the "time" attribute in repair.json. I think it specifies the "repairPoints" over a time interval and auto-repair has a value of "75" whereas the turrets are only "7".
Thanks, I nearly thought about time, but simply rejected possibility of different time interval for different repair systems. Though probably, it is notable during game - while from turrets beam is almost continuous, self-repair flash cloud may be noticed between relatively big periods of time (I initially assumed it is due to target-side cloud randomly shifting). However, I tried to recalculate expected performance for auto-repair: 100*7/75.. == only 9.(3), which is even less than for cyborg engineer (against 13.33 from equation).

Meanhile I tried to make transports armor upgradeable in proposed way - modifying research.json instead of body.json. It works, but somewhat harder, comparing to simple change of transport body class. Probably it is more correct - I remember, how vtols and transports twitched into and out of air after all upgrades research :D (probably due to engine upgrades, applying with multiplication instead of summing).

Post Reply