Verified:

Z [Post Script]

Member

24

Aug 16th 2019, 4:53:26

I can’t post on the UI forum for some reason (which is maybe a bug report itself). So will do here.

It generally feels more modern which is great!

I do think most of the changes so far are technical rather than experiential. For example, I usually want to tech a certain number of turns instead of a certain number of points, say 25 bus and 25 res. I’m a math person, but doing the multiplication of turns and tpt on a phone and making sure I type it correctly just creates unnecessary friction.

That said, sometimes I might actually want points, e.g. getting to 120% strat before an attack. But then I have to go to the wiki, create an Excel formula, figure out how many points I need, tech those points and realize I did something wrong, go back to Excel to find a typo, recalculate how many more points I need, and then go tech those. An alternative would be an option to tech to a certain %, telling you how many turns that would take, and then letting you confirm that - same thought process for buying tech.

Or for attacking, an alternative could be an in-game version of country search based on various filters including DRs, a tap to attack (or spy if you really want to), pre-filling attacking forces based on your op or on parameters you pre-set (e.g. maybe you want to assume 2x allies and 20% oversend, or maybe 5 jets for every $1 of networth), and maybe suggesting and then buying any additional military needed. (I think the preference that already exists for public vs private goods is a simple but good example of the type of pre-set that I’m talking about that approaches the coding from the angle of what the user is trying to do.)

Some of what I suggest blurs the line between putting in the work (to understand strategy and to make sure you pay attention to details) and reducing friction on a phone jumping to other things like calculator - or of stupid mistakes like bad mental math or forgetting someone is a dict. And we can debate how far we go down that path. But I do think it’s be good to include in the UI rebuild what users might actually be trying to do (e.g. tech turns vs points), looking for natural connections across actions (e.g. ops into war room), and bringing in house at least the most basic functions that happens on third party applications or sites. I can provide more specifics if/when I can post on the UI forum :)

Edited By: Z [Post Script] on Aug 16th 2019, 5:06:47
-Z (Post Script)

Bug

Developer

475

Aug 16th 2019, 5:29:05

Z, I think you may need to be a "Member" to post.

I will get pang to look into it.

I think for Tech the plan is to add more details to the "help" section like max tech's and maybe even a tech calculator.

As for using ingame to search and then or input details into grabs/hits etc. i know the wider community will not like that. its a calculation factor that everyone must do themselves. and I am not really ok doing that calculation for you ingame.

But for now i will get someone to look at your access to the forum.

Z [Post Script]

Member

24

Aug 16th 2019, 12:57:59

Looks like I can post with my old LCN alt - so must be the member thing. Thanks!

I’d just ask us to collectively push on things that could reduce friction, make things faster/easier, and keep people in the game - especially when there is no advantage lost (e.g. my tech to x% idea). I’ll post more thoughts in UI.
-Z (Post Script)

Gerdler

Member

2168

Aug 16th 2019, 14:29:59

I dont think the scores list should list DR simply because that is an unknowable unknown right now(while you can calculate a close approximation). I rather have spy ops provide the DR numbers so an investment has to be made to aquire them.

Pang

Administrator
Game Development
5445

Aug 16th 2019, 15:22:40

computing DR for every country in the scorelist is very expensive computationally. no chance we'll be doing that :p

spyop is more reasonable

I updated the forum too so you can pepper the other thoughts on there
-=Pang=-
Earth Empires Staff
pangaea [at] earthempires [dot] com

Boxcar - Earth Empires Clan & Alliance Hosting
http://www.boxcarhosting.com

seangcxq

Member

88

Aug 20th 2019, 4:07:25

Originally posted by Pang:
computing DR for every country in the scorelist is very expensive computationally. no chance we'll be doing that :p

spyop is more reasonable

I updated the forum too so you can pepper the other thoughts on there


Would be odd to spend turns to attain DR info when it can be calculated for 0 turns just by doing some math. What is the actual DR formula used that is so computationally expensive? I have a very close approximation that takes barely any resources.

Gerdler

Member

2168

Aug 20th 2019, 4:43:56

You can approximate it very easily based on a few assumptions, assumptions that are usually wrong(but hopefully close), and thus the approximation is not the exact answer.

The math is done for every landgrab you make for one country, and that is fine, but if the server has to do it for up to 61 countries each time you press the link to the scores list it starts to become computationally heavy. Or when it updates the database of every country on every server so that API's can grab it. It would be something like 3000-3500 countries to be calculated at once and the DR of every hit of the last 72 hours on each of those 3000+ countries has to be calculated. Thats probably 15-30k hits normally (when there are no big wars going on). And then we have not even gotten into the C:C DR, which would have to be calculated for every hit all reset on every country pair separatly; that could be another 100-200k hits. Each of those hits will have to be recalculated at least every 3 or 5 minutes (however often the scores database is updated on the server) as they each will have changed through the DR formula.

Edited By: Gerdler on Aug 20th 2019, 4:49:41

seangcxq

Member

88

Aug 24th 2019, 13:31:52

Originally posted by Gerdler:
You can approximate it very easily based on a few assumptions, assumptions that are usually wrong(but hopefully close), and thus the approximation is not the exact answer.

The math is done for every landgrab you make for one country, and that is fine, but if the server has to do it for up to 61 countries each time you press the link to the scores list it starts to become computationally heavy. Or when it updates the database of every country on every server so that API's can grab it. It would be something like 3000-3500 countries to be calculated at once and the DR of every hit of the last 72 hours on each of those 3000+ countries has to be calculated. Thats probably 15-30k hits normally (when there are no big wars going on). And then we have not even gotten into the C:C DR, which would have to be calculated for every hit all reset on every country pair separatly; that could be another 100-200k hits. Each of those hits will have to be recalculated at least every 3 or 5 minutes (however often the scores database is updated on the server) as they each will have changed through the DR formula.


So it requires recursion?

Even so, it doesn't need to be fully computed from the beginning every 3-5 minutes. Previous scratch can be cached and each update would only build upon the cache. Difficult to discuss in words, but I don't see it as unrealistic. That being said I am not familiar with the actual hardware limits here at EE.

Gerdler

Member

2168

Aug 29th 2019, 4:01:25

The DR changes all the time so it has to be recalculated.

Each hit will have to be recalculated seperatly because the formula calls for different half-lives on all hits.

AtticusRex

Member

104

Aug 29th 2019, 4:15:21

I havent tried to understand that formula, but doesnt your statement that it's unknowable suggest it's too complicated?

A big appeal of this game has always been the ability to calculate factors/outcomes. Things like DR, and even pm regen, have seemed overly complicated.

Maybe I missed whatever development initially led to such complication, but is it necessary?

Gerdler

Member

2168

Aug 29th 2019, 6:04:04

No, it is unknowable in almost all circumstances, but you can make very good guesses like I said and how well we guess makes the better grabber. It also allows, as I suggested before, there to be a spy op that supplies you with another important piece of information about a country at the cost of the investment of spying and having spies.

I really like the way DR stacks up and is calculated right now, but I don't really agree with the formula for how DR acts on grabbing gains. I think the effect on how the returns diminish should diminish more smoothly after like 6-8 DR to 13 DR and that the minimum should be higher set than 10% returns.

Zorp

Member

838

Sep 6th 2019, 0:50:07

DR is 100% accurately knowable server side AFAIK, but not client side because the server doesn't make that information available to the client.

Edited By: Zorp on Sep 10th 2019, 15:15:59

Gerdler

Member

2168

Sep 11th 2019, 11:06:05

Yeah true. Thats why we could have a spy op that gives that information.