[1.04b1] Traffic Signs Numbers issue! (Full Version)

All Forums >> [New Releases from Matrix Games] >> Shadow Empire >> Tech Support



Message


GodwinW -> [1.04b1] Traffic Signs Numbers issue! (7/2/2020 11:20:43 AM)

I've played whole games never bothering to actually do the calculations and just working with the amounts without a hitch.

However, I am now going to take you by the hand along to a journey into the expected values of using Traffic Signs, and you'll see the issue! [:'(]

[image]https://i.imgur.com/wZbEUqN.png[/image]

So, 400 points, and this tooltip can be read 2 ways:

A) Stop 20% of (the LP leaving the Hex in this direction)

The LP leaving the Hex in this direction = 200.
20% of 200 is 40.
Stop 40 LP of the 200 LP going this direction.
So I expect 160 LP to continue and 240 LP now going the 1 other direction.

B) Stop 20% of (the LP) from leaving the Hex in this direction

The LP = 400.
20% of 400 is 80.
Stop 80 LP of the 400 LP from going in this direction.
So I expect 120 LP to continue and 280 LP now going the 1 other direction.

Now a reality check:

[image]https://i.imgur.com/9PffvrU.png[/image]

177? That wasn't an option!

It's closest to interpretation A though. Which to me also makes the most sense anyway from the understanding I have of the system.

And it's suspiciously close to 160 + 16 + 1.6 ( = 177.6).

So is this an indication of a recursive function calculating these?

Let's try another number:

[image]https://i.imgur.com/X1Of8F5.png[/image]

We'd now expect 80% of 200 = 160 to be blocked from the 200 going this direction (interpretation A).

So 40 goes through. And then somehow 40 + 4 + 0.4 = 44.4 to actually be the amount that gets through.

Let's check:

[image]https://i.imgur.com/VvL0ba3.png[/image]

66??

... that's not good! [:D]

Interestingly, it's now closer to interpretation B where 80% of 400 is 320 and so blocking 320 of the 200 (impossible) means blocking it all, leaving 120 leftover that gets added to the free branch making that branch, NW, take 120 more points.

120 + 12 + 1.2 plus the original 200 is indeed the 333.2 points that are leaving in that direction.
Then, to avoid obvious bugs, send the leftover down to the SW (66 points (rounding)).

It's weird!




Malevolence -> RE: [1.04b1] Traffic Signs Numbers issue! (7/2/2020 2:33:26 PM)

How did you get all those images in one post?!?!? [&:]

You didn't use the upload. You used an image site. [8|]

I should have realized.




zgrssd -> RE: [1.04b1] Traffic Signs Numbers issue! (7/2/2020 2:43:02 PM)

I always read it as "stop 20% of what would normally go out that direction".
Normally 200 would go out that direction (as there are 2 roads leading out of the hex and 400 comming into the hex).

So 20% of that is blocked and redirected. So 20% would be redirected. From that, it is only a small math error of 22 vs 20.




GodwinW -> RE: [1.04b1] Traffic Signs Numbers issue! (7/2/2020 2:59:15 PM)

quote:

ORIGINAL: zgrssd

I always read it as "stop 20% of what would normally go out that direction".
Normally 200 would go out that direction (as there are 2 roads leading out of the hex and 400 comming into the hex).

So 20% of that is blocked and redirected. So 20% would be redirected. From that, it is only a small math error of 22 vs 20.


Eh no, 20% of 200 is 40, not 20 :p

So 40 should be redirected to give 240 going to NW, not the 222 of the image.

And that was what I wrote under A).




Vic -> RE: [1.04b1] Traffic Signs Numbers issue! (7/3/2020 12:50:06 PM)

It a weighting system so 20% block means weights 80 and 100.
Resulting in a 177 and 223 split

best,
Vic




GodwinW -> RE: [1.04b1] Traffic Signs Numbers issue! (7/3/2020 3:07:26 PM)

Uhhh. That's not clear to me [:D]
I don't know what you mean.
'Weighting system'?

I spent some time thinking about it, but I cannot arrive at a value of 177. 20% of 200 (one side) is 40. So 160. 20% of the whole 400 is 320.

Closest I can get is that it means that 20% of the LIS normally allocated to the 20% blocked fork is mandatorily assigned to the other fork, so we now have 40 LIS going down the other path, and there's 360 left which is split evenly over the two forks.

That means each path gets an additional 180 points. So the blocked path gets 180 points and the non-blocked path 220. But that's still a few digits off.

Can someone maybe help? I like math and now it's just for my understanding what a 'weighted system' means in direct math. Thanks :)




zgrssd -> RE: [1.04b1] Traffic Signs Numbers issue! (7/3/2020 3:55:00 PM)

quote:

ORIGINAL: GodwinW

Uhhh. That's not clear to me [:D]
I don't know what you mean.
'Weighting system'?

I spent some time thinking about it, but I cannot arrive at a value of 177. 20% of 200 (one side) is 40. So 160. 20% of the whole 400 is 320.


I think I got it:
80+100 = 180
So the split is 80/180 and 100/180

400*80/180 = ~177
400*100/1780 = ~222

They use a lot of weighted lists in Paradox games and their modding, so I am somewhat acustomed to it[:)]




GodwinW -> RE: [1.04b1] Traffic Signs Numbers issue! (7/3/2020 5:24:12 PM)

Nice! Thank you! :)




Gozzon -> RE: [1.04b1] Traffic Signs Numbers issue! (7/3/2020 7:00:24 PM)

Weighted system is good when you do not quite know the amounts or do not care for the amounts and do not try to convert everything to fit 100%. You can do weighted random roll by say chance is 33, 33, 67, 109. You just sum up all the "chances" and then do a random roll from 1 to max number and you got your weighted roll. Saves some calculation

For supplies it is as zgrssd stated: logistic points / sum of weight * weight of the direction. The logistic points / sum of weight basically calculates the "weight value" which is then used to calculate the actual value we are looking for by multiplying the weight (weight is 100 - road block value and this is why I hate the road block since it is inverse of what you are actually looking for).




Page: [1]

Valid CSS!




Forum Software © ASPPlayground.NET Advanced Edition 2.4.5 ANSI
2.28125