[Twinsen] [0.17.74] incorrect statistics

[spiral_power]

I found some incorrect statistics.

This map runs 4,199 hours ( 251,940 minutes ) , including about 500 hours ( 30,000 minutes ) at 0.17.xx .
And it runs at 4kspm (4 rockets/minutes ) except for the first construction hours.

Item: 250 hours
Probably correct.

Item: 1,000 hours
It runs only 500 hours, so counts must be about twice of 250 hours.
Plastic or lower are probably correct.
Green circuit or upper are incorrect.

item: all time
Satellite consumption 122k, 0.4/m is too few, this count is same as 1,000 hours satellite counts ( probably correct ) .
It must be about 908k, 3.6/m
On the other hand satellite production is correct.

fluid: 250 hours
Probably correct.

fluid: 1,000 hours
Water production 21G don't fit consumption 25G.
1G fluid is 25k fluid * 40k, this is not buffer or something.
Older version ( 0.16 ) I found fluid 50 hours statistics incorrect when I use nuclear power generation (much water is handled).

fluid: all time
Probably correct.
Water and crude is somewhat a lot because old factory was inefficient.

Item: 250 hours

item_250.png (183.73 KiB) Viewed 5345 times

Item: 1,000 hours

item_1000.png (179.24 KiB) Viewed 5345 times

item: all time

item_all.png (194.9 KiB) Viewed 5345 times

fluid: 250 hours

liq_250.png (73.97 KiB) Viewed 5345 times

fluid: 1,000 hours

liq_1000.png (64.31 KiB) Viewed 5345 times

fluid: all time

liq_all.png (42.61 KiB) Viewed 5345 times

save data:
https://drive.google.com/open?id=1gkmhG ... OF7AHRzCX3
(I couldn't upload save data to this forum.)

[Twinsen]

How long did the game run during 0.16.51 ?
There have been a large number of fixes in 0.17.xx, so it's possible that any statistics data generated before 0.17.68(last fix) was incorrect.

I will investigate the save, but without a way to reproduce the issues it's hard to understand what went wrong.

EDIT: I reset the statistics and ran the save for some time. Everything was ok, so I believe the problems are due to previous bugs that have now been fixed. The old data unfortunately can't be corrected.

[spiral_power]

0.16.51: 4,199 - 500 = 3,699 hours.
0.17.xx ( only stable ver. ): 500 hours.

I understand it's hard to correct the OLD statstics.

But, this is my latest statistics, items 1,000 hours and 250 hours.
(ver 0.17.74 now)
I'm now going to RAISE a research pace.
1,000 hours blue,yellow, and left orange lines ars flat, it is strange.
Somewhat still wrong.

additional info: 250 hours left orange line ( iron plate ) has odd ceiling 335k.

You mean I need to restart a new game with the latest version or edit the save data for correct statistics?

1.png (179.5 KiB) Viewed 5162 times

2.png (184.81 KiB) Viewed 5162 times

[Twinsen]

u mean I need to restart a new game with the latest version or edit the save data for correct statistics?

There is no way to fix the old data so unfortunately you need to start a new game.


The other things you mentioned are very strange. I will investigate.

[spiral_power]

I see, I accept the old data's statistics incorrectness.
Thank you for your effort.

I'm still going to raise from 4 kspm to 7 kspm, several 335k ceilings appear at items 250 hours.
( This is only additional information, you don't need to repry about above ceilings. )

[Twinsen]

I figured it out. Because I used float, adding one item at a time was no longer possible because of precision.

If you created more than 1398 items per second the 1000h statistics would bug out. (higher for other intervals)

I changed it to double, so now it will work, unless you manage to create 750599937895 items per second.

Technical details here: https://stackoverflow.com/questions/125 ... 77216-in-c

Fixed in Version: 0.17.77

You should see the new statistics being generated correctly, but as before, the old data will still be incorrect until a new save is started.

Thanks for the report.

[spiral_power]

I understand what is wrong, 335k,83k ceilings, copper cables are made double so 166k ceilings.

item: all time
Satellite consumption 122k, 0.4/m is too few, this count is same as 1,000 hours satellite counts ( probably correct ) .
It must be about 908k, 3.6/m
On the other hand satellite production is correct.

But, my first post, the probrem satellite consumption count at all time is wrong may be other cause.
I don't check all stuff but aren't there a trivial mistake ?
( all time satellite consumption count displays 1,000 hours satellite production or consumption count )

[Twinsen]

That's a bug from before 0.17.0 where satellites didn't get counted

See 62606

[spiral_power]

I see, thank you.