Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss

2.7i bug? Build orders not being followed completely

13 views
Skip to first unread message

Tocis

unread,
Jul 19, 1999, 3:00:00 AM7/19/99
to
(Don't know if this gets through as my ISP seems to have trouble with its
news server again - I don't get anything since about 48 hours. If you want
to reply, please do so per email... To...@gmx.net)

Hi,

I am regularly encountering the following phenomenon in a currently running
autohost game and find no explanation - someone please tell me if I
overlooked something or found a bug.

I give a planet build orders (no matter if I build, say, stealth bombers or
chaff) in an amount that will just fit into one year. The planet building
this is maxed out (full population, mines etc.). It is not under attack in
any form, no population or minerals are pulled from it, the research
percentage (0 %) is not being changed. The build order is the first item in
the queue. Everything I know tells me that yes, these build orders will be
carried out to the fullest during the host run.
Yet they are NOT. Always some part of the orders remains for the next year.
Say, I order 190 chaff ships... about 50 or 30 of them (not always the same
number or so it seems) remain in the queue as if I ordered more than one
year's worth of build.

It did not happen in the earlier stages of that game, only since (IIRC)
about 10 to 20 turns. It will not matter much in this game as I am one of
the losers anyway, but still I would like to know what's up.

Any explanations? Is there a chance that I overlooked something, or should I
prepare to send the zipped files to bugstars?

--

Greetings from Tocis (9th member of the Grey Council)
Info at http://geocities.com/Area51/Chamber/8686/greyc.html

Morten Lassen

unread,
Jul 19, 1999, 3:00:00 AM7/19/99
to
>Any explanations? Is there a chance that I overlooked something, or should
I
>prepare to send the zipped files to bugstars?


That propably won't do anything.. They aren't looking at Stars! anymore..

:´(

Sincerly,
Morten Lassen
starp...@danbbs.dk
http://fly.to/starplayer


Wes.

unread,
Jul 19, 1999, 3:00:00 AM7/19/99
to
Only thing I can think of is

1- do you have the "Contribute only leftover resources to reseach"
unticked? and
2- did you play with your reseach levels after assigning you build
instructions?

Doing both these together I have found that your green cue turns blue.

Wes.

Tocis wrote in message <7mudp2$qk$1...@news.dnsg.net>...

>Any explanations? Is there a chance that I overlooked something, or
should I
>prepare to send the zipped files to bugstars?
>

Pier Donini

unread,
Jul 19, 1999, 3:00:00 AM7/19/99
to
In article <7mudp2$qk$1...@news.dnsg.net>,
"Tocis" <To...@gmx.net> writes:

> Say, I order 190 chaff ships... about 50 or 30 of them (not always the same
> number or so it seems) remain in the queue as if I ordered more than one
> year's worth of build.

Did you hit the max number of fleets? Just a guess...

Pier.

James Moody

unread,
Jul 19, 1999, 3:00:00 AM7/19/99
to
On Mon, 19 Jul 1999 07:36:32 +0200, "Tocis" <To...@gmx.net> wrote:


>Yet they are NOT. Always some part of the orders remains for the next year.

>Say, I order 190 chaff ships... about 50 or 30 of them (not always the same
>number or so it seems) remain in the queue as if I ordered more than one
>year's worth of build.

IIRC it's a rounding error. The way the game calculates the build
queues and the actual ships has the rounding in different places, and
it can cause this....

James Moody

(Major Denis Bloodnok)
--
http://www.armed.force9.co.uk ICQ: 7000473
"People who design things to be completely foolproof often underestimate
the ingenuity of complete fools." - The Hitch hikers Guide to the Galaxy

Mario Köhler

unread,
Jul 20, 1999, 3:00:00 AM7/20/99
to
>>Say, I order 190 chaff ships... about 50 or 30 of them remain in the queue

>IIRC it's a rounding error. The way the game calculates the build
>queues and the actual ships has the rounding in different places, and
>it can cause this....

wow, that would be the biggest rounding error in human history :-)...you
order 190 and it "chops off" 30-50...nah, i dont think so.

i agree with something that was posted before: you hit the max number of
ships/fleet

ciao mario

Dirk Thierbach

unread,
Jul 20, 1999, 3:00:00 AM7/20/99
to
Tocis <To...@gmx.net> wrote:
> (Don't know if this gets through as my ISP seems to have trouble with its
> news server again - I don't get anything since about 48 hours. If you want
> to reply, please do so per email... To...@gmx.net)

> Hi,

> I am regularly encountering the following phenomenon in a currently running
> autohost game and find no explanation - someone please tell me if I
> overlooked something or found a bug.

> I give a planet build orders (no matter if I build, say, stealth bombers or
> chaff) in an amount that will just fit into one year.

[...]


> Always some part of the orders remains for the next year.
> Say, I order 190 chaff ships... about 50 or 30 of them (not always the same
> number or so it seems) remain in the queue as if I ordered more than one
> year's worth of build.

> Any explanations?

Some things to check besides those already mentioned:

* Manually check how many ships can be built one year. You have the
min cons, number of mines and efficieny. From that, calculate the
amount of minerals available next year (and compare with the
numbers given in the GUI). Then, determine how many ships can
be built from that amount of minerals, and compare with actual numbers.

* Redo the turn with empty build queue and check minerals on surface
after the turn. Compare with above calculations.

* Make sure you don't have freighters with preset orders that take
minerals off this planet.

* The fleet limit cannot be the cause, because all the ships of same
type should show up in one fleet. So you'll get either none of
them or all.

> Is there a chance that I overlooked something, or should I
> prepare to send the zipped files to bugstars?

Also, make sure to get the .hst file backups of the appropriate year
(maybe after the game has ended) so you can repeat the strange
behavior. Helps both the develeoper in finding errors and you
in figuring out if it is actually an error or not.

- Dirk

Jason Cawley

unread,
Jul 20, 1999, 3:00:00 AM7/20/99
to

Mario Köhler wrote in message <7n1qdq$f29$1...@news.hamburg.pop.de>...

>>IIRC it's a rounding error. The way the game calculates the build
>>queues and the actual ships has the rounding in different places, and
>>it can cause this....
>
>wow, that would be the biggest rounding error in human history :-)...you
>order 190 and it "chops off" 30-50...nah, i dont think so.

Doesn't have to be that big, because the price of chaff is so low. If e.g.
the chaff was listed as costing 2 kt of germ, and you have 380 germ
available, but the cost per ship was really 2.5 germ, then 38 ships would be
left unbuilt (152 * 2.5 = 380). I have encountered in many times myself,
and the practice I adopt to deal with it is to assume the resource cost per
chaff ship is 1 higher than the stated figure. E.g. if the chaff is listed
as costing 12 resources, and I have 1800 to spend, instead of putting in 150
chaff, I will put in 1800/13 = 138, or less. If I have adequate minerals at
the planet (all types) that seems to usually give me all the chaff scouts
built - I can't recall a single occasion when it failed to do so.

But if higher errors are encountered, it could still be rounding, but
applied to each component on the ship rather than the summed cost. E.g.
hull may say 1 kt iron but really cost 1.4, engine can say 1 but really 1.3,
result is the cost of the ship may say 2kt rather than 3 (not sure, as I
haven't tested it carefully), leaving a 35% error rather than a .5 kt one.
The max error size in that case would be 50%, meaning at least 2/3rds of the
chaff would always build (using the displayed cost) - but anything under
that could conceivably be caused by a rounding error in the displayed value
at some stage (with parts listed as cost 1 anyway - 25% error possible with
cost 2 parts, etc).

Naturally, the higher overall tech levels are, the lower the cost for each
chaff ship, so the bigger this effect can look. I have seen it
occasionally, in smaller amounts, with other ship types (typically just one
leftover, sometimes a few when e.g. ordering 10-20 frigates in a turn), but
much smaller errors than I have seen in the case of chaff - which certainly
fits the rounding explanation, and fits few other imaginable explanations.

For what it is worth...

Sincerely,


Jason Cawley

Greg Jury

unread,
Jul 31, 1999, 3:00:00 AM7/31/99
to
I may not know a whole bunch about this game, but all of the responses
so far have been concerned with bugs in the program...

Could it be that someone is stealing minerals off your planet, with a
Robber Baron scanner?

Tocis wrote:
>
> (Don't know if this gets through as my ISP seems to have trouble with its
> news server again - I don't get anything since about 48 hours. If you want
> to reply, please do so per email... To...@gmx.net)
>
> Hi,
>
> I am regularly encountering the following phenomenon in a currently running
> autohost game and find no explanation - someone please tell me if I
> overlooked something or found a bug.
>
> I give a planet build orders (no matter if I build, say, stealth bombers or

> chaff) in an amount that will just fit into one year. The planet building
> this is maxed out (full population, mines etc.). It is not under attack in
> any form, no population or minerals are pulled from it, the research
> percentage (0 %) is not being changed. The build order is the first item in
> the queue. Everything I know tells me that yes, these build orders will be
> carried out to the fullest during the host run.

> Yet they are NOT. Always some part of the orders remains for the next year.


> Say, I order 190 chaff ships... about 50 or 30 of them (not always the same
> number or so it seems) remain in the queue as if I ordered more than one
> year's worth of build.
>

> It did not happen in the earlier stages of that game, only since (IIRC)
> about 10 to 20 turns. It will not matter much in this game as I am one of
> the losers anyway, but still I would like to know what's up.
>

> Any explanations? Is there a chance that I overlooked something, or should I


> prepare to send the zipped files to bugstars?
>

Ken Spoiler

unread,
Aug 2, 1999, 3:00:00 AM8/2/99
to

Greg Jury пишет в сообщении <37A2AA24...@home.com> ...

>> Hi,
>>
>> I am regularly encountering the following phenomenon in a currently
running
>> autohost game and find no explanation - someone please tell me if I
>> overlooked something or found a bug.
>>
>> I give a planet build orders (no matter if I build, say, stealth bombers
or
>> chaff) in an amount that will just fit into one year. The planet building
>> this is maxed out (full population, mines etc.). It is not under attack
in
>> any form, no population or minerals are pulled from it, the research
>> percentage (0 %) is not being changed. The build order is the first item
in
>> the queue. Everything I know tells me that yes, these build orders will
be
>> carried out to the fullest during the host run.
>> Yet they are NOT. Always some part of the orders remains for the next
year.
>> Say, I order 190 chaff ships... about 50 or 30 of them (not always the
same
>> number or so it seems) remain in the queue as if I ordered more than one
>> year's worth of build.
>>


Perhaps, you're right! I am also encountering such a behavior in the game,
hosted by Autohost.

>> It did not happen in the earlier stages of that game, only since (IIRC)
>> about 10 to 20 turns.

Yes, it started to happen occasionaly in 30-40s turns. The behavior is the
following, and this appearing repeatedly, but have no obvious
predictability, but it is really a bug. I did not submit it yet because of
lack of strong evidence.

1) Items, marked in queue as 'green', were NOT completed next turn, with
1-10 resource shortage.
2) There were no population dies/growth, enemy pop-drops, bombing, ANY
fleets in orbit, comet strikes, remote unterraforming, etc.
3) There were no mineral shortages or sudden queue unblocking for other
items.
4) Flag 'do not contribute to research' was ALWAYS ON.

Sometimes research is not completed for 1-40 res too. NO TECH WERE
RESEARCHED OR TAKEN IN ANY WAY THIS TURN (for the simple explanations
lovers, added +10 cost for each tech level plays no role here). Also no
unblocked queues/enemy invasions/everything affecting resources this turn.

>>
>> Any explanations? Is there a chance that I overlooked something, or
should I
>> prepare to send the zipped files to bugstars?
>>

If you have it all with hst file, this will be great! ZIP and send, with
adequate description. My problem is that I have no HST file.

Maybe not a STARS problem :-)) . Perhaps AutoHost computer uses a 'bugged'
Pentium for math calculations :-))) or some clone with the same problem
:-)))

With best regards, Ken.

HeWhoTries

unread,
Aug 2, 1999, 3:00:00 AM8/2/99
to
I also am having the same problem.
Just can't explan it.
Have fun...
Lloyd

Jason Cawley

unread,
Aug 2, 1999, 3:00:00 AM8/2/99
to

HeWhoTries wrote in message
<19990802051748...@ng-cr1.aol.com>...

>I also am having the same problem.
>Just can't explan it.


Um, guys, this is rounding on the price of ships (and bases) - that's all.
It has been discussed before and explained here - first time a long time
ago, again in the last week or two.

The whole "kernal" for executing a stars turn is *not* running when you are
looking at your displays and giving your orders, you know. Just a "front
end", that generates the .x file that game will use when you hit F9 or when
the host hits generate. Those display values are integers, and as far as
the front-end knows, so are the ones used internally. But the real internal
uses exact amounts for things like minituarization of costs (both minerals
and resources). The help file will tell you that the Q gives you an
*estimate* for time of completion, that's it, nowhere are you told the Qs
estimates will be exact, because they aren't. When you execute, the kernal
is not asking the .x file whether your display line was green ;-) It is
adding up the totals for all the parts on the ship, discounting 4% per tech
level over minimum requirements, etc. I don't see this as a bug at all.
You want displayed costs out to the 3rd or 8th decimal place or something,
for every data-point in the game? Mineral concentrations and colonist too?
Ugh, I don't.

To deal with it practically, just leave a few leftover resources when you
order a lot of small ships in one year. The total amounts are low, and will
only be noticeable at all when the cost of each of the design you are
building is very low (like chaff).

For what it is worth.


Sincerely,

Jason Cawley

HeWhoTries

unread,
Aug 2, 1999, 3:00:00 AM8/2/99
to
> The total amounts are low, and will
>only be noticeable at all when the cost of each of the design you are
>building is very low (like chaff).

Thanks, I don't always follow things here.
That explans that.
Have fun...
Lloyd

Don

unread,
Aug 5, 1999, 3:00:00 AM8/5/99
to

> Maybe not a STARS problem :-)) . Perhaps AutoHost computer uses a 'bugged'
> Pentium for math calculations :-))) or some clone with the same problem
> :-)))
>
> With best regards, Ken.

Intel Strikes again!

Don

Don

unread,
Aug 5, 1999, 3:00:00 AM8/5/99
to
You go Jason

Don

Jason Cawley <jason...@email.msn.com> wrote in message
news:uPI6QHN3#GA.464@cpmsnbbsa02...

> order a lot of small ships in one year. The total amounts are low, and


will
> only be noticeable at all when the cost of each of the design you are
> building is very low (like chaff).
>

0 new messages