News:

Simutrans Wiki Manual
The official on-line manual for Simutrans. Read and contribute.

PAK128 Pricing & Balancing - Discussion

Started by DirrrtyDirk, October 10, 2008, 03:21:44 PM

Previous topic - Next topic

0 Members and 1 Guest are viewing this topic.

Zeno

Quote from: prissi on October 17, 2008, 01:53:15 PM
Well, the Napik was done assuming five years to pay back for everything.
That's only a decision we've got to take. I think at the moment the excel has ten years for locos...

Quote from: prissi on October 17, 2008, 01:53:15 PM
With the slower default flow of time, this is just one year three months.
Slower flow? Has the game engine time flow changed? I mean, I know of the bits per month parameter that regulates the real play time, but has anything else been changed in the engine to alter time flow?? So 0,0113 squares/second is still valid using 19 bits per month or not?  :-[

prissi

Yes, but the distance covered per year is now four (or eight) time larger => faster return of investment.

Zeno

#37
Quote from: prissi on October 17, 2008, 02:15:47 PM
Yes, but the distance covered per year is now four (or eight) time larger => faster return of investment.
Four(eight) times bigger...?? I'm afraid I don't understand why at all  :'(

In the excel it's been calculated as 0,0113 sqares/second x 260 seconds/month x 12 months/year ... If 0,0113 is valid, and I hope a year still has 12 months!! So, it only remains 260 seconds/month (it's been used in the excel as DayLength, but I think it's wrong as later it uses DaysInYear=12); Anyway it's not explained anywhere. So I guess a month now isn't 260 real seconds, is it?


Edit: Arrrrg!!! I hadn't noticed!! Four times more distance covered per year means much lower RC!!!  :o Errr... won't really affect, I see :P
Edit2: Is that (4x/8x) true for current stable (r1867), or is a new (nightly) implementation??

DirrrtyDirk

x4/x8 is probably the change in bits_per_month from hard-coded to the new default values (depending on pak) configurable in simuconf.tab
  
***** PAK128 Dev Team - semi-retired*****

Zeno

Quote from: DirrrtyDirk on October 17, 2008, 03:33:24 PM
x4/x8 is probably the change in bits_per_month from hard-coded to the new default values (depending on pak) configurable in simuconf.tab
Then the excel sheet might be correct, as it's written "19 bits/month", as in the default simuconf.tab...  But 19bits/month means 2^19=524 seconds per month, not 260 ???

DirrrtyDirk

262 would be 18 bits per month - and I think that was the hard-coded entry. So maybe the numbers are still from 18, and just the text has changed to 19...?
  
***** PAK128 Dev Team - semi-retired*****

Zeno

Quote from: DirrrtyDirk on October 17, 2008, 04:08:35 PM
262 would be 18 bits per month - and I think that was the hard-coded entry. So maybe the numbers are still from 18, and just the text has changed to 19...?
Mmm... might be. Anyway, some calculations are not based in distance (as average car/wagon income), and changing this value in the excel sheet makes things weird; i guess because of raw data used (saved data). It will need some update ;)

DirrrtyDirk

I know. I've had Napik's Excel sheet for quite a while...  ;)
  
***** PAK128 Dev Team - semi-retired*****

Zeno

Quote from: DirrrtyDirk on October 17, 2008, 04:37:43 PM
I know. I've had Napik's Excel sheet for quite a while...  ;)
Actually I think one of the problems to solve is the average income of a wagon/car is pre-calculated, "raw" saved data. I think is a good way to use prissi's way of thinking on this: first get an average revenue per ton and km and then apply it to driven distance to get the expected yearly revenue.

DirrrtyDirk

Go ahead - as long as it works, pretty much everything is acceptable!
  
***** PAK128 Dev Team - semi-retired*****

Zeno

Can anyone please tell me how to apply bonus levels? I wish to know how to convert the difference of speeds (vehicle_speed-bonus_speed) to the number of levels it represents. I guess if level of bonus is 3 then is applied the bonus 3 times... am I right? Or maybe is applied bonus*3?

prissi

Well, not sure what you want. The money you get is:

revenue*(1+((100*v_max)/v_speedbonus-100)*bonus_faktor/1000)

v_max is the convoi top speed
v_speedbonus the current speed according to speedbonus.tab (or the average speed of all engines of this category)
bonus_faktor is between 0 and 18 in pak64, pak128 may differ ...

DirrrtyDirk

Between 0 and 18 is true for pak128 also.

And about how the numbers work - look a little earlier in this thread: http://forum.simutrans.com/index.php?topic=584.msg4852#msg4852

  
***** PAK128 Dev Team - semi-retired*****

Zeno

Quote from: prissi on October 17, 2008, 07:23:40 PM
revenue*(1+((100*v_max)/v_speedbonus-100)*bonus_faktor/1000)
Uh... seems quite easy...  ::)

Quote from: DirrrtyDirk on October 17, 2008, 09:12:19 PM
And about how the numbers work - look a little earlier in this thread: http://forum.simutrans.com/index.php?topic=584.msg4852#msg4852
Ups. Sorry for not searching before ask... :-[



Btw, I'll take a look... I think I'll have some nice calcs soon :)

prissi

If you look in cars.xls from pak64, there is already the formula used in the income column ...

Zeno

#50
Quote from: prissi on October 18, 2008, 06:51:57 PM
If you look in cars.xls from pak64, there is already the formula used in the income column ...
Thanks, moreover it's also in this excel (somewhere in goods or bonus sheet).

I've been doing a little remake of the excel... but I get revenues much higher than before... I must have made a mistake somewhere...

Anyway, I attach the current xls here (I've deleted it from previous post).




Edit: Actually, now revenues are more credible... they're more like 50 thousand cr/year or 120 thousand cr/year and so. Before they were like five thousand or so! But I get terribly higher RCs (if compared to previous). I think they wouldn't fit at all in ST. But I can't see what can be the causes to this  :-\
If you could take a look, I've updated some explanations to keep it as easy as I could. Also remade the macro so ef speed is calculated a lot faster. Updated xls.

prissi

Well, taking into account that vehicles do not run always in straight lines, need more time than possible with max speed and are not always 100% filled, I used 10% of the maximum revenue for running cost for pak64. Maybe this would give sensible results too?

Zeno

#52
So you suggest to use 10% of theorical revenue? I saw in this sheet was established a "efficiency" percentage, 75% for pass and 48% for cargo (75% loaded passenger in both runs; 95% loaded cargo in 1 run, empty on return = 48%). I think it's the same concept, but maybe you  used a lower value.
Actually, I get much more logical values for RC if I take 25% of income instead of 100%.
Moreover, I've also introduced a new factor: bonus percentage applied. I think I can calculate the revenue using half (f.i.) of the possible bonus (maybe 75% or other value), as trains might not recieve always the maximum bonus value...  ::)


Edit: Added updated xls.

Zeno

Well, now the excel file looks quite nice. I managed to get new values with deviations between 25-30% vs old ones. I've added a couple of new things to the sheet: (In red background) at the begining of the sheet, there are two factors affecting ALL values of the locos, which are bonus and pricing percentage used to calculate RCs; also the average deviation for the whole pak (only locos).
This two factors, are simple: how much of theoric max. revenue and bonus are used to get RC value. Using 12% of the max revenue and around 40% I get quite low global deviation, which means I don't get too away from what it is right now, while I keep % of max revenue near reasonable values (prissi says he used 10% for pak64).
Note the deviation is in absolute value, so it's not a deviation from medium value but measures just deviation importance. The good thing is that the values are now calculated the same way for all locos in the pak, so they should be balanced. I think now it's just a matter of testing in-game and correcting (if needed).
Now the purchase prices bother me, because I used 5 years payback to calculate it, but I *really* think 10 or even more years should be a more realistic value (despite being conscious it will be really hard to start a game with current values if vehicle's prices are two or three times higher). I place you to discuss also this; what do you think of increasing a bit vechicles purchase pricing, and also starting money for the player to compensate this?
I leave the last version of the sheet for you to take a look.

jatypc

Well done. I had a look at the loco purchasing prices and computed them using your Excel file for 10 year payback. In my opinion - and with the playing experience of the current pak128 - the prices are reasonable (I do not know however how significantly the profits changed from old to new system). We might need to adjust the starting capital to account for higher vehicle prices, but not too much - it seems to me that 500 thousands instead of 300 thousands should be more than enough.

The question is, of course, what about the purchasing price of wagons - should it also use in future 10 year payback? It looks sensible.

VS

This is starting to look good! Thanks!

What will be needed to transform the numbers into reality?

My projects... Tools for messing with Simutrans graphics. Graphic archive - templates and some other stuff for painters. Development logs for most recent information on what is going on. And of course pak128!

Zeno

Quote from: VS on October 20, 2008, 01:37:46 PM
What will be needed to transform the numbers into reality?
Mmmm... I'll need all DATs to modify them, and lot of patience I guess... ;D

wernieman

I hope you understand my English

prissi

About the 10%: This I used for engines, for wagon I used 1cr/ton/tile and for trucks 40%, for ships 48-49% and for planes 70% of the theoretical revenue.

gauthier

I haven't read all of the topic so I don't know what are you doing (I tried understand the excel sheet but ...)

I would only know if the prices will increase or decrease ?

Zeno

#60
Quote from: prissi on October 20, 2008, 08:06:47 PM
About the 10%: This I used for engines, for wagon I used 1cr/ton/tile and for trucks 40%, for ships 48-49% and for planes 70% of the theoretical revenue.
Printed! I'll take them as reference values. Actually, I got best results for engines with aprox 10%, so I expect the others will be also quite similar. Thanks a lot prissi.
Btw, I've developed a small tool which I use to massive replace values in DAT files. As I don't know how do you deal with DATs, I can provide the tool to the comunity when finished if you wish so.



Edit:
Quote from: gauthier on October 20, 2008, 08:29:21 PM
I haven't read all of the topic so I don't know what are you doing (I tried understand the excel sheet but ...)
I would only know if the prices will increase or decrease ?
Don't get scared! The prices may be rebalanced in near future; some can go up, some down, but not in a significant amount (20-30% in worst cases). We are trying to establish some basis and rules on how should running costs be calculated for all vehicles.

VS

A side note- originally the excel was designed so that you could export the relevant part to csv and load into simutranslator - and vice versa. Now that it is not certain when it finally works 100%, I would really prefer to have this functionality doubled/backed up...

My projects... Tools for messing with Simutrans graphics. Graphic archive - templates and some other stuff for painters. Development logs for most recent information on what is going on. And of course pak128!

Zeno

Quote from: VS on October 20, 2008, 10:32:24 PM
A side note- originally the excel was designed so that you could export the relevant part to csv and load into simutranslator - and vice versa. Now that it is not certain when it finally works 100%, I would really prefer to have this functionality doubled/backed up...
Uh? I had no idea about this. Could you please give a little detail of this process of data import/export? Is there a tool already done, or it's done with a script, or...?

VS

It's a feature of translator. You can upload a csv file with data for the vehicles. But I don't really know if it works or not, I never used it...

My projects... Tools for messing with Simutrans graphics. Graphic archive - templates and some other stuff for painters. Development logs for most recent information on what is going on. And of course pak128!

Zeno

Quote from: VS on October 21, 2008, 09:03:49 AM
It's a feature of translator. You can upload a csv file with data for the vehicles.
Nice! Is there any documentation on file format expected? I mean the columns, order and so...

Quote from: VS on October 21, 2008, 09:03:49 AM
But I don't really know if it works or not, I never used it...
Uh, oh...  ;D

VS


My projects... Tools for messing with Simutrans graphics. Graphic archive - templates and some other stuff for painters. Development logs for most recent information on what is going on. And of course pak128!

Zeno

Nice, it does exacly the same than my tool (with some more fields), and externally looks like using quite similar formats and so...  ::)

Zeno

I've newly compiled the vehicles.all.pak with 10% revenue used to calc RC and purchasing costs about 2 or 3 times higher; that might be compensated by using higher player cash.

I think these values will be quite equilibrated, as RCs appear to be a bit lower. I'll test tonight or tomorrow (and probably during the whole week) to play with these values and see what happens. If someone wants to test this beta pak, post here and I may publish it on a web/ftp site (german file hosting could be a good option).

prissi

Do not try this on the actual translator, as this only operated on pak with internal reference 0, which is now the base translation texts.

Zeno

Quote from: prissi on October 21, 2008, 04:47:38 PM
Do not try this on the actual translator, as this only operated on pak with internal reference 0, which is now the base translation texts.
Ok. I'll keep using my own tool for a while. If ever it's working back, will be nice to have a couple of ways to do it ::)