Obtaining Better B-Spec Stats ??

  • Thread starter LivingSoul
  • 552 comments
  • 126,911 views
My log is now updated. Current Bspec attempt goes from 1111 to 1511. Result at that point was 9998/100/99/99.

I double-checked the notes, and also verified (with a program) that all my reports of the form B:NNNN/MM/CC/BB(+N/+M/+C/+B) add up properly.

I have not yet verified that my per-race Bspec notes add up properly.

In my case, I definitely collected occasional floating points from some races after starting from a "random" game, but cannot really know how many floating points I was short when I began racing all tracks.

http://www.geocities.com/gt2toxs/gt/logs/gt4-game1.txt
http://www.geocities.com/gt2toxs/gt/logs/

Thanks for all the help. I don't have the energy right now to single-mindedly chase the last two points, however. I'm going to Aspec for a while.

Actually day 1149 is where I actually began systematically rounding up the tracks in the correct four cars.
 
Orion_SR
Actually, I am confident that you are the first PAL player to report maxing out the B-spec points on this board, and no other English language website has the information available here. Wear the laurel while it's still fresh. I'll admit that there must be a huge reporting bias in the data, but reaching 10000 B-spec points is still a rare accomplishment.

Your floating point observations seem to confirm Toga's original predicition that if the milestones were not reached before upgrading, then 7051 landmark would be difficult to reach. I have often criticized the JP tuning set for being weak in the middle and encourage people to experiment with other tunings. Any car or tuning can be used (in addition to or instead of the JP tuning set) on Nurburgring (or any single course with track points), and it won't change the expected landmark of 7051.

Your last track point seems to confirm predictions that battle fractions tend to get dropped on short tracks. I think I had trouble with the Super Speedway and the 350Z. Did you repeat many tracks during your first track point collection? Or target short tracks while looking for that lost point? Also, since your missing 22 floating points held thoughout the collection, how frequently did you collect 14 points on a single track?

First time round i did every track twice making sure i finished within 1 second of the leader. I am not quite sure but i think i got 14 points 4 or 5 times.
When i ended up 1 point short i did go straight for the short tracks and Super speedway was the third one i tried.
 
First of all I want to thank Orion_Sr for his notes in thread B-Spec "Scientific" Study on Jun 04 2005. With that post I have hit the 100 machine point mile stone. I have searched and read allot of posting by members such as ANK, Sportwagon, TigJackson, and others and I want to thank each one of you as well, but the notes from Orion_Sr made it easy to understand and achieved bspec goals.

Now I am working on getting 100 course and battle points so I have a question or two about that.

Orion_Sr in one of your postings you stated that an author on a Japanese website had a database of what he did and also said that itazura_da translated that database. Where can I found that translation?
Orion_SR
Cars and tuning are based on a Japanese website and translated by itazura_da; author is not identified.
http://www.geocities.jp/j8ba7yrg/gt4/7_10000.html

The other question is Orion_Sr stated that there are 58 tracks to use to get the other bspec points Ignore Rally Tracks, Test Course, Tsukuba Wet and Suzuka West. I am playing the NSTC version and I only count 42 tracks. Where are the other 16 or are there certain tracks that is used forward and reverse?

Thank you once again, you, the members and gtplanet has made GT4 such a great game to play.
 
gt4falcon
Orion_Sr in one of your postings you stated that an author on a Japanese website had a database of what he did and also said that itazura_da translated that database. Where can I found that translation?

The other question is Orion_Sr stated that there are 58 tracks to use to get the other bspec points Ignore Rally Tracks, Test Course, Tsukuba Wet and Suzuka West. I am playing the NSTC version and I only count 42 tracks. Where are the other 16 or are there certain tracks that is used forward and reverse?
.

The key translations were quoted in this thread, but there were several follow up translations and clarifications that never made it to this topic. The following link contains a record of the development of B-spec theory and the original translations. Becareful while reading this record, there is a lot of bad theory which was discarded as more data became available. The track list and complete website was never fully translated. I used software translators while searching for more clues and details in the notes at the end of the page, but don't have a record of the translations.

http://andypsionfan.users.btopenworld.com/GT4/bspec08.html

Original JP Website - paste address into translators (only works 1/2 way)
http://www.geocities.jp/j8ba7yrg/gt4/7_10000.html

This link contains the JP predicition table for the final 3000 points. However, this table has not been confirmed by anyone and is not predicted by current Fraction Theory.
http://www.geocities.jp/j8ba7yrg/gt4/3000_1.html

Edit: I should have updated this link a long time ago. Sorry. The JP prediction table is completely accurate and has been confirmed many times. Follow the numbers, don't move on until collecting everything. After collecting 7051 on Nurburgring, every 29 races (375 points) the pattern of +12s and +13s will repeat, and the +14 on the 29th race confirms that no fractions were dropped during that cycle. 7375 is the first hard milestone during the collection. Nurbrugring counts as the 1st 4 tracks of the sweep (7000 to 7051). Below is a link to the JP authors updated Comprehensive skill 10000 templet. The numbers haven't changed, we just understand them better.

http://www.geocities.jp/j8ba7yrg/gt4/10000_t.html
http://translate.google.com/transla...://www.geocities.jp/j8ba7yrg/gt4/10000_t.html

Here is the track list from the spreadsheet I used to chart points. NRX is for normal, reverse, or does not apply. Sort is used to keep the tracks in the order presented by the game. I'm not sure if the tracks are unlocked on the days specified, I didn't have a good reference when I made the list.
Location, Sort, Course, NRX, Unlocked

City, 1, Seoul, N, 1
City, 1, Seoul, R, 1
City, 2, Tokyo Route 246, N, 1
City, 2, Tokyo Route 246, R, 1
City, 3, Hong Kong, N, 113
City, 3, Hong Kong, R, 113
City, 4, Clubman Stage Route 5, N, 1
City, 4, Clubman Stage Route 5, R, 1
City, 5, Special Stage Route 5, N, 57
City, 5, Special Stage Route 5, R, 57
City, 6, New York, N, 1
City, 6, New York, R, 1
City, 7, Seattle, N, 1
City, 7, Seattle, R, 1
City, 8, Cote d'Azur, X, 183
City, 9, Opera Paris, N, 29
City, 9, Opera Paris, R, 29
Driving Park, 3, Motorland, N, 1
Driving Park, 3, Motorland, R, 1
Driving Park, 4, Beginner, N, 1
Driving Park, 4, Beginner, R, 1
Original, 1, High Speed Ring, N, 1
Original, 1, High Speed Ring, R, 1
Original, 2, Midfield Raceway, N, 1
Original, 2, Midfield Raceway, R, 1
Original, 3, Grand Valley East, N, 1
Original, 3, Grand Valley East, R, 1
Original, 4, Grand Valley Speedway, N, 99
Original, 4, Grand Valley Speedway, R, 99
Original, 5, Deep Forest, N, 15
Original, 5, Deep Forest, R, 15
Original, 6, Trial Mountain, N, 1
Original, 6, Trial Mountain, R, 1
Original, 7, El Capitan, N, 1
Original, 7, El Capitan, R, 1
Original, 8, Autumn Ring Mini, N, 1
Original, 8, Autumn Ring Mini, R, 1
Original, 9, Autumn Ring, N, 281
Original, 9, Autumn Ring, R, 281
Original, 10, Apricot Hill, N, 169
Original, 10, Apricot Hill, R, 169
World, 1, Tsukuba, X, 1
World, 2, Fuji Speedway 90s, X, 1
World, 3, Fuji Speedway 80s, X, 43
World, 4, Infineon Raceway Sports, X, 295
World, 5, Infineon Raceway Stock, X, 309
World, 6, Laguna Seca, X, 1
World, 7, Twin Ring Road, X, 211
World, 8, Twin Ring East Short, X, 85
World, 9, Twin Ring West Short, X, 1
World, 10, Twin Ring Super Speed, X, 1
World, 11, Suzuka Circuit, X, 71
World, 12, Suzuka East, X, 1
World, 13, Fuji 2005, X, 323
World, 14, Fuji 2005 GT, X, 141
World, 15, Nurburgring, X, 1
World, 16, Sarthe I, X, 267
World, 17, Sarthe II, X, 351


 
Awesome thanks a million for your tack list Orion_Sr.

I have come across and seen the JP website and point prediction tables before. I guess I am going to have to find and get a translator for the JP website.

Thank you again the track list will help some time.
 
IMG
(dead photobucket links) (or should be)


img.photobucket.com/albums/v494/SportWagon/GT4/Bspec10000Hsm.jpg

img.photobucket.com/albums/v494/SportWagon/GT4/Bspec10000Ssm.jpg

Achieved Saturday August 13, 2005.


I add them as attachments...
 

Attachments

  • Bspec10000Slg.jpg
    Bspec10000Slg.jpg
    40.7 KB · Views: 33
  • Bspec10000Hlg.jpg
    Bspec10000Hlg.jpg
    37.4 KB · Views: 30
Last edited:
On Saturday I went to Super Speedway and did Family Cup races with the Lupo, Miata and RSC'02 and saw no visible change.

When I came 3rd by less than 1/2 second in the 350Z Concept LM (TS3) I got the last two points, much to my relief.

Pulling out an earlier save, I couldn't get past 9999 trying the 350Z alone.
 
SportWagon
I got to 9998/100/99/99 :banghead: :banghead: :banghead: :banghead:

A question would be, how likely is it that the two points are spread out in tiny fractions across many tracks?
Now that you're one of the experts, would you care to answer your own question?

I would suspect that over the next few days, a lot of people with 9999 B-spec points will be making a few runs on Super Speedway. BTW, both of my 10000th points came from loosing races. 2nd in the test game, and 3rd in my clean up game.
 
Orion_SR
Math Review
58/2=29
Track:3000/58= 12 and 27/29
Course:2000/58= 8 and 18/29
T-Battle:1000/58= 5 and 9/29
Actually, 1000/58 is 4 and 9/29. Well, that is, if you divide by the 4 categories as well (as you must do for the other two numbers).

Well, I don't think the bitmap is actually as large as 290000 bits.

I offer as a hypothesis that it's more like 8 bits Course per category, and 4 bits T-Battle per category, (total 2784) multiplied up to give the even 3000. It might be 10 and 5 multiplied down, but 8 and 4 would by a byte and a nybble--a naturalish programming decision. That 8 and 4 would give a granularity of 125/116 (3000/2784), or slightly more than one. If they do maintain granularity finer than that, it's not obvious where and when it was (not) given.

If the granularity is much less than this, I think it would be highly likely that many people would lose just a few 29ths here-and-there scattered across many tracks, and it would be very difficult to clean up and get 10000 points.

With granularity greater than 1, no special fudging would be needed to make sure 10000 points was "really" 10000 points. If it's not that large, one wonders how many "10000" could really be only slightly more than 9999+15/29. Actually, thinking about it (see below), granularity greater than one-half is sufficient.

Hmm. There is your observation of category points changing without the large total changing. I believe that proves the granularity is less than 1 (but it could be arbitrarily close to one), so my hypothesis needs revision. 10 and 5 would probably be sufficient, but seem unlikely choices. 16 and 8 might be the next likely choice.

10 and 5 would give 3000/3480 or 25/29 as the minimum increment.
16 and 8 would give 3000/5568 or 125/232 as the minimum increment.
(3480 is 4*(10+5)*58; 5568 is 4*(16+8)*58)
It's not impossible Course and T-Battle points are scaled differently, from independent bitmaps, even.

125/232 is still greater than 1/2, and so the total before 10000 would be 9999+107/232 which would definitely round down even without fudging.

25/29 would make sure that the fractional parts of the numbers is always a number of 29ths. I'm not sure that's important.
 
SportWagon
Actually, 1000/58 is 4 and 9/29. Well, that is, if you divide by the 4 categories as well (as you must do for the other two numbers).
Oops, I really messed up the math that time. The errors have been corrected.

B-spec math sure gets confusing in a hurry. I strongly suspect that the different types of points aren't very different from each other because they seem to have inter-related influences on skills. I'm not sure how to explain this - since any combination of 100 whole course points and/or battle points "adds up to" 1 course skill, the points can't be all that different in how they fulfill the requirements to build skills. This leads me to believe that T-battle "beads" (bead = smallest B-spec unit) are bitmapped just like course beads, except they also have an influence on battle skill.
 
Orion_SR
B-spec math sure gets confusing in a hurry. I strongly suspect that the different types of points aren't very different from each other because they seem to have inter-related influences on skills. I'm not sure how to explain this - since any combination of 100 whole course points and/or battle points "adds up to" 1 course skill, the points can't be all that different in how they fulfill the requirements to build skills. This leads me to believe that T-battle "beads" (bead = smallest B-spec unit) are bitmapped just like course beads, except they also have an influence on battle skill.
Do you have the details for your time the total skills points did not change, even though one (or more?) of the category totals did change? I.e. what were the four Bspec numbers before and after the change?

I believe knowing that we could deduce some things about the minimum increment (granularity).
 
SportWagon
Do you have the details for your time the total skills points did not change, even though one (or more?) of the category totals did change? I.e. what were the four Bspec numbers before and after the change?

I believe knowing that we could deduce some things about the minimum increment (granularity).

I don't think it will help. I may haveIseen it twice, but only once for sure. I might have records of which skill and point somewhere, but it's such a rare and uncontrolled event that I don't see any hope from learning anything.

1000/232=125/29=5*25/29=25*5/29

I suppose the minimum B-spec unit for track points could be 5/29th or 25/29th.
 
To all who have maxed out your B-Spec points.

What colour helmet do you get and is he a much better driver than at 7,000 points?
 
slotcarrod
To all who have maxed out your B-Spec points.

What colour helmet do you get and is he a much better driver than at 7,000 points?

I never noticed the colour of his helmet i will check tonight. His driving skills are still fairly bad.
 
At 10000 points:

The "Pass" button still has a large negative effect on driving skills.
At Infineon Raceway, it seems to me that:
- "Pass" on can cause wipeouts in several spots no matter what pace you choose.
- You can do a circuit at 4 on blue or yellow tires, but green tires require 3 for a couple of spots (driver overcompensates for the difference between cold and warmed-up tires)
- You must use pace 3 if you are following another car or being followed closely, at least through three tricky spots (early off-camber right-hander, second of the next two 90-degree right-handers joined by a short straight, and central left-hand 180 at the bottom of the hill).
- You will usually get through the final hairpin on pace 4, but sometimes you drift wide on exit and lose time; it still seems best to come in on pace 3, and up the pace to 5 when braking is finished; attempting to leave "Pass" on all the time will result in very bad driving through this corner--Bspec actually doesn't seem capable of passing unassisted here unless it happens to rear-end the passee first

So, to win the Infineon Enduro, I pretty much took 2.75 hours of real-time, "managing" the car the entire time. (turbo 4 Bentley Speed 8)
 
Orion_SR
SportWagon
Do you have the details for your time the total skills points did not change, even though one (or more?) of the category totals did change? I.e. what were the four Bspec numbers before and after the change?
I don't think it will help. I may haveIseen it twice, but only once for sure. I might have records of which skill and point somewhere, but it's such a rare and uncontrolled event that I don't see any hope from learning anything.
If you knew the exact number at which it occurs, we could possibly reject some "granularity" values as not allowing the achievement of N-1/2+eps, N+1/2-eps for that particular N. (Because it wouldn't be possible to have an even multiple of that value to give you an even 10000, and also allow the two successive numbers to round to the same number).

Your observed occurrences were in games in which you had achieved the base 7051+21/29 were they?
 
Here is proof of my score.


And Slotcarrod this is what my helmet looks like.

 

Attachments

  • Jay H Score.JPG
    Jay H Score.JPG
    54.8 KB · Views: 80
  • s helmet.JPG
    s helmet.JPG
    46.7 KB · Views: 70
Congratulations to you.

Looks like it's about time for you to start playing a different game.
You've accomplished everything, haven't you?

So what's it like in the year 2015? ;) (2005 + 3722 days)
 
ANK
Here is proof of my score.


And Slotcarrod this is what my helmet looks like.


Thanks ANK! 👍 My helmet was blue and now it is more black, I don't know if there is anything to it? I was just pondering why? :)

SportWagon: I can't say for ANK, but when I get all the points, I will tune all my cars and enjoy racing them. 👍 By the way, I am over 4000 days and all the dealers have the same cars for sale that they had in 2005!! What the hell??
 
SportWagon
Congratulations to you.

Looks like it's about time for you to start playing a different game.
You've accomplished everything, haven't you?

So what's it like in the year 2015? ;) (2005 + 3722 days)

I am trying to get the last few cars think i have got 721 at the moment.
Next week i have a weeks holiday from work so i will probably get a new game (but i won't stop playing GT4 just yet). Not sure what game to get out of these 3.
Resident evil 4, Tales of symphonia, or Forza.
 
I modified the title, but a search restricted by title still doesn't seem to find it. :(

You need to reduce the size of the results somehow, e.g. by restricting to GT4 Forums, or restricting by date.

I put "Bspec" in a new post title in that thread, but you seem to need to apply even more restrictions to find the thread via that word in that title.

When searching for common words at this board, it's important to restrict the size of the set returned, or it's difficult to tell which results get omitted.

Note that my new thread has been linked into the stickied/locked FAQ thread.

I.e.
LOOK HERE BEFORE YOU ASK – Many FAQs at https://www.gtplanet.net/forum/showthread.php?t=53246


By-the-way, "maximizing" with a "Z" is arguably the correct spelling, even in Britain.
(Contrary to what the above FAQ thread uses).
 
SportWagon
By-the-way, "maximizing" with a "Z" is arguably the correct spelling, even in Britain.
(Contrary to what the above FAQ thread uses).
Ah, well. This is why I usually avoid discussing proper spelling. I suspect the topic will get re-indexed as it is used and be found more quickly by the engine. But I think you caught my main point; we want people to find your topic quickly if "B-spec" is the only search criteria used.
 
Orion_SR
B-spec: 10000/100/100/100 synthesis achieved May 6, 2005 using variations of the method described on the JP website. I was involved in two B-spec theory threads on GameFAQs.com and refer you there for more details about the theories and how they developed. You are welcome to borrow anything posted by Orion_SR in your efforts to learn more about B-spec. I made what I hope are more precise measurements of the B-spec points. These are the basic concepts of my current point theories.
What was your B-spec mileage?

And, of even more interest, what was your B-spec mileage the time you re-achieved synthesis after starting with a new game?
 
I have a question about race order. I have been studying this thread for a while now, and have been using Orion_S notes to get to 10,000. I went through all of the tracks using the 4 different classes. Until the end when I only got to 9463 and realized that I was using the wrong class of car. I was using the street Toyota RSC instead of the Rally RSC. My question is do you think that I can use the Rally RSC only and go through the tracks to get to 10,000 or do I have to use the other three classes again along with the Rally RSC one at a time for each tracks?
 
gt4falcon
I have a question about race order. I have been studying this thread for a while now, and have been using Orion_S notes to get to 10,000. I went through all of the tracks using the 4 different classes. Until the end when I only got to 9463 and realized that I was using the wrong class of car. I was using the street Toyota RSC instead of the Rally RSC. My question is do you think that I can use the Rally RSC only and go through the tracks to get to 10,000 or do I have to use the other three classes again along with the Rally RSC one at a time for each tracks?

I'll assume you have read the notes and understand the difference between track points and floating points, and the different cars and tunings required for the different points. Since the magic tuning set was discovered on a Japanese website, very little testing has been done on other methods of collecting points, so I have very little data to support this hypothesis on track points.

Let's say you found a new set of cars and tunings that when raced on Test Track (a course without track points) all 7000 floating points are quickly collected. Those cars are then sold. Now it's time to collect track points.

Pick any car at random and race it on your favorite track. That car should be able to collect all 750 points for it's class. Pick any other car and enter it in one race on the same track. If it doesn't earn any points, sell it and try another. Based on your experience of cars and power, and an understanding of the basic principle involved, it shouldn't take long to find your favorite set of 4 cars that can collect all of the track points. Track points aren't very picky. Every car seems to qualify for one of the four classes on each track.

Well, that's the theory anyway. It's never been tested, everyone is using the same set of cars. To apply this to your data, I would guess that the car you chose earned 13 (maybe 12 or 14) track points on each track. If it had earned nothing on every track, you would have noticed the problem sooner. So you probably don't need to repeat the 58 races with the RSC Rally Raid to get missing track points. Those two cars are similar enough I would guess they earned the same track points. Besides, you are well within the 750 points that would have been missed. (This guess turned out to be incorrect, the cars don't earn the same track points.)

However, the tuning requirements for floating points, especially the RSC (and Miata) on Nurburgring are very specific. It is very unlikely a different car would have earned the same batch of floating points. A similar car would earn some, but probably not all. So take the RSC A (default) and RSC B (+Medium Racing) and run a few races on the Ring. See how close that can get you to your goal.

BTW, what are your skill levels? This data is used to determine which points might be missing. Also, I'm assuming you didn't make the 7051 landmark because you picked up this thread mid-game.
 
After reading through this thread I'm still unsure if the B-spec driver needs to have many points to be successful. My co-driver has about 6200 points and he's good enough to win all the 24h endurances. That's all I expect from him as I don't have the possibility to do them myself.

So, does he get much better with more points or is it all about statistics?
 
Back