B-Spec Explored - How to get the best from Bob

[Famine]

My apologies for keeping this thread locked and enticing, but I need to bring all the information I've been collecting into one place and a semblence of order. That'll take quite a bit of time and require quite a few posts before it's ready...

I still have some data left to gather, so some of the post 1 links will remain unlinked for now. I hope to complete these shortly.


The recent Seasonal Events, not to mention the Signature Edition Challenge way back when, piqued my interest in Bob* because, largely, we have no real idea how B-Spec actually functions. We have our feelings and we suspect that doing certain things elicits certain responses, but ultimately how Bob behaves can be a bit of a mystery. The La Sarthe Seasonal in particular has highlighted how seemingly random Bobs can be, with maxed-stat, level 40 Bobs in their prime failing to do what other people's semi-retired 34s can achieve - or even long-in-the-tooth level 30s. I've personally noticed a huge difference between the abilities of a Bob I sent out to pasture some time since and his successors who cannot hold a candle to him.

So on a spare account I've decided to do the best I can to research Bob and how you can get the best out of yours. This thread will serve as an accumulation of all the data I have collected along with more general B-Spec information that I hope everyone will find useful.

My Test Bobs and Methods

• Meet the team
• Testing methods
  • Character
  • Mood
  • Bob's key stats
• Training

General B-Spec Topics

• Creating your B-Spec Bob
• What do Bob's stats mean?
  • Braking
  • Cornering
  • Acceleration
  • Physical Strength
  • Mental Strength
  • Character
  • "Condition"
• Training Bob
• Remote Racing
• B-Spec Career Mode
• Racing Bob
  • Instructions
  • Mood
  • Strength

* For the uninitiated, the shorthand way of referring to a B-Spec driver is "B-Spec Bob" or "Bob". This is simply for the reason that "Bob" is shorter than "My B-Spec Driver", can refer to any driver regardless of the random name you've chosen for him and is now just about universally understood to mean an AI cretin you can instruct. GT4 also had B-Spec drivers who weren't lucky enough to get names, so "Bob" kind of stuck.

 

[Famine]

Meet The Team​

Initially, six Bobs were created specifically for testing. Three were designated for offline use only and three for training through remote racing Bobs. Two were generated to have very hot characters (one offline, one remote racer), two were very cool drivers (one offline, one remote racer) and two were middle-of-the-road (one offline, one remote racer). Though it is not possible to randomly generate the drivers in such a fashion that I could get the three pairs to be of identical ability and character, I've kept it as close as I can. This is, however, still a variable and should be kept in mind throughout.

K.Snow - Cool, Offline
Detail screen: Braking 9%, Cornering 8%, Accuracy 8%, Physical Strength 16%, Mental Strength 5%
Overview screen: Speed 8%, Mental Strength 5%, Physical Strength 16%
Character: Cool (3%)

D.Hoffman - Cool, Remote Racer
Detail screen: Braking 9%, Cornering 8%, Accuracy 8%, Physical Strength 16%, Mental Strength 5%
Overview screen: Speed 7%, Mental Strength 5%, Physical Strength 16%
Character: Cool (3%)

D.Solis - Hot, Offline
Detail screen: Braking 9%, Cornering 8%, Accuracy 7%, Physical Strength 16%, Mental Strength 5%
Overview screen: Speed 8%, Mental Strength 5%, Physical Strength 16%
Character: Hot (91%)

B.Roth - Hot, Remote Racer
Detail screen: Braking 9%, Cornering 8%, Accuracy 8%, Physical Strength 16%, Mental Strength 5%
Overview screen: Speed 7%, Mental Strength 5%, Physical Strength 16%
Character: Hot (90%)

H.Craft - Middling, Offline
Detail screen: Braking 9%, Cornering 8%, Accuracy 8%, Physical Strength 16%, Mental Strength 5%
Overview screen: Speed 7%, Mental Strength 5%, Physical Strength 16%
Character: Hot (57%)

F.Roth - Middling, Remote Racer
Detail screen: Braking 9%, Cornering 8%, Accuracy 8%, Physical Strength 16%, Mental Strength 5%
Overview screen: Speed 7%, Mental Strength 5%, Physical Strength 16%
Character: Hot (54%)

Their ability percentages were calculated by specific measurement - yes, I used a ruler on my TV. Once I realised absolute values were of no meaningful use, I created a graded measuring card that allowed me to quickly determine what an individual Bob's ability percentages were without ripping my screen to shreds. This is incredibly useful and I'd suggest anyone wanting to do similar tests does the same.

Error margins for initial measurements are +/- 1mm in 75mm (the size of the bars on a 32" LCD screen), or +/- 1.5%. Error margins for the graded card are +/- 2%. Assume a 2% error on all measurements.

Percentages for the character bar are expressed from 0% (Cool) to 50% (Middle) to 100% (Hot).

 

[Famine]

Testing Methods - Character and Mood​

In order to test Bob's abilities as best we can, we need to minimise the variables at play. He needs testing at the same track repeatedly and, since we're looking at his racecraft, we need the same opponents. Since we can predict grids in GT5, that's not a problem, though there is a slight variable in that the AI cars don't necessarily behave in identical manners each time. I don't want Bob to interact with the other AI more than he has to so the ideal grid is one he can pass with relative ease - which means a low level B-Spec race.

For all tests I've chosen the Beginner FR Challenge at Trial Mountain. Six laps of the circuit on Sports Medium tyres or worse (no wear) against a field of middling FR opponents.

The Car​

On the same basis of minimising variables, Bob needs a car that will allow him to get past the AI field quickly even at low levels, but also not destroy in the process with his ineptitude - too many crashes may invalidate the times. I've opted for the Lexus IS F Racing Concept '08, which is very stable indeed and comes with adjustable downforce front and rear (which has been adjusted to 15/30 - the minimum settings). It's still enough of a car for Bob to get himself in trouble with if he's not careful, but we shouldn't see too many crashes.

The Test Cycle​

Each Bob will be asked to drive the race nine times (54 laps) at level 0 with 0xp. They will drive the race thrice being kept cool and forced to slow down, thrice being kept hot and forced to speed up and thrice being kept as close to the middle as possible with the pace buttons. Overtake commands and pit stops will not be used, though Bob's abilities to pass by himself will be noted.

After each individual race the console is reset and a backup save restored to prevent Bob gaining experience.

Once a Bob's test cycle is completed, he will be levelled up 10 levels - using B-Spec offline if he is an offline driver (each Bob will use the same race to gain their experience) and Remote Racing if he is a remote racer. He will then repeat the nine race sequence, his times logged, his race abilities noted and then levelled up 10 levels again. This will repeat until I have run out of inclination to do it.


This will provide us with information both on how Bob's character affects his race at different levels and how his mood affects his race at different levels, along with a minor insight into how he performs with increasing ability statistics.

 

[Famine]

Testing - Bob's Key Stats​

Once the above tests had been completed, some of the Bobs were retired and replaced with fresh level 0 Bobs with varying abilities. Each new Bob was generated to have acceptable statistics for all of the key stats except one, which would be as low as I could manage to get it. In the absence of the ability to hotlap Bobs and evaluate their performance, Bobs were then re-run in the above race with more particular attention paid to their overtaking, their braking, their cornering and their acceleration. This test is, unfortunately, qualititative and not quantitative but should generate a sufficiently clear picture to determine what it is that each statistic does to their on-track performance.

A.Smith - Cornering-deficient Bob
Detail screen: Braking 11%, Cornering 0%, Accuracy 6%, Physical Strength 12%, Mental Strength 7%
Overview screen: Speed 5%, Mental Strength 7%, Physical Strength 12%
Character: Not recorded

R.Fink - Accuracy-deficient Bob
Detail screen: Braking 16%, Cornering 16%, Accuracy 2%, Physical Strength 17%, Mental Strength 4%
Overview screen: Speed 7%, Mental Strength 4%, Physical Strength 17%
Character: Not recorded

F.Lynch - Braking-deficient Bob
Detail screen: Braking 6%, Cornering 16%, Accuracy 10%, Physical Strength 17%, Mental Strength 2%
Overview screen: Speed 8%, Mental Strength 2%, Physical Strength 17%
Character: Not recorded


In order to better determine the effects of Physical and Mental Strength, higher level Bobs were run through the Indy 500 race (as part of their training mentioned below) as hot as possible and the time taken to deplete Physical Strength fully, to start depletion of Mental Strength and the time taken to deplete Mental Strength fully were logged.

 

[Famine]

Training​

In order to establish if there is any quantifiable difference between training a Bob through remote racing and training one through B-Spec career races (or Seasonal Events), the original cohort of six Bobs were divided into two groups and each group trained with one method.

Level 0 to Level 10
The offline Bobs were levelled up through two further attempts at the first race, using an AEM S2000, each netting 2584xp (including the 200% bonus), to give them 5168xp in total, followed by a run at the GT World Championship in a borrowed Formula GT for 17,450xp. This gave them a total each of 22,618xp, 55% of the way through level 10.

The Remote Race Bobs were run in grids on the Autumn Ring/Civic Remote Race in a grid rigged to gain them the minimum possible XP so as not to overrun the level (levelling them up to 5, as per the original plan, proved impossible even without the 200% multiplier). Consistency here is very hard to ensure due to the varying prizes available and any overrun required an earlier gamesave to be restored to bring the Bob back down to level 0 to start all over again.

The three level 0 Bobs found themselves up against a very high level Bob from the host account and racked up 24,160xp, 23,120xp and 23,120xp between them to put them 60-70% of the way through level 10...

Level 10 to Level 20
The offline Bobs were levelled up to 15 first to check their stats, by a single run in Like The Wind at Daytona Road for 53,558xp. Following this they were each put through three runs at the Dream Car Championship at Circuit de la Sarthe in the borrowed Formula GT. Each netted 77,188xp for a total of 231,564xp and a grand total of 307,730xp putting them 80% of the way through level 20.

The Remote Race Bobs were run in grids on the NASCAR/Daytona Remote Race in grids designed to gain them sufficient XP. As before consistency here is very hard to ensure due to the varying prizes available and any overrun required an earlier gamesave to be restored to bring the Bob back down to level 10 to start all over again.

They required eight races each in this event, netting about 210,000xp in the process (210,760xp for Froth, 213,000xp for Hoffman and 216,840xp for Broth), putting them 50-65% of the way through level 20.

Level 20 to Level 30
For reasons which will be explored in a later post [link], the different training methods were abandoned (along with the offline Bobs) and Bobs were levelled up first to Level 25 with a single run at the Grand Valley 300 (1,598,400xp) in a borrowed Formula GT to check their stats and then up to Level 30 with a single run at the Indy 500 [info].

 

[Famine]

Creating Your Bob​

The Bob creation screen has, as far as this guide is concerned, just the one useful button:

There is a backout button on the right, a button for selecting Bob's wardrobe from your gear menu, a button for randomly selecting Bob an outfit from your gear collection, a button for randomly generating a name for Bob (formed of a random first initial and, allegedly, a random real-life racing driver's surname) and the leftmost and most important button to randomly generate Bob's stats that you can see on the right.

The top three sliders, of "Braking", "Cornering" and "Acceleration", are combined in the Bob overview to form the driver's "Speed" rating (on a similar scale). The next two, "Physical Strength" and "Mental Strength" govern aspects of Bob's endurance. The bottom one is Bob's character, from Cool to Hot.

We'll visit what each bar influences in later posts, but for all but the Character bar, bigger is better. However, each bar seems to have maximum and minimum values at the point of Bob's creation. These are:

• Braking: 8% to 18%
• Cornering: 0% to 18%
• Acceleration: 0% to 10%
• Physical Strength: 15% to 18%
• Mental Strength: 0% to 7%

There appears to be no set amount of overall "ability" to be distributed across the abilities - I've logged Bobs with an average score of 7% up to Bobs with an average score of a little over 12% and I have no reason to suppose these are either the lowest or highest possible stats.

Ultimately, these stats determine how fast Bob goes and as he levels up they will increase. Thus the most you can extract from Bob initially, the better he will perform at lower levels - my most recently created Bob started off with stats of 18%, 13%, 10%, 16% and 5% for an average of 12.4%.

 

[Famine]

What Do Bob's Stats Mean?​

The three deficient Bobs were tested in the same Trial Mountain race at as near to centre mood as possible. Behaviours and speeds were logged, but this comprises a non-scientific and qualitative test.

Braking
All three Bobs would brake on the main straight at approximately the same speed and braking point, though the braking deficient Bob - F.Lynch - would habitually brake a car length earlier and from a slightly lower speed (139mph compared to 140mph). This Bob would also give peculiar "confidence lifts" in the middle of the corner, suggesting at an in-built uncertainty of when to accelerate and when to decelerate. The Braking stat thus seems to govern both braking distances and accepted maximum speeds on the circuit.

Cornering
Two of the Bobs would happily run through the wide curve at the bottom of the hairpin at 56mph in 2nd gear. A.Smith, however, would only drive through it at 54mph and he was the Cornering-Deficient Bob. The Cornering stat would appear to govern in-corner speeds.

Accuracy
There were no significant differences between the Bobs with normal Accuracy and Accuracy-Deficient Bob R.Fink. In fact R.Fink contrived to be the quickest Bob and the best passer, reaching the lead two corners before F.Lynch and a full sector before A.Smith. At this time I do not know what the Accuracy stat governs.

Physical Strength and Mental Strength
An odd pair of stats that seem to govern quite a great deal of things.

Primarily, Physical Strength tells you how long a driver can drive for. This stat depletes during a race - though the depletion rate, helpfully, varies depending on the race. You may find a driver can only do 4 laps of Autumn Ring Mini, but he's fine doing 6 laps of Grand Valley in a different series. It will always return to normal before the driver's next race (if he's driving in tandem with other Bobs in an endurance, he will recover this stat while not racing but it takes some time). As this stat gets low it will also start to deplete Mental Strength - presumably on the basis that the driver needs to concentrate more when tired.

Hard, close racing and lots of pace up commands will deplete Physical Strength (and Mental Strength) much more quickly than just driving. As both stats reduce during the race, it seems that his core stats (Braking, Accuracy, Cornering) are also reduce - though the nature of this change is hidden - and he becomes less able to drive. He is prone to making more errors and, eventually when both are depleted, he will become incapable of driving fully in the red. The lower the initial stats, the more quickly this occurs.

It also seems that Physical Strength governs a Bob's overtaking ability. This requires more testing for confirmation.

Character
The character and mood test revealed absolutely no difference in pace under any conditions between Bobs with functionally identical Braking, Cornering, Accuracy, Physical Strength and Mental Strength stats. Character is fundamentally unlinked to on-track pace:

Pace Down
D.Hoffman - Cool; Averages - L0 1'56.589, L10 1'54.330, L20 1'52.473
K.Snow - Cool; Averages - L0 1'56.602, L10 1'54.415, L20 1'52.614
H.Craft - Middle; Averages - L0 1'56.928, L10 1'54.334, L20 1'52.641
F.Roth - Middle; Averages - L0 1'56.956, L10 1'54.322, L20 1'52.591
B.Roth - Hot; Averages - L0 1'57.011, L10 1'54.356, L20 1'52.553
D.Solis - Hot; Averages - L0 1'57.284, L10 1'54.614, L20 1'52.777

L0 range 0.60%, L10 range 0.25%, L20 range 0.27%

Middle
D.Hoffman - Cool; Averages - L0 1'48.185, L10 1'47.148, L20 1'45.324
K.Snow - Cool; Averages - L0 1'47.975, L10 1'47.216, L20 1'45.433
H.Craft - Middle; Averages - L0 1'48.099, L10 1'47.173, L20 1'45.548
F.Roth - Middle; Averages - L0 1'48.222, L10 1'47.175, L20 1'45.569
B.Roth - Hot; Averages - L0 1'48.237, L10 1'47.314, L20 1'45.502
D.Solis - Hot; Averages - L0 1'47.976, L10 1'47.262, L20 1'45.572

L0 range 0.23%, L10 range 0.16%, L20 range 0.13%

Pace Up
K.Snow - Cool; Averages - L0 1'50.927, L10 1'42.891, L20 1'41.595
D.Hoffman - Cool; Averages - L0 1'51.776, L10 1'43.089, L20 1'41.881
H.Craft - Middle; Averages - L0 1'51.918, L10 1'42.765, L20 1'41.736
F.Roth - Middle; Averages - L0 1'52.224, L10 1'42.971, L20 1'41.749
B.Roth - Hot; Averages - L0 1'52.555, L10 1'43.071, L20 1'41.785
D.Solis - Hot; Averages - L0 1'51.161, L10 1'42.977, L20 1'41.656

L0 range 1.47%, L10 range 0.19%, L20 range 0.28%


I think those numbers - gathered over a thousand laps - pretty solidly confirm that the driver's Character does nothing at all for pace. What Character does seem to control is how quickly and how effectively Bob responds to your instructions (for his Mood) and conditions around him.

When immobile on the grid, cool and middle Bobs showed very little change in Mood, whereas hot Bobs would immediately start moving into the red before the race had even begun. Hot Bobs also responded much more rapidly to pace change instructions given, whether Up or Down, as well as becoming much more agitated much more quickly when in traffic. However, Hot Bobs also would change their mood quickly independently, with player instructions followed and then quickly ignored after the 10 seconds instruction cycle, according to their own present situation - even cooling down very quickly when not in traffic. Cool Bobs would react very slowly to player instruction, with commands followed for up to a minute, and their own situation.

Character thus controls the rate of change of the Mood bar according to its two inputs - driver situation and player instruction.

Hot Bob: Rapid rate of change of Mood; Personal situation much more significant in altering Mood than player instruction; Instructions habitually ignored after 10s; Need significant player attention to be effective.
Middle Bob: Average rate of change of Mood; Situation and instruction inputs weighted equally; Instructions followed for ~30s; Player does not need to babysit much.
Cool Bob: Very low rate of change of Mood; Player instruction much more significant in altering Mood than personal situation; Instructions habitually followed for up to a minute, requiring some babysitting to normalise pace after Up/Down commands to prevent them continuing to change Mood for too long.

"Condition"

I am not able to ascertain at this time what function Condition plays. In testing, no Condition resulted in a change of laptimes or behaviour. It is possible that a Bob in Good Condition is better at overtaking (suggested by observations in the B-Spec Sarthe race), but not necessarily certain.

 

[Famine]

Training Bob​

The character and mood test revealed absolutely no difference in training under any conditions between Bobs with functionally identical Braking, Cornering, Accuracy, Physical Strength, Mental Strength and Character stats. How the Bobs are trained is fundamentally unlinked to on-track pace:

Cool
D.Hoffman - Pace Down/Online; Averages - L0 1'56.589, L10 1'54.330, L20 1'52.473
K.Snow - Pace Down/Offline; Averages - L0 1'56.602, L10 1'54.415, L20 1'52.614

D.Hoffman - Middle/Online; Averages - L0 1'48.185, L10 1'47.148, L20 1'45.324
K.Snow - Middle/Offline; Averages - L0 1'47.975, L10 1'47.216, L20 1'45.433

D.Hoffman - Pace Up/Online; Averages - L0 1'51.776, L10 1'43.089, L20 1'41.881
K.Snow - Pace Up/Offline; Averages - L0 1'50.927, L10 1'42.891, L20 1'41.595

Middle
H.Craft - Pace Down/Offline; Averages - L0 1'56.928, L10 1'54.334, L20 1'52.641
F.Roth - Pace Down/Online; Averages - L0 1'56.956, L10 1'54.322, L20 1'52.591

H.Craft - Middle/Offline; Averages - L0 1'48.099, L10 1'47.173, L20 1'45.548
F.Roth - Middle/Online; Averages - L0 1'48.222, L10 1'47.175, L20 1'45.569

H.Craft - Pace Up/Offline; Averages - L0 1'51.918, L10 1'42.765, L20 1'41.736
F.Roth - Pace Up/Online; Averages - L0 1'52.224, L10 1'42.971, L20 1'41.749

Hot
B.Roth - Pace Down/Online; Averages - L0 1'57.011, L10 1'54.356, L20 1'52.553
D.Solis - Pace Down/Offline; Averages - L0 1'57.284, L10 1'54.614, L20 1'52.777

B.Roth - Middle/Online; Averages - L0 1'48.237, L10 1'47.314, L20 1'45.502
D.Solis - Middle/Offline; Averages - L0 1'47.976, L10 1'47.262, L20 1'45.572

B.Roth - Pace Up/Online; Averages - L0 1'52.555*, L10 1'43.071, L20 1'41.785
D.Solis - Pace Up/Offline; Averages - L0 1'51.161, L10 1'42.977, L20 1'41.656


What this means is that you can train Bobs as quickly as you like in any race and the resulting Bob will be no different from any other Bob with the same stat values trained in any other way. Just the stats, ma'am.

*B.Roth was prone, moreso than other drivers, to colossal crashes when Pace Upped and low on Strength. There is significant inconsistency with low level drivers running Pace Up, but the two Hot drivers end up directly comparable.

 

[Famine]

Remote Racing​

Please refer to this most excellent thread for all information about Remote Racing. There is little need for me to reinvent the wheel for this.

 

[Famine]

B-Spec Career Mode​

Although there's little to be said about this - it's quite hard to get wrong - there is an interesting snippet about how Bobs gain experience and some of the multiBob events.

It's a commonly held misconception that Bobs acquire levels at lower XP values than your overall B-Spec level (or A-Spec level), because they seem to level up quicker past level 25 than the overall level does. This is, as I say, a misconception.

In fact it's just a symptom of how Bobs acquire experience - at level 25 you unlock Endurances and, typically, Endurances are multi-Bob events. The listed XP offered for the event is how much you acquire for your B-Spec level and it's multiplied by the number of Bobs allowed then distributed over the Bobs according to how much of the race they drove.

This means that if you put one Bob into the Tsukuba 9hr endurance race - an event for four Bobs - you will get 1,933,494XP (including the 200% login bonus) for your B-Spec level but your Bob will get 7,733,976XP. This will increase a level 0 Bob to level 33 - whereas players would expect him to rise to level 28. Please see hubble618's XP thread.

 

[Famine]

Racing Bob​

(to be added)

 

[ElementRacer]

Sorry this is late.

The bottom line is, if I throw a level 0 driver into a 24 hour race and gain X amount of levels, the skill acquired is no different to if I leveled Bob up progressively, like you would your first driver?

 

[LeGeNd-1]

Sorry this is late.

The bottom line is, if I throw a level 0 driver into a 24 hour race and gain X amount of levels, the skill acquired is no different to if I leveled Bob up progressively, like you would your first driver?

There will be a difference, I tested this personally (nothing comprehensive like Famine, mind you). I maxxed out 1 Bob to level 40 with just 2 races (Nurb 24h + 4h with 200% bonus), and even though he has the same stats as my other level 40 Bob (with hundreds of races under his belt), he is a bit slower. Especially when he's overtaking or in traffic. Their raw pace is similar, but Bob's who have been through more races are more 'competent' at dealing with other cars.