Average A-spec points

[Der Alta]

Average A-spec points

A-Spec and B-Spec mean nothing to me. Seriously, at 30% thru the game, I haven’t used B-Spec at all. As it is, the A-Spec points are all over the place. I can post an easy 120 point win in one race, and fight like hell to win an 8 point race elsewhere.

With the broad range of A-Spec points available, noone really cares that you’ve managed to compile 10,000 A-Spec points. No one really cares that you’ve done 5 races or 500 races to obtain those points.

An easy way to rank yourself, and compare to others would be a listing of Average A-Spec points. For instance, If I were to post up my average as 78, and Famine posts his as 158, we’d have a very clear indication of how much better he is than I.

So, my thought would be to have PD include this average to help you compare to others when Gt5 comes about. Racing On-line would be much fairer (yes, some will sandbag the races) and you’d have a much more enjoyable experience when racing against others in real time.

Of course, this is a moot point if PD included that already, and I haven’t found it yet.

 

[DeLoreanBrown]

👍^^^👍 der alta

i was considering something similar as a formula in 4, it would help for the 'constrained' but Serious User.

 

[Dave A]

I agree, an average a-spec points would be a great addition in thoes terms. Having 10k points as you said, is pointless, you could have done them in only a few races and the mission hall, or it could have taken every race in the game.

 

[Der Alta]

I seriously wish that PD had done something with them. As it stands now, they're all but useless. Would it have taken another 2 hours to code the Points per Races and make a display for it?

 

[Tenacious D]

Hopefully A-Spec and B-Spec points (if included) will be sensibly calculated in GT5 so that won't be much of an issue. But to be sure, it's just a few lines of code to calculate it, paint a location and display it.

I wouldn't doubt that since Gran Turismo is all Polyphony does, it takes its criticism pretty seriously, and with such a super system, PD will give us a true Gran Turismo vision.

 

[Dave A]

PD do plenty of other titles, they used to do Motor toon GP, they do Z.O.E which Kaz wan't to make another one of, and now they do Tourist trophy as well amongst othet titles they've developed. Gran Turismo is just their hit series, the one they focus on the most.

 

[Tenacious D]

Darn these long hours. It only took me about a day to even forget that PD has that Superbike games in it's list. I need to become independently wealthy so that if I'm up till 3 am or so, it's because I'm writing music or fiction.

 

[SportWagon]

Computing average A-spec points isn't as simple as taking the current total number of A-spec points and dividing by either the number of races done or the number of A-spec races done.

Currently you "win the A-spec points only once". So if you've done a 2CV race for 94pts, if you do it again for 104pts it's in some sense like only a 10point race. To compute a meaningful average A-spec point value they'd need to keep a separate counter of "true total A-spec points", in which the total would be added each time, not only the "increment". If they didn't do that, they'd discourage people from re-entering races. (Because once you'd won the maximum possible A-spec points for an event, all you'd be doing is increasing the number of races (the denominator) and therefore lowering your average).

If the game were to count each event a maximum of once, (i.e. "done" or "not done") that only provides extra meaningful comparison between two players who have done vastly different proportions of the races. That is, as the percentage complete approaches 100, all you're doing is scaling the A-spec total as it is currently implemented. That is, if two players have 100% (or close to it) their relative A-spec scores (leaving aside quibbles about Dodge Rams and Toyota Prii) are currently fairly meaningful. Each must have at some time completed relatively equivalent races to get the same total. Now, true, one might have immediately racked up the score and the other gradually inched it up increments with successive attempts. And yes, the extra computed average (i.e. using the "true total") would indicate that difference, but an average computed just on the basis of the number of completed events would not.

This is leaving aside the fact that not all races with the same value are really equivalently difficult. And, yes, ignoring the anomalous car entries like the Dodge Ram and Toyota Prius.

So a true "A-spec total" and "A-spec average" could be useful, as long as they did get rid of those wildly anomalous cars. (Otherwise you could artificially inflate your total by repeating bogus 200pt races). (And they'd need to do something about 250pt missions too; either (probably) count only once, or perhaps count as 60 instead (or perhaps a range from about 10 to 250 depending on the mission)).

Currently you can compute some sort of average by dividing the A-spec total by the percentage completed.

E.g. suppose you have two players, one at 100% with 20,000 A-spec points, and the other at 20% with 5,000 A-spec points.

20,000/100 = 200 points per percent
5,000/20 = 250 points per percent (sic)

I.e. the player at 20% appears "better". I.e. they appear on track to get 25,000 points at 100% instead of 20,000. Okay, so I guess they could have easily added that calculation to the status display. It's rather difficult to succinctly describe or label, however.

 

[Dave A]

To put it simply, you have an extra counter that adds up all the points you've ever got (so if you do the same race twice you get the seconds race's points added in full) and divide that by the number of race win's you have. That gives you a true average points per race.

 

[SportWagon]

Yes, but they went to a lot of trouble to not count the A-spec points twice, assuming the comparison of totals would be sufficient, since your total indicates the sum of the best results you can achieve. So to have added the extra total would have been contrary to the notion they had for A-spec points. It would mean that if you learned to be better later, you would need to start a new game in order to be able to have a good average.

They would have needed to consider the possibility of overflow, but, true that's not really a concern. They should easily be able to allocated storage to allow for an A-spec total for up to almost 10,000,000 events--far more than anyone's going to do.

But before considering adding such extra calculations, it would be far more important for them to improve the calculation of the A-spec points for each event (avoiding artificially lowering the A-spec points offered with weak also-ran cars, and abolishing the Rams and Prii anomalies).

 

[Dave A]

It's not hard to simply create two seperate counters, one with only the best points each race as it is, and one with all the points ever totalled. That's all they'd need to do.

 

[Hun200kmh]

I agree with SportWagon. The "A-Spec average" is very simple to obtain in GT4, just using this formula:

A-Spec points/ game completion (%) = A-Spec average

The only flaw here is that we will never know how many times each guy entered a race before he got the most A-spec points he could possibly win. But it isn't really a flaw, if we don't care about the learning curve, and just want to know how much that guy was able to achieve.

(besides, if we don't use the save/load scheme we'll know the win/lose ratio, and that's also a clear indicator of the history behind the "a-spec average"!)