My Diaries

A Lurid Tale of Obsession, Depravity, Wits and Attempted Wit

Sat Jun 12, 1999

Trying to Express My Thoughts about Race-Series Prize-Cars

In the Normal and Tuned Series I think the cars are awarded a bit more evenly than true random distribution would give. It only took me 10 wins, which would be on the low side of expectations if the cars were truly awarded randomly. (Try rolling a die, and see how long it usually takes you to get all six numbers; this morning I took 19,13,13,17).

According to a semi-brute-force program I wrote, you have only the following chance of getting all six on or before:

And after that, I need to work around an overflow problem before I can report further. (Or remember some less brute-force probability theory). The above intuitively makes sense to me, but I may have made a mistake. As you get more prize cars (or die rolls), it actually becomes more likely you will get ones you already have, making it quite difficult to actually collect all of them.

And I seemed to win the last car after I took the new car I had been using (Viper RT/10) and racked up some other series wins for it.

In order:

Apart from the Viper, all other cars were quite old and had somewhat extensive racing histories.

A car which has been winningly raced in a lot of other series seems to have a better than random chance of getting a new prize car in a "target" series. I suspect records of these wins are stored in the "shared" portion of the cars records, ie with the "lifetime" supply of parts, since it seems cars in which another "instance" has done well will also do well.

The Normal series came off even faster, with 8 wins (i.e only two duplicates). I don't have detailed notes here, but I won with Viper GTS several times, then lost with a Griffith 500 at one point, and won a duplicate car after that with the Griffith 500, and won the final car first time out with a Griffith 4.0, after preparing in Sunday Cup and Clubman and FR. (Race-modded Griffiths of both types had previously raced extensively in vain, hunting for the elusive yellow ConceptCar).

In contrast, some of the "International Series" prize cars seem really "locked out" (or others "locked in"!) for a long time. In a "test" game a prize ConceptCar, not raced in any other series, won a white/blue Viper GTS-R in 5 out of 5 (not all consecutive) USvsJP entries, while a race-mod Viper and a race-mod GTO'95MR both turned up black FTO's first time. That Viper first won the yellow ConceptCar for me, in that game, after I first won all of GT league (including GT-I) with it, plus FR, and then won UKvsUS three times, getting three different cars. So, later I tried doing the same thing with a race-mod Viper in my "main" game, and, sure enough, a yellow ConceptCar was finally won there!

I cannot believe this is all due to faulty use of faulty random number generation routines, as is suggested by at least one FAQ.

I know for a fact that racing models are not penalized, prize-car-wise, for entering non-racing series. (This is, however, about the only definite fact I can state).

I'll post about this to r.g.v.s soon.

Here is the same table for 4 cars (or 4-sided dice)

If prize cars were awarded randomly, and you entered an International Series 15 times, you would have a 95% chance of getting all four cars by that time (and a 5% chance of not doing!)

Here's the numbers for 2 car series.

I.e. by 6 or 7 tries, you really should have got both cars. But I seemed to get both cars in two tries in all both of those races. Which the above says is somewhat unlikely.

(Though the probability of it happening is 1 in 4, so it's not very unlikely).

(This table's easy; # of combinations is 2^n, and there are only two combinations at any n which do not include both, namely all the first, or all the second).