# Is it Just Me or is S1 Road and Street Very Rare? I Actually Did the Math

So I’ve been playing FH4 for a while now and 90% of what I do is online Team Adventure (I like pain). One thing I have noticed during this time is just how rare S1 Road/Street races are. It’s gotten to a point where some days, it is actually faster to finish a full 5 race offroad game than to just find one S1 Road/Street. This is also a consistent theme and pretty much happens every time I play. The severity fluctuates, on a good day, it takes only around five minutes. On a bad day, I’ve actually sat for 45 minutes to find just one S1 Street race. I got so done with it I actually started recording every time I wanted to find an S1 Road/Street race and made a terrible video showing how impossible it is. You don’t have to watch it as I’ll break down my results down below, and the video is pretty bad. It’s just to serve as a basis for my quick maths that I attempted to do to prove that S1 Road/Street is harder to find than a needle in a haystack.

So, just to get things straight. There are a few types of races you can get in Team Adventure/Ranked's MM:

B: Dirt

So, assuming the matchmaker is fair and treating all races/classes as equal, there should be a roughly 9% chance to get each class and race (100/11), and roughly an 18% chance to get S2, 36% chance to get S1, 36% chance to get A, and 9% to get B. There should also be a roughly 27% chance of getting Road, Street or a Dirt race, and roughly a 9% chance of getting Crosscountry since Road, Street and Dirt appear in three out of the four classes while Crosscountry can only appear in two.

Now let’s compare this to what I got in the video.

In the second clip I got to 11 races (once again, including the sole S1). This time I got 0 S2 Road, 1 S2 Street, 2 S1 Road, 0 S1 Street, 2 S1 Dirt, 0 S1 Crosscountry, 0 A Road, 1 A Street, 0 A Dirt, 3 A Crosscountry, and 2 B Dirt. Percentage wise that is 0% for S2 Road, S1 Street, S1 Crosscountry, A Road, and A Dirt. 9% for S2 Street and A Street, 27% for A Crosscountry and 18% for S1 Road and B Dirt (this one doesn’t add up to 100% cause I did some heavy rounding on all the numbers, true for most of the calculations I did).

On the final clip. I went through 18 races (with S1 Street included). 3 S2 Road, 1 S2 Street, 0 S1 Road, 1 S1 Street, 2 S1 Dirt, 3 S1 Crosscountry, 3 A Road, 1 A Street, 0 A Dirt, 2 A Crosscountry, and 2 B Dirt. A 16.6% chance for S2 Road, S1 Crosscountry, and A Road. 11.1% chance for S1 Dirt, A Crosscountry, and B Dirt. 5.5%% chance for S2 Street, S1 Street, and A Street. 0% chance for S1 Road, and A Dirt. Class wise we get a 22% chance for S2, 33% chance for S1, 22% chance for A, and 11% chance for B.

So if we total up all three of the games I recorded. We get:
6.6% chance of getting S2 Road, 5% for S2 Street, and a general 11.6% for S2.
3.3% chance of getting S1 Road and Street, 11.6% for S1 Dirt, 10% for S1 Crosscountry, and a general 26% for S1
20% chance of getting A Road, 5% for A Street, 10% for A Dirt, 18.3% for A Crosscountry, and a general 53.3% for A
6.6% chance of getting B Dirt, and a general 6.6% of getting B.

So if we look at what we’re supposed to get above, 9% for each type, only the S1/A offroads achieve that, while the Road/Street/B gang fall below. We can also see that S1 Road and Street is ridiculously hard to get. Hell, B Dirt is just as common as S1 Road and Street COMBINED. A Road and Crosscountry also pop-up way too much. A in general is ridiculously overinflated. Remember, there should only be a 36% chance to get A, which is the same for S1 since they both have the most amount of possible races with 4 variations. Yet, A is 17% more than the supposed average and 27% more than the supposedly equal S1.

Now, if we also divided them by race type (Road, Street, Dirt, and Crosscountry), we see that there is a 29% chance of getting Road, 13.3% chance of getting Street, 28.2% chance for Dirt, and 28.3% for Crosscountry (again, not a perfect 100% due to rounding). We can clearly see that Street is severely neglected, despite the fact that there are 3 possible classes to get Street with (S2, S1, and A), which is tied with Road and Dirt, and more than Crosscountry (which can only appear in S1 and A).

So what do you make of all of this? Firstly, it’s hard to say this MM is “random”. Despite what I said in the video (I did say it was bad), B and S2 are not over-represented, but under, and A and S1 are over-represented and most likely receiving some sort of boost in the MM or getting heavily influenced due to other game factors (for example, if more people quit mid-game in A and S1, then there might be more open A and S1 games, however, in my experience, this is not the case and people quit in every class equally). Street is also receiving some kind of penalty in the MM cause frankly an 11.7% difference from what they’re supposed to get (25%) is beyond the margin of error. And dear god I don’t know what in the world is going on with S1 Street and Road to make it appear nearly 3x less than it’s supposed to. Your guess is as good as mine.

Now obviously, I could just be paranoid and there is nothing wrong with the MM. However, the end results still speak for themselves and I am definitely not the first person to notice that FH4 gives a lot more offroad races than on-road races. The time it takes to also get an S1 Road/Street should also be taken into consideration. On average of the three videos, it takes 10 minutes and 20 races just to find one S1 Road/Street. It is very easy to actually finish an S1 Road/Street race quicker than you can find one. So in the end, I’m curious, did anybody else get results similar to mine? Is S1 Road/Street insanely hard to find and all you ever get is A class? Because I constantly hear people complain about the MM but I don’t ever see much proof. If enough people are interested and I have absolutely nothing to do, I might even just start recording every single time I go looking for a race to get a better and larger sample data.

TLDR: Yes, S1 Road and Street is unusually hard to find and is most likely the reason why offroad appears more often than on-road races.

Thank you for coming to my Ted Talk.

Thanks for doing the math, it seems to match up with my much more limited observations (one adventure per month for the playlist). I’ve been annoyed at how many times offroad events have shown up, largely because they takes twice as long to complete as the others with the awful freeroam rushes nearly doubling the number of events I have to sit through.

Seems like there’s someone at PG who is obsessed with offroad. Not that I’m against offroad (dirt is my favorite race type) but the CC tracks are almost universally awful (especially at higher classes) with all the unrealistic jumps and wreckage (and trees), and freeroam rush suffers from many of the CC handling issues (especially since tuning between dirt and CC vehicles is quite different in most cases) and even more trees.

yeah i only get cc from s1

I’m not sure if the number of observations is giving a statistically significant result. I’ve done one adventure a month for the last 4 months and have had road or street every time, and have had 1 x A, 1 x S1 and 2 x S2.

One way that you can look into statistical significance is by doing a Monte Carlo simulation in Excel. If you google, you’ll find numerous resources telling you how to do this, e.g.

So what you would do is build a simulation that randomly chooses from the 11 options with equal probability for each one, and does it the same number of times as you did it. You’d then repeat this many times to get an idea of what these equal probability outcomes look like in practice. People tend to have an incorrect idea of what true random behaviour looks like. For example, if they try to randomly pick lottery numbers, they avoid adjacent numbers when in fact adjacent numbers quite commonly occur in true random selections.

It’s possible to do statistical analysis to work out confidence intervals and statistical significance directly using equations, but it’s much harder to teach someone how to do that because it requires more understanding of the characteristics of the data, and there are so many pitfalls to learns about. Monte Carlo simulation is much easier to grasp and apply correctly with minimal knowledge and understanding.

1 Like

I tend to get CC most of the time, but I do get S1 and A class as well as B rally. I do notice that there#s a reduction in winter racing based on my minimum 10 team adventures per week.

I haven’t noticed S1 road/street being particularly rare when compared to A and S2 road/street. I haven’t been tracking but my feeling is that they come up fairly evenly for me.

I do wonder though if the aim is to split evenly between cross country, dirt and road/street (ie 1/3rd each between those categories). That could explain why the two cross country events appear to occur a lot more than frequently than other events.

Quite possibly. I’ve been doing street races in ranked now because if I just wait for A and S1 road only, they’re just hardly ever coming up. I even did an S2 road yesterday which I would normally avoid like the plague. I actually did okay and won it surprisingly, with T10’s JONK1969 in 2nd. But it’s too fast for me to ever really feel like I’m fully in control.

Online racing is popular because the unranked lobbies are frequently quick to get into and often full. But ranked is a ghost town, especially before the evening European time. (Yes I know people work during the day etc. but right now millions of adults are furloughed and millions of kids are off school but it’s still a ghost town).