It's not hard to make educated guesses. 100 lobbies with 8 people in them for an average of one hour is 19,200 visits per day, with some of them being multiple visits from a single person. An average of a half hour gets you 38,400 players etc. Given that many of them are probably repeat visits from avid players, the numbers are likely much lower.
We also know that the participation rates in Seasonal events for GT5 hovered around 80-120k for most seasonals which represents from about .66 - 1% of the user base. That's total participation over the life of the Seasonal. Doesn't matter how you slice it, the numbers are nowhere near big enough to justify characterizing online as anything but a niche' part of the game and nowhere near the norm.
You didn't read it fully?
Ok. Lets work with the assumption that GT5, 2 years after release and 10mi sales, has 120k unique users playing and doing an online activity.
What is its retention rate? Is 120k above or below average for games? What about racing games?
What is retention rate anyway?
Retention rate is the percentage of a school's first-time, first-year undergraduate students who continue at that school the next year. For example, a student who studies full-time in the fall semester and keeps on studying in the program in the next fall semester is counted in this rate.
It works the same for games. Developers expect that a number of people that buy their game is going to drop it after a certain period. For exemple, they'd have a nice graphic with accounts that downloaded updates.
They would know: "Hey, after 2 years, X users are still updating."
That is the general "playing the game" number. If the game communicates with a server even if you are not participating in an online activity such as the Seasonals, they would have a number of unique users monthly/weekly/daily: "Hey Y unique players turn the game on every day".
That's more representative of a game's health than units sold in terms of USER RETENTION.
And yet we have nothing that allows us to say: No the majority of players that still load the game, play only the offline portion. Or the reverse.
So even if AC sold 300k, and we can do a napkin 'educated' math of 20k unique users still accessing the game once a month, we cannot say what percentage participates in online activities such as PvPs or just sit around doing hotlaps. We don't have these metrics.
We don't know what the average user lifetime for racing genre games, and I haven't come across breakdown of feature uses.
A recent exemple of a fully online game. Star Wars BF.
Sold 13mi+ copies.
In a day it's peak is 50k with the hourly average at ~20k. If we draw the ceiling at every hour having different users than other, that would give us 480k unique players per day. We can go a step further in imagination and say that every day in a week sees different players. That would make 3.3mi unique players.
Of course, these numbers are far from reality.
But still on that exercise it would mean that less than 25% of people that bought the game that is exclusively online still play it.
To then 13mi+ copies is meaningful in terms of making bank, but in content deployment, server management and game environment health, it's meaningless.
While observing how games in almost every genre structured themselves to present online and multiplayer capabilities is an indicator that yes, that online presence is the normal. Multiplayer is the normal.
And again, if you say "but in the simulator racing genre!", it begs more information or a frame of reference. Which you cannot provide.
Four points made and every single one of them incorrect.
- Its nose to tail yes, but not as the result of a late brake (but I guess that's your psychic ability at play again).
- At least once during the preceding straight I am ahead of him and I hit the apex first (as your own gif shows)
- Back with the psychic powers again.
- After the curve at least a car width is between us when he leaves the track (as your own gif shows)
So quite frankly you now seem to have resorted to mind reading and sheer fantasy to try and prove a point.
1) You can switch to the AI going for an "earlier brake point" then because it's clearwhen the nose of his car dives and when you engage the brake to avoid collision.
I admit that either way (you going for a later brake than him, or him going for an earlier) I don't know which lap it was for the state of the tires or even the optimal brake point for that car.
Either way that's a close to contact situation. All right...
2) I see the moment you meant, but nope, camera angle tricked you.
Aligned at straight.
His rear axle slightly ahead at near entry.
Still slightly ahead at entry.
3) I suppose you are talking about the intent to overtake? There were only three phrases.
4) It's a car space in relation to the white line that determines the track and at curve exit.
Even though you are still wrong, I'll repeat myself about the crux of the issue.
The BMW at Snetterton shows what users do to AIs and they have to cope with that.
At the curve start, the yellow car who is following the normal line (outside, inside, outside)
The AI seemingly understands he's ahead of you (or is not aware of you until he tries to close in) because he positions himself to take the inside towards the apex, but once he notices you he swerves back and goes off-track.
The video shows you attacking the AI from a position where you simply have no gains other than forcing him out of bounds. Since you hold your pace, you throw the decision to him of avoiding you. That's textbook scenario for complaints of AI cutting or crashing into (although in this situation he avoids without stepping of the gas).
Then he barely loses distance after his grass arc, meeting you close on his return. Which again he has to correct since you are doing your line close to the edge. I said that at this point he'd be wrong to come in at that speed, but "defensive driving" tells you to be aware of those around you at all times.
Gotta pass around that responsibility.
Players take this constant aggressive behavior because there are minimal to no risks or consequences.
That's why you can ABUSE the AI opponents.
So you will have no problem at all citing the warnings and infractions they received.
Me
Was he flagged or punished? Probably not, interpreting all that is up to the officials
Now the key bits you seem to have missed are that a maneuver needs to "hinder" another driver. Now given that I was behind him (and had no other driver directly behind me) please explain how my change of direction to the inside on the straight hindered him?
The second key point is "deliberate crowding of a car beyond the edge of the track", given that I was six foot away when he left the track that could not be described as either deliberate or crowding.
So I ask once again, which part of the sporting regulations does it fall foul of?
I quite happily acknowledge is an aggressive move, but you are the one who was quite happy to infer it was an actionable maneuver. If that was not the case then you would not have used terms such as "
Til there officials would be "all right...", which would indicate that the part following would not be 'all alright'.
The key bits you missed are "liable to" and "such as".
Liable to hinder doesn't entail
hindering, it entails the "possibility to hinder".
And "such as" denotes the exemplification. "Deliberate crowding of a car" is the exemple of a "maneuver that is liable to hinder".
I thought you spoke and read english!
And yes, keeping a line where you'll visibly not produce an advantage except if you force the action of avoiding contact to the opponent is "liable to hinder" him.
---
But perhaps you'd exit the curve faster than him... Although again, his grass arc doesn't tell the same story. But I'll be fair and say that's just funky AI tire properties.
And I was going to say that at 0:22-23 it still looked like you moved towards the track edge to block him on his return, but after much freeze-framing and finger measuring, it's the camera shake and wheel alignment that threw me off.
You just kept going straight on the outside.
AI reaction is spontaneous, its reaction to what the driver is doing and can't know what the driver is intending to do in advance (unless its acquired your psychic abilities).
???
First "AI reaction" is not spontaneous, it's deterministic and it follows a cascade of decision processes and hierarchies/priorities.
It can be "emergent", but still follows the same principle in general. That's why they are so expensive to compute, and games have rather simple decision making processes.
But still that misses what I wrote by a mile.
Risk aversion has to be programmed onto AIs, IT's not some spontaneous reaction.
The same above follows though.
A basic made up exemple of a priority list and decision cascade.
Priority order: 1) Keep line 2) Keep speed 3) Avoid contact 4) Stay inbound.
He'd go: Can I keep in line? Yes? No?
If yes, Can I keep speed? Yes? No?
If no, returns to "keep in line?"
If no, "Can I keep speed?"
If yes "Can I avoid contact?"
And so on in the space-time point he has to take that decision til it breaks the chain.
You can write something that will heavily prioritize "Avoid Contact" and then the following action will always take the measure to Avoid Contact.
That's why Risk Aversion is not spontaneous, this process is established by the programmer.
But you probably knew that, right?
Now I have no intention of attempting to read you mind, so what I would love to know and see is how you would overtake, and given your stated criteria I don't expect to see it being on the inside line (despite that being perfectly legal), as the result of you driving alongside the car on the straight before the corner, impeding the other drivers line at all or coming within six foot of the other car once you have passed the apex.
I look forward to seeing the video soon.
I'd likely just try to do a faster line on the outside since the big straight (where acceleration can produce ground gain...) is right after the curve you took the inside with a softer curve following that AIs tend to blip brake.
But I'm not little Senna, I'm just little Rubinho.
