It's funny how the people defending Onlive fail again and again to understand what's input lag, how it works, why it's bad and how it's different from regular lag.
And it's even funnier how those people immediately proved my above statement to be 100% true.
EDIT: Oh and the Netbook thing: Don't buy a netbook if you want to play games newer than X-Com and CivII and Majesty 1. They are not made for gaming. It's like buying a DVD player and complaining that it doesn't run Halo 3.
You guys are doing a lot of bashing on something before it's even come out/been used. My suggestion is to see how it is when it is released before condemning it instantly.
If someone announced a perpetual motion machine that would revolutionize how energy is produced, you better damn well believe I'd bash on it before seeing it.
This is the same category - they are making claims that trump basic physics, so it gets bashed until proven otherwise.
How does playing a game over cloud computing (Something that already exists.) trump basic physics? Have you guys thought that maybe the box they give you to access it might have a bit of hardware acceleration/software on it that helps with reducing lag? Sorta like how you can get better adapters for your PC so your PC's connection has less lag?
There is a fundamental difference between input lag and network lag. This has been said many times already in this thread, and it is key point as to why people are rightfully scoffing at this.
Yes which is why i said something about the hardware. You seemed to have missed that part though. Like i said, you guys are all bashing on something you as of yet, know basically nothing about. So it's all just a lot of pulling things out of asses at this point. I myself am excited about the future prospects of such a service/device and am looking forward to see how it works out, for better or worse. I tend to not formulate opinions on things i haven't even given a chance yet.
Hardware cannot compensate for input lag. That is the fundamental problem, and why people are saying attempts "compensate" along those lines are akin to solutions for a perpetual motion machine. There is no way to compensate for input lag unless you very literally have a time machine of some form, because user input is not predictable. The issue of specialised "hardware" doesn't come into it.
The scenario is this:
1) Player sees visual output
2) Player responds to visual output with action
At this stage you've got the start of your input lag
3) Player input travels across the internet to server location, which takes x miliseconds.
4) Instruction is accepted and executed, which happens effectively instantaneously.
5) Resultant image is rendered, formatted, and compressed prior to sending. Given promises on their front, I'm willing to accept this will take a negligible amount of time as well (I'm doubtful that's true, but I'm willing to accept it for now).
6) Image feed is streamed back to the player over the internet.
7) Player sees results of input they provided, responds further.
Therein ends the input lag. In-between phases 2 and 7 you've got a gap where the user has provided input, but the simulation is still running without knowing about what the user input is. In actual terms this means the user is effectively giving input several hundred milliseconds behind where the simulation actually is. The simulation is responding to instructions that are based on where the simulation was a few hundred milliseconds ago.
The real question is how significant this input lag will end up being, and whether it actually ends up affecting the gameplay in real terms. But you can't "compensate" for input lag. It's already obvious that for some games even severe input lag will still be functionally playable. I mean what's 1/3rd of a second or more when you're playing King's Bounty? In others, particularly reflex based games, it can make all the difference. I mean, I cannot imagine playing even a game like Super Mario World with a significant input lag, that would pretty much destroy the whole thing. FPS's might be OK (depends on the game and the extent of the input lag), but fighting games are out. Similar with RTS's depending on gameplay style, games that are heavily micro dependant wouldn't be nearly as viable. And any semblance of online play is pretty much out the window because you're now placing input lag on top of the regular internet communication lag you'll be having (unless again, the game is turn based).
EDIT: And this isn't even me saying that OnLive is going to fail, I'm one of the few who actually believes it might have a shot if it starts off slow and tries to remain around until such a point where the idea is a bit more viable (expanding to other services would be a good start). But you can't just will the potential lag issues away, those aren't in OnLive's hands, not completely. Especially relevant on a service that costs so much on a monthly subscription on top of buying and renting the games for usage.
It would be awesome if they partnered with GoG, made a good dosbox or VM setup and made a system to do succession games with it. All those old games with hotseat capability suddenly get playable again. Especially if only one of the participants has to own the game in order to play it in a group, you know, just like it used to be for hotseat games? Older games have ideal resolutions to play on an Iphone, except that you have the fat-finger problem.
Sorry, I didn't mean that to be a sleight or anything. It's just that so far these carefully managed stage presentations are all we've seen, and anything else to do with the actual trials has been under NDA so we've got nothing about how well the system actually works. So I kind of got bored of watching them "showcase" OnLive again and at this stage I'm waiting for the real deal to come out so we can have a proper look at it.
Sorry, I didn't mean that to be a sleight or anything. It's just that so far these carefully managed stage presentations are all we've seen, and anything else to do with the actual trials has been under NDA so we've got nothing about how well the system actually works. So I kind of got bored of watching them "showcase" OnLive again and at this stage I'm waiting for the real deal to come out so we can have a proper look at it.
fair enough
if it even comes close to the 80ms lag they mention i am so getting this
Sorry, I didn't mean that to be a sleight or anything. It's just that so far these carefully managed stage presentations are all we've seen, and anything else to do with the actual trials has been under NDA so we've got nothing about how well the system actually works. So I kind of got bored of watching them "showcase" OnLive again and at this stage I'm waiting for the real deal to come out so we can have a proper look at it.
fair enough
if it even comes close to the 80ms lag they mention i am so getting this
80ms is nothing. Like i said as long as im not trying to compete online and just dicking around, or playing a single player game, then this will be great fun.
And besides, isn't 80ms the top-end goal they are reaching for? I understand the input lag and the internet lag and speed of light yadayada, but it's not like these guys are idiots. They're working with ISP providers to get how the crow flies connections and priorities. It's not like they don't understand speed of light limitations or anything.
Sorry, I didn't mean that to be a sleight or anything. It's just that so far these carefully managed stage presentations are all we've seen, and anything else to do with the actual trials has been under NDA so we've got nothing about how well the system actually works. So I kind of got bored of watching them "showcase" OnLive again and at this stage I'm waiting for the real deal to come out so we can have a proper look at it.
fair enough
if it even comes close to the 80ms lag they mention i am so getting this
80ms is nothing. Like i said as long as im not trying to compete online and just dicking around, or playing a single player game, then this will be great fun.
As long as the single player game doesn't require precise timing or fast reactions. So no first person shooters, no Bayonetta, no fighting games, no rhythm games...
So there's no way to code/adapt 80ms input lag on the server side? I guess it'd only need to be 40ms (one way) but still. Just modify a rhythm or button mashing game to higher tolerances of input lag. Is that more difficult than it sounds?
So there's no way to code/adapt 80ms input lag on the server side? I guess it'd only need to be 40ms (one way) but still. Just modify a rhythm or button mashing game to higher tolerances of input lag. Is that more difficult than it sounds?
There are ways of compensating for input lag, yes.
But with respect to rhythm games (or anything that requires precise timing), the problem is not solely with the fact that there is perceptible input lag but the fact that this lag is also variable. From moment to moment as you are playing the game, the amount of input lag for any given action is subject to your ISP's service, traffic on the routers between you and Onlive, congestion on Onlive's servers, interference with your wireless internet (if you use it). You can't compensate for input lag that could change with any input.
So there's no way to code/adapt 80ms input lag on the server side? I guess it'd only need to be 40ms (one way) but still. Just modify a rhythm or button mashing game to higher tolerances of input lag. Is that more difficult than it sounds?
It is. At that point you're actually talking about recoding the game itself, but even then, it's not a question of tolerances as such. To take Street Fighter 4 as an example, presumably you're still going to be able to do a Hadoken regardless, as long as you input the command properly (this is presuming your latency is constant throughout, and not varying, which is another fairly big hurdle). The issue is that it'll be coming out x hundred milliseconds later than you wanted it to. There's no way to account for that because there's no way for the game to intelligently decide "OK he's going to want to do a Hadoken at this point" and execute it for you before it's received your actual input. I mean, I'm just learning SF4, I'm a complete novice at it and I still whiff the occasional special, and I regularly mess up combos. Heck on rare occasions if the pressure's gotten to them, even the pro players do. Add in input lag on top of that and your likelihood of that happening increases a LOT all of a sudden.
And I mean, Street Fighter 4 is actually pretty dull to play singleplayer, it's all about playing with other people. Which is why it becomes an issue for games like that, even with as "minor" as an 80ms input lag, that seriously screws up the timing a ridiculous amount.
For most other games, I'm not so sure 80ms input lag is going to be an issue. But frankly, I'm seriously doubtful they're going to achieve that kind of input lag.
So there's no way to code/adapt 80ms input lag on the server side? I guess it'd only need to be 40ms (one way) but still. Just modify a rhythm or button mashing game to higher tolerances of input lag. Is that more difficult than it sounds?
It would still be 80ms because that's the amount of time between when the visual stimulus is created and when the input feedback is received.
You are also saying that one of the base principles of the system, that the games will "just work" with OnLive is wrong.
The reason I don't care for the service is that it is the *height* of DRM. If it picks up a little bit, you will never be able to own a game ever again. You thought Ubisoft was getting draconian with their latest DRM exploits? Imagine how those corporate suits must be salivating at the thought that not only can they ensure that no one pirates the game BUT they can make you pay monthly for the privilege of playing it.
jonxp on
Every time you write parallel fifths, Bach kills a kitten.
3DS Friend Code: 2707-1614-5576 PAX Prime 2014 Buttoneering!
The reason I don't care for the service is that it is the *height* of DRM. If it picks up a little bit, you will never be able to own a game ever again. You thought Ubisoft was getting draconian with their latest DRM exploits? Imagine how those corporate suits must be salivating at the thought that not only can they ensure that no one pirates the game BUT they can make you pay monthly for the privilege of playing it.
It pretty much is the golden egg laying goose yeah. Once a month, ka-ching! Although to be honest I'm not sure such a system, even if it takes off, will really replace people owning consoles and games. Far more problematic that singleplayer is multiplayer, to the point where I don't think it'll be viable over such a service. And without that, well you've immediately got problems.
So there's no way to code/adapt 80ms input lag on the server side? I guess it'd only need to be 40ms (one way) but still. Just modify a rhythm or button mashing game to higher tolerances of input lag. Is that more difficult than it sounds?
It would still be 80ms because that's the amount of time between when the visual stimulus is created and when the input feedback is received.
You are also saying that one of the base principles of the system, that the games will "just work" with OnLive is wrong.
The reason I don't care for the service is that it is the *height* of DRM. If it picks up a little bit, you will never be able to own a game ever again. You thought Ubisoft was getting draconian with their latest DRM exploits? Imagine how those corporate suits must be salivating at the thought that not only can they ensure that no one pirates the game BUT they can make you pay monthly for the privilege of playing it.
I agree with the DRM part. However, I'm still holding hope for a rental service of some sort. Pay X amount per month, get access to Y number of games. Or even X amount per month for unlimited rentals. It would be like netflix streaming except for games. Which I hope for, and is where I see this system shining. Will it happen? I don't know.
They have talked about the "OnLive Game Portal", which will be different from their $15 per month regular service. You can pay per-game to rent (and play free demos) instead.
RandomEngy on
Profile -> Signature Settings -> Hide signatures always. Then you don't have to read this worthless text anymore.
So there's no way to code/adapt 80ms input lag on the server side? I guess it'd only need to be 40ms (one way) but still. Just modify a rhythm or button mashing game to higher tolerances of input lag. Is that more difficult than it sounds?
It would still be 80ms because that's the amount of time between when the visual stimulus is created and when the input feedback is received.
You are also saying that one of the base principles of the system, that the games will "just work" with OnLive is wrong.
The reason I don't care for the service is that it is the *height* of DRM. If it picks up a little bit, you will never be able to own a game ever again. You thought Ubisoft was getting draconian with their latest DRM exploits? Imagine how those corporate suits must be salivating at the thought that not only can they ensure that no one pirates the game BUT they can make you pay monthly for the privilege of playing it.
I agree with the DRM part. However, I'm still holding hope for a rental service of some sort. Pay X amount per month, get access to Y number of games. Or even X amount per month for unlimited rentals. It would be like netflix streaming except for games. Which I hope for, and is where I see this system shining. Will it happen? I don't know.
Sounds like Gametap to be honest. Only without the streaming.
subedii on
0
TetraNitroCubaneThe DjinneratorAt the bottom of a bottleRegistered Userregular
So there's no way to code/adapt 80ms input lag on the server side? I guess it'd only need to be 40ms (one way) but still. Just modify a rhythm or button mashing game to higher tolerances of input lag. Is that more difficult than it sounds?
It would still be 80ms because that's the amount of time between when the visual stimulus is created and when the input feedback is received.
You are also saying that one of the base principles of the system, that the games will "just work" with OnLive is wrong.
The reason I don't care for the service is that it is the *height* of DRM. If it picks up a little bit, you will never be able to own a game ever again. You thought Ubisoft was getting draconian with their latest DRM exploits? Imagine how those corporate suits must be salivating at the thought that not only can they ensure that no one pirates the game BUT they can make you pay monthly for the privilege of playing it.
The DRM part is exactly what I find worrisome. Developers and publishers are probably thinking this is where they need to move. Not only do they avoid the possibility of piracy, but no one, every, anywhere will be able to get ahold of the game code (barring some internal OnLive / Developer leak). That means that there's no possibility for leaks, no possibility of resale, no possibility of modding. They'd avoid people stealing their game (or even theoretical people stealing their game, which some devs like Crytek baaaaaw over constantly), they'd be able to charge much higher prices further after release, and they'd never find their game on the 11 o'clock news because some horny teenager made a 'nude patch' that was never part of the intended release. The big hook here is that this form of DRM provides an exceptionally appealing convenience to a large audience - the ability to play current games without putting down money for a computer. Alone the DRM would never work, but with that kind of added benefit a lot of players would be more than willing to swallow the negatives.
To be clear, I think it's a disturbing trend to put consumers in a situation where they pay for everything and own nothing. I like being able to mod my games, and I like being able to tweak settings to my preference (They've confirmed that there will be NO possibility of modifying your graphical options with OnLive, which you might expect by the nature of the service). But if this service takes off, and some big third party title makes a ton of money on an OnLive exclusive release? You can bet the other publishers and developers will be beating down the door.
But, that's all theoretical, assuming the technical difficulties inherent to the service don't hamstring the entire idea.
Hey, you know what I'm in the mood for? A crackpot theory. Here goes!
Ok, so this whole latency compensation thing is basically impossible unless they've done one of two things:
• Invented a time machine
• Invented a way to read your mind
The first I can't think of any way to fake, but the second... Hear me out. What if, for each game, OnLive has taken huge recorded input dumps of hundreds of test players, crunched the numbers for timing between movements, directions that follow directions, pacing, reactions, and so on, and made a highly sophisticated heuristic profile that records your input patterns for the first few seconds of a game, runs vast, complicated algorithms on the sample to match you to the overall game profile, and then uses this data to constantly predict your next twitch 80ms in advance.
The above is completely insane, highly unlikely, and the only way I can imagine OnLive "compensating" for input lag.
There's not 'no possibility' of modding. The mod tools could also be offered through the service. So you load up unrealEd or hammer in 720p and get cracking.
Actually, it's not so crazy to have them offer something with a bit more freedom, not for modding tools per say, but maybe an FTP service so you can get your own games or video files on there. It's hard to get the security issues solved right though, video card drivers are full of holes.
Guys, 80ms is probably the smallest amount of input lag you'll see. It's the time they're saying their servers will take to process your input and show you the result. It does not factor the distance beyond 1000 miles.
OnLive will only be good for turn based games and Peggle.
So imagine my surprise today when I get the following in my inbox:
Congratulations, your pre-registration for the OnLive Game Service has qualified you to receive a special offer! You originally signed up to receive three months of free membership to the OnLive Game Service, but we are pleased to provide you with an even better limited time offer:
By completing your registration now and joining us as a launch member, you're eligible to receive ONE YEAR FREE membership to the OnLive Game Service, plus ONE FREE GAME of your choice*, with no commitment for further spending. This registration is only open through June 3rd, 2010, so act quickly!
A free year and a game to test out the service? Sure, why not! This is a great way to get people to actually take time to try the program for themselves, since it removes any cost to them outside of time.
edit: Oh, I did have to provide a credit card as proof of being 18 or older and to give them an avenue for future game purchases and to renew the monthly membership after the year. Fortunately for me, my credit card expires before a year from now, so I won't even have to worry about that.
edit: I tried to go to the URL they sent me again after completing my sign-up, and they're single-use. Sorry, otherwise I'd post it up here.
I'd love to try this free pre-registration subscription offer. Except when I visit in Opera, the speed scan doesn't work. When I visit in IE 8... it says my browser is not supported (yet lists IE8), so I visit in Firefox... and the same thing happens.
Well that's junk. Oh well, I'll know within the first week if the service is worth keeping or not anyway. Try it on my laptop and wife's netbook. See if it works as advertised.
Someone might have done this already, I'm not sure. Let me know if I've slipped up. This seems feasable in metropolitan areas, but form your trailer in the middle of the desert, it probably wont turn out too well.
Speed of light: 3.00x10^8 m/s
Speed of electricity in coax cable: ~2/3 of this [1]= 2.00x10^8
I just got the offer for one year + game and signed up. I'm anxious to use the service on my laptop, plays video fine but barely runs newer games. If the games look good and are playable this $400 laptop is going to getting a lot more use. Game prices will probably be the biggest hurdle, seems like any game you buy is just a fancy rental like steam. In my case however it might make sense on a few games.
I just got the offer for one year + game and signed up. I'm anxious to use the service on my laptop, plays video fine but barely runs newer games. If the games look good and are playable this $400 laptop is going to getting a lot more use. Game prices will probably be the biggest hurdle, seems like any game you buy is just a fancy rental like any other digital distribution service. In my case however it might make sense on a few games.
Posts
And it's even funnier how those people immediately proved my above statement to be 100% true.
EDIT: Oh and the Netbook thing: Don't buy a netbook if you want to play games newer than X-Com and CivII and Majesty 1. They are not made for gaming. It's like buying a DVD player and complaining that it doesn't run Halo 3.
Hardware cannot compensate for input lag. That is the fundamental problem, and why people are saying attempts "compensate" along those lines are akin to solutions for a perpetual motion machine. There is no way to compensate for input lag unless you very literally have a time machine of some form, because user input is not predictable. The issue of specialised "hardware" doesn't come into it.
The scenario is this:
1) Player sees visual output
2) Player responds to visual output with action
At this stage you've got the start of your input lag
3) Player input travels across the internet to server location, which takes x miliseconds.
4) Instruction is accepted and executed, which happens effectively instantaneously.
5) Resultant image is rendered, formatted, and compressed prior to sending. Given promises on their front, I'm willing to accept this will take a negligible amount of time as well (I'm doubtful that's true, but I'm willing to accept it for now).
6) Image feed is streamed back to the player over the internet.
7) Player sees results of input they provided, responds further.
Therein ends the input lag. In-between phases 2 and 7 you've got a gap where the user has provided input, but the simulation is still running without knowing about what the user input is. In actual terms this means the user is effectively giving input several hundred milliseconds behind where the simulation actually is. The simulation is responding to instructions that are based on where the simulation was a few hundred milliseconds ago.
The real question is how significant this input lag will end up being, and whether it actually ends up affecting the gameplay in real terms. But you can't "compensate" for input lag. It's already obvious that for some games even severe input lag will still be functionally playable. I mean what's 1/3rd of a second or more when you're playing King's Bounty? In others, particularly reflex based games, it can make all the difference. I mean, I cannot imagine playing even a game like Super Mario World with a significant input lag, that would pretty much destroy the whole thing. FPS's might be OK (depends on the game and the extent of the input lag), but fighting games are out. Similar with RTS's depending on gameplay style, games that are heavily micro dependant wouldn't be nearly as viable. And any semblance of online play is pretty much out the window because you're now placing input lag on top of the regular internet communication lag you'll be having (unless again, the game is turn based).
EDIT: And this isn't even me saying that OnLive is going to fail, I'm one of the few who actually believes it might have a shot if it starts off slow and tries to remain around until such a point where the idea is a bit more viable (expanding to other services would be a good start). But you can't just will the potential lag issues away, those aren't in OnLive's hands, not completely. Especially relevant on a service that costs so much on a monthly subscription on top of buying and renting the games for usage.
screenshots, an interview, anything?
edit:looked at their site
supposedly this goes gold june 17th, we shall see then
i hope this is real
it really would be a game changer
EDIT: at about 20 minutes in hes playing crysis on an iphone....
its ok, i'll be interested enough for both of us
fair enough
if it even comes close to the 80ms lag they mention i am so getting this
80ms is nothing. Like i said as long as im not trying to compete online and just dicking around, or playing a single player game, then this will be great fun.
80ms is the very opposite of "nothing".
Tofu wrote: Here be Littleboots, destroyer of threads and master of drunkposting.
But with respect to rhythm games (or anything that requires precise timing), the problem is not solely with the fact that there is perceptible input lag but the fact that this lag is also variable. From moment to moment as you are playing the game, the amount of input lag for any given action is subject to your ISP's service, traffic on the routers between you and Onlive, congestion on Onlive's servers, interference with your wireless internet (if you use it). You can't compensate for input lag that could change with any input.
It is. At that point you're actually talking about recoding the game itself, but even then, it's not a question of tolerances as such. To take Street Fighter 4 as an example, presumably you're still going to be able to do a Hadoken regardless, as long as you input the command properly (this is presuming your latency is constant throughout, and not varying, which is another fairly big hurdle). The issue is that it'll be coming out x hundred milliseconds later than you wanted it to. There's no way to account for that because there's no way for the game to intelligently decide "OK he's going to want to do a Hadoken at this point" and execute it for you before it's received your actual input. I mean, I'm just learning SF4, I'm a complete novice at it and I still whiff the occasional special, and I regularly mess up combos. Heck on rare occasions if the pressure's gotten to them, even the pro players do. Add in input lag on top of that and your likelihood of that happening increases a LOT all of a sudden.
And I mean, Street Fighter 4 is actually pretty dull to play singleplayer, it's all about playing with other people. Which is why it becomes an issue for games like that, even with as "minor" as an 80ms input lag, that seriously screws up the timing a ridiculous amount.
For most other games, I'm not so sure 80ms input lag is going to be an issue. But frankly, I'm seriously doubtful they're going to achieve that kind of input lag.
It would still be 80ms because that's the amount of time between when the visual stimulus is created and when the input feedback is received.
You are also saying that one of the base principles of the system, that the games will "just work" with OnLive is wrong.
The reason I don't care for the service is that it is the *height* of DRM. If it picks up a little bit, you will never be able to own a game ever again. You thought Ubisoft was getting draconian with their latest DRM exploits? Imagine how those corporate suits must be salivating at the thought that not only can they ensure that no one pirates the game BUT they can make you pay monthly for the privilege of playing it.
3DS Friend Code: 2707-1614-5576
PAX Prime 2014 Buttoneering!
It pretty much is the golden egg laying goose yeah. Once a month, ka-ching! Although to be honest I'm not sure such a system, even if it takes off, will really replace people owning consoles and games. Far more problematic that singleplayer is multiplayer, to the point where I don't think it'll be viable over such a service. And without that, well you've immediately got problems.
I agree with the DRM part. However, I'm still holding hope for a rental service of some sort. Pay X amount per month, get access to Y number of games. Or even X amount per month for unlimited rentals. It would be like netflix streaming except for games. Which I hope for, and is where I see this system shining. Will it happen? I don't know.
Sounds like Gametap to be honest. Only without the streaming.
The DRM part is exactly what I find worrisome. Developers and publishers are probably thinking this is where they need to move. Not only do they avoid the possibility of piracy, but no one, every, anywhere will be able to get ahold of the game code (barring some internal OnLive / Developer leak). That means that there's no possibility for leaks, no possibility of resale, no possibility of modding. They'd avoid people stealing their game (or even theoretical people stealing their game, which some devs like Crytek baaaaaw over constantly), they'd be able to charge much higher prices further after release, and they'd never find their game on the 11 o'clock news because some horny teenager made a 'nude patch' that was never part of the intended release. The big hook here is that this form of DRM provides an exceptionally appealing convenience to a large audience - the ability to play current games without putting down money for a computer. Alone the DRM would never work, but with that kind of added benefit a lot of players would be more than willing to swallow the negatives.
To be clear, I think it's a disturbing trend to put consumers in a situation where they pay for everything and own nothing. I like being able to mod my games, and I like being able to tweak settings to my preference (They've confirmed that there will be NO possibility of modifying your graphical options with OnLive, which you might expect by the nature of the service). But if this service takes off, and some big third party title makes a ton of money on an OnLive exclusive release? You can bet the other publishers and developers will be beating down the door.
But, that's all theoretical, assuming the technical difficulties inherent to the service don't hamstring the entire idea.
Ok, so this whole latency compensation thing is basically impossible unless they've done one of two things:
• Invented a time machine
• Invented a way to read your mind
The first I can't think of any way to fake, but the second... Hear me out. What if, for each game, OnLive has taken huge recorded input dumps of hundreds of test players, crunched the numbers for timing between movements, directions that follow directions, pacing, reactions, and so on, and made a highly sophisticated heuristic profile that records your input patterns for the first few seconds of a game, runs vast, complicated algorithms on the sample to match you to the overall game profile, and then uses this data to constantly predict your next twitch 80ms in advance.
The above is completely insane, highly unlikely, and the only way I can imagine OnLive "compensating" for input lag.
PSN:RevDrGalactus/NN:RevDrGalactus/Steam
OnLive will only be good for turn based games and Peggle.
A free year and a game to test out the service? Sure, why not! This is a great way to get people to actually take time to try the program for themselves, since it removes any cost to them outside of time.
edit: Oh, I did have to provide a credit card as proof of being 18 or older and to give them an avenue for future game purchases and to renew the monthly membership after the year. Fortunately for me, my credit card expires before a year from now, so I won't even have to worry about that.
edit: I tried to go to the URL they sent me again after completing my sign-up, and they're single-use. Sorry, otherwise I'd post it up here.
Anyone else having this issue?
Subscription-based services are allowed to spam expiration dates until they figure out what your new one is. It's to facilitate uninterrupted service.
Speed of light: 3.00x10^8 m/s
Speed of electricity in coax cable: ~2/3 of this [1]= 2.00x10^8
Order of magnitude of latency acceptable: 10^-3 s
10^-3 s * 2.00x10^8 m/s = 2.00x10^5 m/s = 200 km.