Hard to say precisely, as a lot of this is placebo, and being rigorous in measurement is unpractical.

My purely subjective exp is that I moved from ADSL to Fiber, and at the same time the routing to DE has improved. Ping went down from 48 to 8. I used tn -20 before and I eventually removed it as I felt it was there were confusing situations where things should have been smoother.

I also have fiber (35dl/4up) and I'm from Austria but my average ping to DE is ~46-48 , how the hell do you get 8 ms? Are there other commands necessary besides maxpackets (set to 125) and rate(25000) ? Or is there any software which could improve my ping?

fiber in austria with 35/4?
UPC?

If so, that's cable and even their so called fiber products aren't proper fiber connections. UPC also has the worst routing of any provider and usually has bandwidth problems in the evening.

Definitely called fiber power on their website but I can't find my product atm , they got 2 new lines (75/7.5 and 150/15) which take up the old links there, think they had 35/4 and 100/10 before.

Well I wanted to change my provider since I had bandwidth problems every other evening for a few weeks last year but I couldn't really find anything better, best A1 connection is slower than mine and drei is about same speed/price , what provider do you have or could recommend me (I live in Graz so there prob are more available in Vienna)?

Speed wise you probably won't find anything that can compete with UPC for the price but for the ping pretty much anything else is better. There might be some Graz specific providers that i don't know of that might be good but i can't help you there.

I'm pretty much stuck on tele2 16/1 cause i don't want to lose my semi good ping.

Doesent UPC provide somekind of hybride, like FTTC or something?They sell this down here, maybe thats why pious ping is so crappy.

They are calling it "fiberpower" in austria but if anything it's probably FTTN. UPC has always been a cable company in vienna. My guess on why their ping is usually shit is because they route through their own net throughout Europe instead of the best possible route. They also have had problems with handling the bandwidth of their own users in the past which probably contributes. Imho one should stay away from them if you care about ping.

Must be the isp routing. I bet your packets go through Amsterdam. Nothing much you can do... Maybe a vpn, but I wouldn't know one.

almost

http://tools.pingdom.com/ping/?target=91.198....id=6848266

What is low ping to you? Also if your ping is below 40 I personally I wouldn't use it but if your ping is 40 or higher -10 is probably a good setting. -20 is good for 60+. This is just my opinion.

-20 for every ping.

it kinda aims for you

Binding timenude -20 on zoom, the ultimate.

why dont u just download an aimbot and be done with it

also have a macro that starts utorrent on the same key XD

utorrent? boy, there are proper programs to jam your connection!

;) I was just trying to remember what the windows kiddies were using when I last checked. I just use transmission on Debian and my windows box is just for quakelive pretty much.

u use transmission and call utorrent users kiddies, whahahahahaa

I used rtorrent for a long time but installed transmission when magnet links were gaining traction and rtorrent didn't have proper support for them yet. I don't really torrent too much so I never arsed myself to go back. Any particular clients you'd recommend for debian?

was more of a mockery since last time i used transmission it was VERY basic compared to utorrent.

Oh yeah it is basic for sure, that's actually why I installed it.

using -16 on 60 ping, all the maxpackets 63ers become too untollerable otherwise

cl_autotimenudge 1 and no worries :(((( i feel bad for using it now, since i started with playing on laptop with 0 tn

Im asking specifically for the pings i get on SWE/Nor servs. I get 8/20 there and was wondering if i should remove it altogether when I play there.

no need for timenudge if your ping sub20 imo

any ping under 60 and timenudge is detrimental imo, the only weapons it helps are RL/PG

They should ban timenudge users im serious. It causes warp and is killing quake for me. I See really bad players capping in ctf because They just warp trough shaft and rail and rockets are so hard to hit Its like playing against cooller ... While they lets say hit 20 shaft vs you An experienced player who could win with mg from his shaft so to say. Also Tons of warping people from poland and russia etc Its An invasion i really like them as persons but i hate them in game because 9 outta ten times they cap because of warp. THIS WOULD NOT HAPPEN ON EQUAL (LAN)PINGS

Timenudge does not cause warping.

It's an entirely client side, visual cvar.

Timenudge does not cause warping.

Watch out with the formulation.

It can cause warping but it will never make the user warp to others. The warping involved here is that the TN user's opponents and teammates may appear skippy to HIM. I think all the confusion comes from that.

that is the definition of client side :)

also factually incorrect.

How so? Isn't the whole idea behind timenudge to update the visuals faster with the new information rather than interpolate between that and what was available before to smooth it out?

it changes the timestamp of the sent packet. Which is used to calculate the unlagged correction on the server.

Now I've never seen the actual code which the server uses for unlagged correction but I've tested it with a few people and it works like this.
I pointed this out to ID back in 2009..... And I wrote about it on esr in 2011 (I think).

Ok so the client sees things N ms in advance and tells the server to take account of that with the bigger timestamp so it doesn't look too far behind in its gamestates. So far so good; that effect doesn't affect other clients.

Now what other use of this timestamp could make someone skippy? If it were used to accept/reject a packet, one could create artificial jitter by toggling timenudge on and off constantly. But packet sequence numbers are still used for that, right?

It does effect the other client, in general the lg / rail (of the person using tn) will be more accurately drawn on the other clients screen. (as the shots are "less unlagged")

I'm not 100% on the 'packet sequencing' but what do you think happens if I say I'm at X on time T then walk for 25 seconds and then say I'm on X+(v*25) on T+25... compared to walking the same 25 seconds and saying I'm on X+(v*25) on T+5 ? :p

But in general I don't think that people try to "create artificial jitter" by doing this there are much easier and more effective ways.

I'm not 100% on the 'packet sequencing' but what do you think happens if I say I'm at X on time T then walk for 25 seconds and then say I'm on X+(v*25) on T+25... compared to walking the same 25 seconds and saying I'm on X+(v*25) on T+5 ? :p

Probably nothing because your timestamp shouldn't be used for that. It should only be used for correct backward reconciliation. That said, it probably doesn't even need a timestamp, only your ping and your cl_timenudge value.

As far as I know, in VQ3, packets were not associated with a "time", they were associated with a sequence number. Every time the client sent a new packet, it gave it a new sequence number, superior by 1 to the last. The server, upon reception of a packet, looks if its sequence number is higher than the most-up-to-date packet received from that client and discard it otherwise. Look at Netchan_Process(netchan_t*, msg_t*) there :

https://github.com/id-Software/Quake-III-Aren...net_chan.c

There is no need for a timestamp here. The server only took account of information updates and only worked on its current snapshot. There would be no such thing as changing your position at a different time; it just sets your most up-to-date position on the current snapshot and that's it. The potential problem I was talking about earlier was if these hypothetical timestamps were used instead of the sequence number to determine which packet is the most recent. But I don't think that's the case; there would have been no reason to change something that works unless there were some deeper changes (like updating different snapshots according to a packet timestamp or something).

It has been several years since I looked at the code.

And well I don't want to make a huge post (atm) on the q3 code as I might mistake some details but you should look at the following lines and re-evaluate.

On the client this happens:
cl.serverTime = cls.realtime + cl.serverTimeDelta - tn; (cgame.c)

When you press a button this happens:
// send the current server time so the amount of movement
// can be determined without allowing cheating
cmd->serverTime = cl.serverTime; (cl_input.c)
(notice cmd->serverTime)

These cmds get used on the client in an pmove action. (in order to get to the next frame).

They also get stored in a (sequenced) message. Which gets send towards the server.
However the server reads each "cmd" as a cmd (also with cmd.serverTime which is the time which the client attributed to that cmd).
And this gets used in the same pmove action to update the server status.


There are some checks though on the server if the servertimes actually don't run backwards (in which case they are ignored)...
Which as you can reason about might cause warp. (a cmd which has been made just after changing the timenudge can create a cmd which has a smaller servertime than the previous command and thus it gets ignored).

I was more accurate on the specific lines of code in the past though.....
I might write a better reply if you have problems with this one.... but I need a bit of time for that again.

Last time I delved into the q3 code to figure this stuff out I spent like 15 hours of reading the code / understanding it / checking if I was really right etc.

But anyway that the server *does use* the clients "cmd.serverTime" should already be fairly obvious considering Carmack's / TTimo's coment on the cheating part.

So I took the time today to go look that in the code. You claim that :

Which as you can reason about might cause warp. (a cmd which has been made just after changing the timenudge can create a cmd which has a smaller servertime than the previous command and thus it gets ignored).

Because of that :

cl.serverTime = cls.realtime + cl.serverTimeDelta - tn;

But that can never happen because of what follows this very line :

// guarantee that time will never flow backwards, even if
// serverTimeDelta made an adjustment or cl_timeNudge was changed
if ( cl.serverTime < cl.oldServerTime ) {
cl.serverTime = cl.oldServerTime;
}
cl.oldServerTime = cl.serverTime;

https://github.com/id-Software/Quake-III-Aren...ame.c#L976

That said, even if toggling timenudge constantly could make you warp to others, that's a bit far from the usual thought that normal usage of timenudge can make you warp.

you: "It can cause warping but it will never make the user warp to others."

me: "incorrect"

you: "how so"

me: "points towards a case where changing tn makes a player warp".

you: "That said, even if toggling timenudge constantly could make you warp to others, that's a bit far from the usual thought that normal usage of timenudge can make you warp."


I don't see where we disagree?
I'm just saying that timenudge is not purely client side and it can possibly even cause warp.

That statement didn't show disagreement. Only that the random player that says normal timenudge users warp because of it are clueless.

However, I'm not sure it can cause warping even while toggling.

that would kind of depend on if the "movement cmd's" also have a time and if that's used.

I thought they were but I don't really want to say so and be wrong so if you want to be sure then check it ;)

More importantly, the outgoing packets are created and sent in CL_SendCmd() prior to the timenudge altering the cl.serverTime timestamp that is used for rendering in CL_SetCGameTime():

// send intentions now
CL_SendCmd();
// resend a connection request if necessary
CL_CheckForResend();
// decide on the serverTime to render
CL_SetCGameTime();

orly?

Errh, please force
com_maxfps 125
cl_maxpackets 125
rate 25000
snaps 40
on all players...
Also require a vid_restart on FPS changes, if you don't want to force FPS,

Snaps is removed from clients, and forced internally to the servers framerate, which is 40 for QL servers.

rate seems not to make much difference, and maxpackets minimum is 63 which doesn't cause any problems. It just feels a bit laggier for the client.

Maybe you could cause yourself to look choppy with a very low framerate (30 is the minimum) but I doubt it.

If you're looking for a solution to warping players, you're looking in the wrong place.

Maybe you could cause yourself to look choppy with a very low framerate (30 is the minimum) but I doubt it

You can.

I was suggesting standardization to limit variables.

>You can.

It wouldn't be warping though, it would be very steady and easy to predict, plus it would be a huge inconvenience to the player using it.

>I was suggesting standardization to limit variables.

What for? We know that no combination of these causes the warping everyone is familiar with, nor can any combination of these prevent a player from warping.

alias +wololo "com_maxfps 30"
alias -wololo "com_maxfps 125"


Just for the sake of standardization

Test it. See if it causes warping.

Even if it does, It most likely won't look anything like the warping you're thinking of.

It won't cause warping it it'll be more like jerkiness

I do find that people who have less regular movement also have all settings to a minimum...
Not sure why it feels that way and I don't really know how to show that it does matter... (it can ofcourse be that most people reduce the values when they have crap i-net. So it might just be small a correlation rather than a causation)

I do find that people who have less regular movement also have all settings to a minimum...

Well maxpackets 63, and rate 16000 are the default. Are you sure they just aren't newbies who haven't changed any settings?

You can't "force quake" to "force less than 30fps" but if you just start a background process which takes a lot of cpu I'm pretty sure you can get below 30 fps np.

How could anyone play like that? Plus by this method they would likely warp with any netcode with the exception of maybe netquake.

On the 1st point... I don't really play vs "newbies". Also I can't really tell if newbies move irregular because they move too slow to notice it. But when you get a pretty high velocity it often seems to me as though they often have some mini-warps (like missing a few frames in the movement).
I'm alergic to those things heheh :p

On the 2nd point.... I'm a very logical person and that sometimes gets taken against me.
I'm just pointing out that there is an exception to what you said.

On the 1st point... I don't really play vs "newbies". Also I can't really tell if newbies move irregular because they move too slow to notice it. But when you get a pretty high velocity it often seems to me as though they often have some mini-warps (like missing a few frames in the movement).
I'm alergic to those things heheh :p

You know good players that voluntarily play with rate 16000 and maxpackets 63?

On the 2nd point.... I'm a very logical person and that sometimes gets taken against me.
I'm just pointing out that there is an exception to what you said.

Sa'll good, I'm no stranger to pedantry myself.

No I don't know any active player who does this... it would be kinda funny if someone did this on purpose such that he becomes less hittable... and then say it to his opponents. Wouldn't it?

It's like saying: "Hey I just won because I cheated." people don't work that way....

I do know that there are several players who will do just about anything to get wins.... it wouldn't suprise me in the least if people attempt this.

But anyway... there are much more simple ways to get free wins online...

I do find that people who have less regular movement also have all settings to a minimum...

By minimum settings what do you mean?

the lowest netsettings you can have (maxpackets / rate).

Ok, so you know players then (not necessarily good, or newbies) that deliberately use rate 16000 or 8000, and maxpackets 63?

As I already said in comment 98.
No I don't know of any active player.... but it wouldn't suprise me in the least if there were a few.

And I do know that some people have played on those settings..... But they felt like their connection would become unstable otherwise (or so they said).
But it's hard to say if they had those small little anoying warps because their connection was already unstable or that they warped because of the settings.....

also theres no actual limit to the lack of frames?
You can't "force quake" to "force less than 30fps" but if you just start a background process which takes a lot of cpu I'm pretty sure you can get below 30 fps np.

actually in QL it's not an entirely client side visual cvar.

It shouldn't cause warping though (on any normal use).

their warp is very probably not due to any timenudge setting but more due to bad connections / torrents / uploads / downloads / stuff like that.

tn = - (lowest ping on lagometer) / 2
pwnage

is that a real equation or just something u made up

that's basically autotimenudge formula, but much more stable one with lower pings

So -25? 8 (

This doesnt make sense because on UK for example, the lowest ping i get on the lagometer is around 40.

-20 on 40 ping? isnt it a bit too much?

no it's not too much

It actually isnt that much. His "formula" kinda works. While having 30 ms on SWE, i tried -15 TN with cl_truelightning 0 and drawcrosshair 0. The result was something else. It was like the LG was shot on a bot match in Q3 (the lg still doesnt keep u with the xhair, but its almost there). Try it. It's not tha hard to test.

But I heard somewhere that TN is pretty much a double edge sword. You have to dodge earlier because of it (don't ask), you are a bit easier to hit (imo) and you are a magnet to plasma.

it's funny that in another thread you're arguing with me on netcode issues....

I was just saying that (again, as in the other reply, im totally drunk atm) I think that u have a silly excuse for not playing, and I was just trying to understand WHY you dont play, whereas I just explained something I experienced here.

since im a high pinger ( ping jumps from 80 to 150 and even more, depends on the mood of my connection ) i use timenudge -20, i have to say that it is really helpful but kinda confusing at the same time, it helps a lot when using LG but kinda confusing when using the RG cause with timenudge 0 you aim at the enemy = it hits the enemy while with -20 you have to aim where the enemy was like half a second ago to actually hit him, it helps when you fire at the enemy and he just hides really quick, with -20 the rail will hit him while with 0 it won't .

it is hard to explain but i have to say that timenudge helps the high pingers but makes the game lags a bit and get you in some confusing moments sometimes :P

if the game lags, you have too much tn

No such thing as too much tn. I played on african servers with ~230 ping and cl_autotimenudge 1.

It was barely playable, but still, it was playable. At best, I could do a 70% height rocket jump.

Maybe it does help you at 80+ pings, I don't know. But I think your findings about firing 'around the corner' are not precise, or rather, it's got nothing to do with timenudge, but the unlagged code itself.
When I'm going out of line of sight I still get railed, and I also do the same to others all the time, no timenudge and low/normal pings. That's just how it works, due to buffer/delay in netcode, which makes it possible to have a mostly balanced playing field up to 80ms.

timenudge is useless if you have a ping <75ms.

uuh. not true..

why would you use timenudge when there is the anti-lag up to 80ms ?

because there is no anti lag up to 80 ms, its more like up to 70.

also even unlagged goes wrong at higher pings and especially tracking gets erratic so you need tn to level it up.

also dont you ping much lower then that, how do you know how 80 ping would play?

He probably just assumes like most of ESR.

Well he is right though, netcode unlagged 'buffer' is 80ms. I don't have a link for you but that is official information as far as I remember. I have played on all kinds of pings, and my experience seem to reflect that.
I ping 80 to RUS and sometimes 80 to UK when my routing is broken, and I don't aim any differently there. At 130 to east coast I have to of course predict opponent and shoot in front of model.
Makes sense to me.

As for timenudge, I have experimented with it quite a bit in the past, and found that it didn't make any difference, other than making models unsmooth and warpy(especially if you play at 120hz).
Maybe it could help reduce the number of occasional railthroughs - I'm not sure - but even then I'd find it more disadvantageous than anything. Also I believe that railthroughs are a product of minor hiccups in your/opponent's connection(ping fluctuation/packet loss).

"netcode unlagged 'buffer' is 80ms."
sorry but I honestly don't care about any documentation when the difference between 70 to 80 is so major (and with major i mean "slightly shit" to "fuck this shit" :)) the difference between 60 to 70 is far less then this so even if unlagged does still work up to 80, it really cant do its job any longer.

"I ping 80 to RUS and sometimes 80 to UK when my routing is broken, and I don't aim any differently there"
are you implying you aim the same way, or better phrased, you perform the same way (damage output over accuracy) as you when you ping 50 etc, or even at 60?

"found that it didn't make any difference, other than making models unsmooth and warpy(especially if you play at 120hz)."
this is personal preference.

"Maybe it could help reduce the number of occasional railthroughs - I'm not sure"
wrong, they result in more of them, but the pro for it is that the enemy model still is portrayed in the right place most of the time due to tn trying to correct enemy position on your client.

"I believe that railthroughs are a product of minor hiccups in your/opponent's connection(ping fluctuation/packet loss). "
partly true ofc, also a bit of a "duh" statement. but not all are due to this.

Well I guess we just have a disagreement then, or had different experiences. I think the difference starts to become major after 80.
I know you pinged 90 to DE often, and still do a ton of damage, and seen some ppl with 100. So I don't think 70 to 80 could be major, and keep in mind opponent's ping is part of the equation, and all combat situation will be different.

Sorry what I meant was by not aiming differently, is that I still aim at the model as with any ping other than 80+ where you have to start aim in front, the higher your ping is obviously.

Don't know how you could call it personal preference when it simply makes it harder and take more effort to track unsmooth model, and exaggerates those who already warp to begin with, but fair enough.

As fair railthrough, ok you say not all are due to hiccups. So what else can it be, netcode or bug. You also say TN will increase railthroughs not decrease. Well TN will show you less updates of the player model but they'll be slightly more accurate, so care to elaborate why you think that's the case? Sounds illogical to me.

- I do a ton of damage because I have memorized alot of situations, have enough practice with the ping to somewhat adjust my aim to it, and especially in pubs I just abuse powerups easily because so few ppl in pub actually use or even time powerups well (hello p9 pubs where I get ALL powerups including enemy regen)

that being said you don't assume I am performing the same way I would with low (50 etc) ping when you say "tons of damage" I hope? Cause I don't, far from it, both my aim and combat capabilities are hurt badly when I ping around 80 or more.

- The problem with unlagged is that you do aim at the model somewhat but the closer you are to the opponent distancewise the more you see it doesnt work right. the higher you ping the more distance this becomes correct. If you try cpma sometime with the same high pings you can see how much better that mod handled it vs ql (ql is still not bad at all though)

- the unsmoothmodel is still positionally more correct so you only have to adjust to the warp it induces, normally without timenudge you require much better dodge reading which can be very hard, so the payoff is more compared to no nudge

- the railthroughs are mostly all due to netcode, especially the ones midair, the higher you ping the more you will usually get, timenudge doesnt actually change the mechanic but due to the clientside warp, even when you adjust your aiming to it you will have alot of occurances.

- Sure, I don't mean that, just purely hitting regardless.
When was the last time you played on lower ping, then? What server, and opponents?... What do you base your comparison on? I'm really curious. Because the way you make it sound like how disadvantaged you are at 70-80 ping, and how much better you'd play at 50 or lower, is frankly really exaggerated...
I'm not saying there isn't a difference, there is(and again opponent ping dependent), but not as HUGE as you claim it to be.

- Didn't play Q3 so no comment. Although "doesn't work right" isn't saying much.

- Without TN you require much better dodge reading? What...
It's the opposite, if anything. Dodging should be easier to read when movement is more fluid and smooth. That's almost like saying you can track better at 60hz, or track a player better who has permawarp with low net settings or whatever.

- Ok first of all, let's not call TN warp, because it's different from what people usually identify as 'warp'(that actually goes through the server). Let's call it less fluid or less smooth, less updates seen of the model but potentially more accurate. Now, just because you have this 'warp', everything else remains the same - you try to rail the model, and if you hit on client but not on the server, you have a thru. But, because of the potentially more accurate models, you might get less of these.
Your previous claim that TN usage gives you MORE thru-s, is untenable.

word of advice... most people don't care about what is actually happening... they just think they are correct no matter what anyone else tells them.

Yea well the problem is, from my experience(both ingame and real life) most people have no idea how to properly analyze things and make solid conclusions.
Let's say that you want to examine the effects of TN. So what do you do. You get a test partner, go 1v1 LG on Thunderstruck or something like that. Make sure you both have similar pings, with little to no jitter. Later on you could test different pings. Agree on a specific pattern of tracking and movement that you will be repeating. Then you can alternate TN, turn it on/off for either player, and see the results.
Even in this staged scenario, it can be really difficult to properly deduce what is happening, so that begs the question: how could players possibly make valid comparisons from just playing the game normally....

That's why, when people come and say "Oh I just set TN -20 and hit much better" and things of that sort... I have to take it with a grain of salt.

tbh I'd just download the q3 code and analyze that ;)

I don't use timenudge if my ping is under 60. If it's over that, I use -10, and if it's around 80 I use -20.

on a related note, does cg_projectilenudge make any real difference?
I usually play on 60ish ping, been using tn -20 for a year now, but recently my game experience has become more and more laggy.
I tried projectilenudge but hadn't noticed any real effect.

afaik cg_projectilenudge has been disabled for some time now

so is it possible that some players are using timenudge to achieve a kind of "mini-warping" that makes them difficult to hit with hitscan?(RG/LG)

or is that just the interwebs being bad/unfortunate circumstance?

if so is there any kind of crackdown by id on these people

Timenudge can't do that.

So nowdays autonudge 1 is better for high ping cause it can go lower then timenudge -20 ?

Edit: I play in Poland alot (60-80) ms ,i like the depressing atmosphere there ! Great stuff

why is it depressing? i somewhat like polish servers because the very different playstyles compared to swe servers fx.

Was joking around abit ;) If you mean different in a sneaky +backish kinda way .....YES !

if done right, otherwise i hate that.

-20 timenudge on all guns (any ping) except LG/PG/MG because I need the fluid looking player model for tracking aim.

thats quite dumb.

ya but it works

ur using tn for RL and RG. thats wrong.

u want to see the full animation of the model when you shoot rockets or rails, because both weapons heavily rely on predicting the enemies movement

give it a try mate

dood i know, but it works for me

placebo mb?

I tried all different kinds of timenudges, but this is the only combo that helps increase my accuracy on all guns, placebo or not, It's what I use.

Liner also does TN 0 on the LG.

Although I've seen him with the default config and he plays almost the same so I really don't think config is such a huge factor for him. Just knows how to aim and dodge and that's that.

liner has said lg is all cfg

but it's because quakelive lg is all mouseaim, so he figured he'd practice the most ridiculous dodge pattern and then aim on it, so he started at like 30 acc lg then gradually built it up to 35-37 acc while dodging like a shortbus kid

which is why you can't lock him but he can lock you

1v1 lg fight with him in a tunnel or something and i'd wreck him, but not eyetoeye or thunderstruck where he's got room to dance

I use -20 on all pings. I keep it on and forget about it, thinking about changing it would affect my concentration.

Even when i played on perth servers on 10 ping id still have -20 :D

Damn straight. Nudgers represent!

-7 on daily basis

as soon as i switch so a server with ping 40+ i go for -12

+ even on lowping against warping mofos i always go at least -12 up to -20

dunno if it's only placebo or not but it normally works out

Whos still timenudging in 2021 ???

https://www.youtube.com/watch?v=dMCOnEsU7Ys

100%. Kinda aims for you.

Now that's an expression I haven't heard in a long time, a long time.

it won't help you beat cyber, that's for sure

-10 because corona

I bet Evil had used timenudge once. And that's why he got constant ddoses, lost to Rapha most of his finals, left Quake and moved to USA to feed the gulls.