tubechallenge.com

Congratulations on completing the entire network! And it's very kind of you to reveal your entire route.

Doesn't Bob Robinson's time include the East London line?

"Also something else happened here but i dont think its appropirate [sic] for the forum."
Toilet issues?

Starkey7 wrote: Doesn't Bob Robinson's time include the East London line?

I assume Andi means beat Bob Robinson's record for total number of attempts which was 51, rather than beat Bob's time.

Many congratulations to Andi and Sarah on getting this round on what was a tough day and unpleasant weather conditions. Welcome to the full network club and like many of us you probably now have the "bug" to beat the network........
It is quite a good "buzz" just getting around and you never know Sarah may come out to play again as you soon forget the pain........

I was also talking to Soup Dragon and the others that we would soon have Bob Robinsons attempts record within sight, only another 44 to go..........

Many thanks its only encouraged me to go out more I still have times to set for many challenges mouse,zone 1,alphabet and dlr but its the full network i find easier.

as for being the second attempt i really look at it as my first the other was unplanned no door positions accommodation at 1 end was unknown didnt know some of the run times even though on that one it looks like i made a train no else has ever got?so i abandoned it even though i could have gone on.

i still have i think 12 days left holiday before april 1st so i be out again soon

Stakey 7 wrote

Congratulations on completing the entire network! And it's very kind of you to reveal your entire route. Very Happy

Doesn't Bob Robinson's time include the East London line?

"Also something else happened here but i dont think its appropirate [sic] for the forum."
Toilet issues?

Im here to help fellow tubechallengers so if they wish to use feel free to do so although i tweak it slighty as u can see even non runners could match Sarah run times so its a good figure for them.Bob Robinson 51?attempts im looking to beat i think thats is a goal which is achievable and lets just say anybody who picked a water bottle at Mill Hill East it didnt contain apple juice.

Well done indeed, Andi and Sarah!

Interesting to get a full route published. I ran it through my "Very Realistic Simulation Based on Not-So-Fast-Running" software. My conclusion is that this is the best route I've analysed apart from my own ones... Furthermore, I get an expected arrival in Upminster on the 23:11 - some 13 minutes outside the provisional GWR. I think part of that is due to that I assume the pessimistic JP times on the Richmond-Hatton Cross link. I have Richmond arrival at 19:08 - how does that compare to your plan Andi?

BTW Mine and Lars' tentatively planned comeback route should have us arrive at HT123 at 22:48 (not taking into account recent/upcoming time table changes for Pic Line) - just 10 minutes within the provisional GWR...

Some interesting analysis there Hakan, this seems to confirm that the provisional GWR time set on Dec 29th is indeed going to be a tough nut to crack............

Going Underground wrote:Some interesting analysis there Hakan, this seems to confirm that the provisional GWR time set on Dec 29th is indeed going to be a tough nut to crack............

Yes, I think so! I'd rate it as being "10 mins" better than your existing GWR (relatively speaking, of course), since our expected time back then was 18h flat.

well my plan obviously im not going to give everything away I planned to be in Richmond before that my estimated plan was 17h12mins obviously with delays I was expecting to lose a little for delays but there was plenty of potential to gain some time as well i thought it would even it out around GWR.

Due to a little bad luck we just caught the bad jubilee line bit at the wrong time in estimation looking back at the times we collected and the connections i could have made on my own i lost 55 mins or close so it would have been very close to GWR but as my plan said

Richmond-Hatton Cross 34 mins planned
on the attempt it was
Richmond-Hatton Cross 32 mins

hwolge wrote:
BTW Mine and Lars' tentatively planned comeback route should have us arrive at HT123 at 22:48 (not taking into account recent/upcoming time table changes for Pic Line) - just 10 minutes within the provisional GWR...

I also am having problems being within the 17hr 30 time with the new Piccadilly line timetable, the 8 minute wait combined with the fact you only have a train every 10 minutes that you can catch makes it even worse. My route planning though does work on the fact of guarenteed connection times often leaving 4 minutes for cross platform changes rather than risking putting in the 1/2 minute connection in the plan I could possibly of made on the day.

I did also look at starting at Heathrow T4 as this elminates the problem of having to wait at T4, if you start there. But the District line didn't really work that well with the Piccadilly that early in the monring.

Going Underground wrote:Some interesting analysis there Hakan, this seems to confirm that the provisional GWR time set on Dec 29th is indeed going to be a tough nut to crack............

...or just that the Brown-Starkey route really is the best route!

Starkey7 wrote:

Going Underground wrote:Some interesting analysis there Hakan, this seems to confirm that the provisional GWR time set on Dec 29th is indeed going to be a tough nut to crack............

...or just that the Brown-Starkey route really is the best route!

It's not bad, there are a still a couple of aspects I am not 100% happy with so with a little "tweaking" and given a day with good service I think it will satnd up pretty well......

hwolge wrote:Well done indeed, Andi and Sarah!

Interesting to get a full route published. I ran it through my "Very Realistic Simulation Based on Not-So-Fast-Running" software. My conclusion is that this is the best route I've analysed apart from my own ones... Furthermore, I get an expected arrival in Upminster on the 23:11 - some 13 minutes outside the provisional GWR. I think part of that is due to that I assume the pessimistic JP times on the Richmond-Hatton Cross link. I have Richmond arrival at 19:08 - how does that compare to your plan Andi?

BTW Mine and Lars' tentatively planned comeback route should have us arrive at HT123 at 22:48 (not taking into account recent/upcoming time table changes for Pic Line) - just 10 minutes within the provisional GWR...

I'm aware of knowing very little about Hakan's program, other than that it is based on Genetic Algorithms. I do know quite a bit about solving related problems, and generally this involves two fairly distinct parts

(1) Identifying and searching through possible solutions (i.e., routes)
(2) Assigning a score - in this case an expected time (or statistical distribution of times) to each solution (route)

There are (combinatorially) many possible routes, though a very large majority are not serious contenders. Assuming that the search part of the program has some inherent randomness (as GA's do), it should be possible to get an idea of how complete the search is by whether or not different & better solutions are found if the program is run for longer.

Not sure whether Hakan would be prepared to comment on this?!

Oh, I'm happy to comment in general terms. First of all, I have of course a multitude of programs, some more, some less related to each other. In one extreme I have a GA-type of search algorithm program that has been very helpful in coming up with a few novel approaches to the GWR. That one is a bit stupid in that it uses the time tables as exact numbers and cannot handle simple variations, like:

"When changing at Earl's Court, I'll do Kensington (Olympia) if that train arrives firat, otherwise I go to Wimbledon and do K(O) later"
or just:
"take the first SB ELL train to either NC or NCG and then proceed to the other one"

The other extreme is the one I used above, where I enter a specific route, with variations (to some extent) and evaluates it with assumed distributions of delays etc (quite heuristicly estimated, but somewhat vindicated over the past few years). Sample syntax for the input file is:

Finsbury Park
Walthamstow Central
Leytonstone
Epping / Hainault
Woodford
#IF VISITED ("Hainault") THEN
  Epping
  Stratford
#ELSE
  Hainault
  Stratford
#ENDIF
Waterloo

where "Epping / Hainault" means "I intend to go to either of these places based on which train that leaves first", then two lines down, I say, "Well, if i have visited Hainault already, then I take the next Epping train (and then back to Stratford), else I'll go via Hainault.

The output from this program is two-fold. Firstly I get a compact list, where I can follow the route (with variations) and see the expect 30/50/70 percintile arrival times and the weighted average. Secondly I generate a "complete" time table with all reasonable connections in a "standard" time table fashion (some 10 pages long). That part is quite useful when actually going out there to execute...

HÃ¥kan, just like Ikea you are legendary.