Zezo forum:

Post by BooBill | 2020-06-25 | 11:21:32

The wind direction in the router and in the dashboard don't appear to match.

At the 1110Z (10 minutes into the race) the router is expecting winds 252� at 6.2kt. Dashboard is showing currently 245.2� @ 6.8kt. That's really significant as it puts you either side of the max twa of 120.

15 Comments

Post by Hardtack | 2020-06-25 | 11:58:47

It seems that the 06z forecast jumped north by 20� w.r.t the 00z forecast.
I think that at 11am (UTC) zezo already switched to the new forecast while VR is still interpolating between the two.

I don't have light sails, but I'd LOVE to see the other boats change between LG/CO/Spi ;-)

Post by BooBill | 2020-06-25 | 12:13:11

Some guys had it figured out. Started 120 TWA, sailing a little above the layline initially, after shift they've been able to come back to the layline at a better angle and did it all on LG.

Those of us who didn't figure it out got screwed with an unnecessary sail change and slower angle to the mark. That will probably cost a mile to the front by the first mark. You know how hard it is to overcome that in a race like this.

Post by YourMomSA | 2020-06-25 | 13:18:10

This is a common situation, and has been discussed in several other threads. When the forecast suddenly changes dramatically, three different elements each handle it differently:

-Zezo
-VR user interface
-VR actual server (seen in dashboard)

VR should be able to solve the difference between their two, but it isn't reasonable to expect Cvetan to reverse-engineer VR's exact programming to make Zezo stay perfectly in sync. The three eventually all merge. How long it takes depends on how big a change occurred in the new forecast. And yes, sometimes one environment suddenly "figures it out" and makes a big jump.

This means it's important, if you want to do really well, to recognize when these situations are present, pay attention as much as you can until they sync up, and not navigate based on just one of the three. You need to monitor all three and make decisions as it evolves. Zezo's a great tool, but sometimes it should be ignored in favor of what's actually happening in the game.

In this case, I noticed over the last two days that I kept shifting between wanting my start to be TWA 120 versus a straight line to the first corner. It was significantly changing with each forecast update. So I wasn't surprised when it shifted again just before the start, and I routed based on what I was seeing in the game rather than Zezo's recommendation.

Post by BooBill | 2020-06-25 | 13:51:01

Thanks for that @YourMomSA. I wasn't finding fault with Cvetan, if anything the router had the correct winds, the game did not. I'm a real sailor, so actual wildly at odds with forecast is something I expect.

I saw the exact same things and debated the same strategies you did. Ultimately I went with the rhumb line plan, I see you went with the 120TWA plan. If you hadn't answered the question, I was just about to specifically address it to you. How did you know? As soon as the first server update occurred I realized the problem, but that was too late, I couldn't avoid a sail change.

If I understand you correctly, you just chose and happened to choose right, whereas I did not. There's no information source I'm missing?

Post by YourMomSA | 2020-06-25 | 14:27:39

I'm in the US, so the start time was 6 AM in my time zone. I don't lose sleep for the game very often. Usually only in coastal TWA situations (like short-tacking a coastline). I usually miss the wind update that occurs at 4:30 AM my time and adjust when I get up the next morning. No different here. I went to bed with a straight line set to the first corner, but with it in my mind that depending on the 4:30 AM update, I might prefer TWA 120.

I happened to wake up at 5:50, so I checked it on my phone. Zezo said compass 130 was TWA 129 and VR UI said it was 109, and I didn't have dashboard and I didn't want to disturb the family by walking to a computer... So I took a "trust neither" attitude toward it. Reverted to my personal preference to avoid sailchanges, and also relied on knowledge that this boat likes TWA 120 better than anything else in these windspeeds. So I set it to TWA 120 and went back to sleep, knowing my alarm was set for 6:40 to get up in time for work anyway.

So... how did I know? I didn't. But it wasn't dumb luck either. It was falling back on experience under uncertainty, which gave me better odds of being right.

FWIW... I doubt the difference will be big. A sailchange at 6 knots boatspeed with pro winches costs about 0.06 nm. Although the boatspeed at TWA 120 is notably better than at TWA 115, so maybe it's worth 0.1-0.2 nm.

Also... It's important to be true to your own personality and preferences under uncertainty. If you're happiest hand-steering based on the VR UI until things sync, great. If you're happiest following the Zezo route and hoping VR syncs with it fast, that's fine. If you're happiest sifting through the dashboard truth to figure out what that means relative to UI and Zezo to try to mentally route it, have fun trying. But in the end, you're always going to be wrong sometimes, and you'll be more content with your mistakes if you followed your own instincts.

Post by LinusVanpelt | 2020-06-25 | 14:38:19

For my part, at the start, zezo was blocked at a forecast +9h. I hesitated "a long time" between TWA 90 & 125. Finally observing situation on VR & thinking after, i started at 118� to avoid a change sail ;)
Just a good feeling without any other information, but if you have some, i take ;)
Good wind all

Post by MidnightFoiler | 2020-06-26 | 06:19:36

I had similar process to YM while trying to get our 2yr old to go to bed! Things kept changing on me and in the end I just put 120 into VR and saw that was pretty close and I could avoid a sail change.

We can go either side of Ille d'Yeu on the way back right?
Despite the lines in VR and dashboard suggesting otherwise. Only 1 gate there and we already went through it.

Post by BooBill | 2020-06-27 | 15:23:47

So, I did a little post-race analysis on the last leg. I really wish better tools were available to do this, hint hint.

I rounded the last mark placed in the mid 30s. It's really difficult to tell exactly where you are, but I benchmarked against other people that were within a few boat lengths at the time. I found 3 strategies.

Winds were expected (and experienced) to head us about 10 deg and build significantly as we made our way north.

The obvious strategy was the rhumb line. Sail the shortest distance. It all could be done on C0, so no sail changes. Starting a little abaft of the beam and finish a bit forward. YC6211 (hi I see you around here) appears to have done that and gained 2m20sec on me and about 20 places in the fleet.

My strategy was to bear off coming around the mark, get into the wind sooner, head up after half an hour to sail a bit above the line, then bear off and go for a blast the last 5 miles. Sailed the longest distance as a result and lost about 20 positions. BTW, this is what the router suggested.

The third strategy was to head up immediately after the mark as much as land would allow and then go for the blast reach at the finish. Toppen appears to have taken that strategy. He finished about 4 minutes and 42 positions ahead of me. Clearly the correct strategy in hindsight.

I'm interested if anyone else looks at this kind of stuff and finds it valuable? Anybody see it differently than me?

Post by YC6211 | 2020-06-27 | 23:47:32

So I actually did come up a little bit after the rounding mark. There was a balance that needed to be struck between getting into better wind and not sailing too much extra distance. I think it was very boat dependent. The polars on these boats show that the boat gets faster and faster the deeper you go (until about 120TWA) so sailing higher into more pressure is also sailing slower. If we were sailing something like Class40s where the fastest point of sail is closer to what we had, it would have been a no brainer to go futher west into wind.

So essentially what I did was split the strategy. I sailed 98TWA which was about 5 degrees below course on fixed TWA. Once i'd lifted to where the bow was pointed at the finish I changed to fixed heading. Then I did a big bear away 2 minutes out to gain speed at the end.

The stratagy seems to have worked as rounded in 22nd (I think) and ended up 17th!

Post by LinusVanpelt | 2020-06-27 | 18:00:54

Hi i'm interested but unfortunately with the same observation, same strat�gie and therefore the same conclusion (about 20 places lost). It's only when i wake up one hour before finish line, after a short sleep, i saw the damage : too late ! just after finishing i still wonder why not have chosen this strategy now evident but a bit late for me: clairvoyance/talent of those who knew how to see it at the right time !

It was my second "big mistake", the first was on mark 1 where i prefered assuring to not landing, the passage marks partially masked the coast line. Not aground but few minutes lost irratrappable in a such race with such sailors. After i rased the stones, don't missed buoys, and final we are knowing, so i'm really happy from my result in top 100.

A very interesting race, learning using different means and almost making choices between them : VR, zezo, feeling,...
I do my experience.
Thanks all sharing your visions, analyses.

Post by YourMomSA | 2020-06-27 | 18:22:55

It's very hard to gauge exactly how things worked out because it's so hard to tell how you're doing relative to other boats at any point before the finish. But I took the "sail high early" approach more aggressively than Toppen, and I think it worked out positively. My recollection is that the vector into the building breeze was something like WSW, so I wanted to sail as closely as I could do that without sailing high enough to be slow relative to the fleet. So I settled on TWA 95 for 2-3 hours. Then I bore off to straight-line it to the finish at 11 PM my time, and went to bed. I think I rounded the last mark around 28th-32nd, and I was in something like 41st when I bore off. Finished 19th, so I guess it worked out ok.

This race, though... never had enough dynamics to spread the fleet out. 2-3 minutes at the finish was a huge margin in this one. I haven't re-checked, but I think the margin between the ~30th that I rounded the mark in and the 19th that I finished with was maybe 1:30 or so. In a fleet like that, risk mitigation has to also be part of the plan. Deviating and being wrong can be very costly.

Even if Cvetan had a post-race analysis function, I don't think it would answer this particular question. Zezo isn't granular enough in its analysis, nor capable enough of precise coastal navigation, to pick up on a 1-minute difference over a 3-4-hour leg.

Post by BooBill | 2020-06-27 | 20:45:30

So much for my thought that the tricky VMG bit through the gap occurring middle of the night for those in Europe would advantage us on the other side of the pond. If anyone ahead of or around me touched a rock or sailed a sub optimal VMG then I didn't see it. It all came down to that last reach. Call it right and it was worth 2m20 on the fleet, call it wrong, same thing.

The Dashboard doesn't continuously collect data on those you are following, so, practically the post-analysis couldn't be done there. It's something VR itself would have to support. A slider to control that stupid animation that replays the race (which has a way of crashing my iMac BTW) would be very helpful.

Post by MidnightFoiler | 2020-06-29 | 01:13:19

I would say you lost a little on heading up a bit late at the corner of the island and then going low was not right. VR dashboard map is a nice way to zoom in and see the finer details for a post mortem - just click the boats in the game for the tracks to appear - yours is green.

I was cooking birthday lunch for my wife and niece so had no to time to check much. Experience/gut said go twa to the finish or a bit higher. I compared straight, twa and a few degrees higher than twa and then down again half way in VR router. I knew zezo was too constrained by land to be any use - you can check this by routing to a point on the land you think it should be shaving and it will turn abruptly to get there.

twa seemed slightly better and higher than twa seemed the same so I went with twa. I think the wind update half way gave us a few more degrees so I feel it was about right. I went maybe a degree higher than Arylla and gained a little but not enough for the win ... so happy he finally won one as we often chat - he invited me onto the top of the podium to enjoy the champagne :-)

Interesting that YM did so well with quite an aggressive move west for a short leg. Could just be everyone else being lazy and going straight though ... people often get lazy towards the end I think.... unless they are vying for a win or PB ;-) A good opportunity to make places.

Post by BooBill | 2020-06-29 | 03:18:11

Thanks @MidnightFoiler for the suggestion of looking at the Dashboard map to see the tracks. The colours don't stand out that well, but the detail is more precise than the tracks in the VR client.

Interesting that the dashboard map shows none of us coming close to land around Ile de Re, whereas in the game client it looked as close as you could get. Is the landmass in the game expanded a bit to only allow areas with enough depth to actually sail? Could we have gotten away with getting closer?

My strategy of sailing a bit of an S wasn't just blindly following the router. I always put thought into why the router wants to do something. In this case, bearing off for speed to get to the wind sooner made sense to me, so I went with it. I kind of do trust the router to do the math though. The added speed should have made up for the added distance. Didn't work out though.

For a race that was low on strategic moves and more of a test of how precisely can you maneuver, it sure seems to have been made or lost on that one last strategic decision.

Good race.

Post by MidnightFoiler | 2020-06-29 | 07:02:14

The land is just different in VR. In most areas you can trust the outline in the game but occasionally not! There are tricks to getting the waypoints nice and close ... perhaps your team can enlighten you ... or you come on ours :-P Either way a practice with a 2nd boat or perhaps on a record race is a good idea.

I did realise you weren't just following the router but thought the S would pay. I actually think Arylla won on the long leg from the top by doing a slight S which made sense from the shifts etc ... I went pretty much straight and he gained... it was night for us in Australia.