Honda has revealed that following positive progress on the dyno, it was only when it hit the track that it realised how dire its situation was.
"We introduced a new power unit concept this year, so I would almost call this Year One again," says Yusuke Hasegawa. "Our aim is to develop the 2017 concept into the 2018 season and hopefully 2019 as well.
"So the engine weight, centre of gravity and the combustion concept is all going in the same direction as the other three engine manufacturers.
"It was good for us to do that," he continues. "We can modify the specification of some of these parts to catch up with the other three engine manufacturers.
"Last year the engine concept was completely different, so with minor modifications we could not duplicate the same type of performance. That's why we really needed to change the whole engine concept this year.
"Roughly speaking, it takes almost a year to design a completely new engine. So that's why we started 2017's engine development last May.
"So, this year's power unit was built up at the end of last year. In parallel, we also conducted some mono-cylinder tests and some experimental tests on other engines. However, when we fired up the complete engine for the first time, we could see it wasn't delivering the durability or performance in accordance with our expectations. We also found many minor issues. So we needed to modify tiny bits.
"After resolving these smaller elements, we started to test the full concept at the start of this year, call it Spec Zero as it was the initial one - and before the first winter test we confirmed that it ran on the dyno.
"Of course, at that moment we knew that the power was not delivering to our target. Then, at the Barcelona test, we found more issues on the car, such as the oil tank issue. It was a car-related issue. This is not a complete engine issue, but of course it is very important."
For those wondering why these issues were not spotted before the car hit the track, according to Hasegawa the answer lies in the difference between testing on the dyno and running in a brand new car on a track.
"Many items we could not test on the dyno, so it is normal that we need to check some functions in the car. The oil tank is one of the biggest items, so we have a rig for the oil tank but we cannot recreate the same types of G forces and conditions as in the car. Of course, by design we have to consider the actual car situation in theory, but sometimes it is not always the same situation so that is why we had some issues with the oil tank first.
"The second issue was down to the vibrations. On the dyno, the model is stiffer and heavier, so it doesn't create any synchronised vibrations, but on the car - with the gearbox and the tyres - there is a much lower level of inertia. Low inertia does not always create vibrations but it's completely different from the dyno and that's why we suffered a huge vibration on the car. Of course, we were aware some level of vibration would come in the car but it was much bigger than we expected."
While Honda began work on the current issues, there was still the need for in-season development and attempting to find a greater rate of improvement than its rivals. However, as Hasegawa explains, this is easier said than done.
"It depends on the items," he admits, "something like the combustion system takes longer for testing and confirmation. So we cannot set a target just for two weeks later, normally such an upgrade takes something more like half a year.
"But it is also important to make those kinds of developments as well, so actually the Spec 3 power unit is like that. This development started in March, more than three months ago."
Spec 1 was the version raced in Australia, while Spec 2 was then introduced at the Spanish Grand Prix where Fernando reached Q3. Spec 3 had its first run in Baku before being raced in Austria and taking Vandoorne to Q3 for the first time at Silverstone.
"Spec 2 was introduced in Barcelona, and we were already working on Spec 3," says Hasegawa. "Even at that time we knew there were some good items that we were working on but they were not ready at that moment, so we needed another few months to introduce them."
In previous years, development was restricted by the infamous token system, but 2017 has seen that system removed, and as a rsult opened up the opportunity for Honda's Sakura HQ to push on with this year's concept.
"If we still had the token system we wouldn't have been able to change the whole engine modification for this season and also introduce the Spec 2 and Spec 3 power units," admits Hasegawa. "I haven't counted how many tokens it would cost so I don't know for sure, but maybe it would be difficult to modify and introduce the current Spec 3 engine using last year's token system."
Asked how often developments being introduced to help close the gap, Hasegawa admits: "Almost every race, but we don't always call it an upgrade because sometimes it is a countermeasure for the durability or reliability issues. Also for weight reduction we improved many areas.
"A good example is that last year we changed the induction system from aluminium to carbon. That type of update is easier to introduce, but the core performance has not changed. So when the material changes or something like that, that's a different type of update.
"I get excited when we bring an update," he admits, "because it can lead to increased expectations of scoring points or similar. But still we need more from our other updates to catch up to the level of Mercedes and Ferrari.
"I am definitely confident that we are closing the gap to the leaders," he insists, "so, from that point of view, our speed of development is good. But at the same time it is natural for the follower because you have a target that you know is achievable."
While work continues on the Spec 4 unit, development is simultaneously ongoing for 2018, as Hasegawa believes the current concept holds major potential for the coming seasons.
"We don't stop developing, we need to keep updating. Of course the performance and results are the most important things but it's all learning for the future too. Compared to last year we needed to modify the engine concept, but next year we will keep the same concept.
"It's good that we can use the same concept," he admits, "because this year's development and improvement is directly connected to next year. So that means we don't need to stop the current development, and from that point of view we have already started next year's design."