Mystery of investment index that defied gravity

Below is what Brian Woods, who some on this forum may know, posted this afternoon on LinkedIn about an index devised by BNP Paribas.

The Index, which Colm Fagan and I call the Deep Value (DV) Index, was created by BNP Paribas. We have completed a detailed study of its performance.
The graphic below illustrates the mystery.
BNP Paribas launched the DV index in July 2015. In their presentation they explained how, by using a novel technique of enhancing the number of cum-dividend shares in a price only index and other aspects such as low volatility, options on DV would be cheaper than on its benchmark EuroStoxx50 (ES50). When they simulated this DV creation back to 2001 they found the remarkable gravity defying result that DV had grown to c.176% while ES50 had declined to c.74%, a relative outperformance against its benchmark of c.237% despite the inbuilt cum-dividend drag!!
For the first few years post-launch DV matched its benchmark despite this inbuilt drag. Then (possibly with very unfortunate timing) Broker Solutions launched the Secure Accelerator Bond (SAB) series in Ireland. This would pay 40% plus, if DV simply did not finish below its starting value in five years. 1,304 back-tests in the brochure showed that the 40%+ would always have been achieved. Colm and I complained formally to the Central Bank that the brochure was potentially misleading but it was a few years before they got round to addressing the issue, too late for the investors in SAB.
Back to DV vs ES50, after the initial quiet period of going live, gravity asserted itself and despite the benchmark very comfortably surpassing its initial value in the 5 years, DV finished well below its benchmark and crucially below its starting level, leaving SAB customers at a loss!! We are aware that this has completely mystified customers and advisors alike.
The before and after launch experience are so starkly different that Colm Fagan and I decided to test the statistical significance of the difference. We used the actual experience since launch as our data base and boot strapped a million simulations of the pre launch experience. The best result of the entire million simulations of relative performance was 202%, way below the actual 237%.
It could well be that our method is seriously flawed but can AAM contributors who are more expert than us on statistics comment on this?
Activate to view larger image,

And here's what I posted in reply:

Brian, as far as we know, everyone who bought this product lost money. They (and their advisers) believed the story as told by BNP Paribas.
Can BNP Paribas explain how the index defied gravity for 14 years, beating the EuroStoxx50 hands down (when no products were linked to it), then lagged it spectacularly (when products depended on it). If BNP Paribas can't or won't answer our question, will the advisers who sold the product, believing the performance story, or the regulators who oversee BNP Paribas, force them to answer it?

It seemed to match the eurostoxx 50 fairly well until 2020 then the dramatic divergence, we all know what happened in 2020 ,covid, then inflation then the rise in interest rates. So it looks like this fund was doing something silly with government bond investments that was great during the negative interest rare era but then terrible after that ended

@joe sod I think the main thing that OP is mystified at is the out of the park difference between the proprietary index before its launch and after its launch when it became live and the basis for actual retail structured products. @Colm Fagan in 2019 estimated that the "drag" (which incidentally BNP fully concede exists) of the index against its benchmark was around 3.5% p.a. In horse racing parlance it was carrying a handicap weight. As you point out, it seemed to broadly defy this handicap for a few years before the inevitable kicked in. But in statistical terms this was well within range. To defy the handicap over 15 years would be on the cusp of credibility but to actually grow by 76% while its benchmark fell by 24% is in the equal DNA territory of probabilities. Can you think of some paradigm shift in that period that could cause such a black swan with polka dots occurrence?
I don't think the index had any bonds but it had defensive type stocks with good dividend payments so similar characteristics to bonds.