signal discrepancies

[Derrick]

I’ve noticed some discrepancies in signals lately and I’m wondering if someone can explain what may be happening. The Nasdaq 100 strategy is one example.

The November email had:
20% CA + 20% ESRX + 20% FOXA + 20% SBUX
However, QuantTrader currently says Nov. was:
FOXA, ROST, SBUX, WLTW
Same discrepancy in December:
email was ESRX, FOXA, MDLZ, WBA
Quantrader: FOXA, KHC, MDLZ, SBUX

The issue is more pronounced when trading Strategy of strategies on Quanttrader. I’m not understanding why those allocations seem to change so often. For Example, At the beginning of October it allocated one third to MYRS, now Quanttrader shows BUG in that space. Since October was such a bad month this created a substantial difference in the returns between the conservative BUG vs. aggressive MYRS. As a result of discrepancies like these, I did not get anywhere near the same returns for 2018 that Quanttrader shows. I have been a subscriber since 2015 and I do not remember ever encountering this issue to this extent.

Thank you.

[Vangelis]

Derrick,

Since you have been with us you know our main strategies are changed/updated rarely. Maybe every few years. This past two months we have introduced 2 new strategies, and a new Hedge sub-strategy that affects all existing ones. When QT runs a backtest with the latest parameters, it just does that. Produces a backtest. So all allocations are based on the current parameters. So under the new Hedge strategy, N100 would have invested in TLT not UGLD. That does not mean that the signals issued at the time (UGLD) were wrong. It just means that we (and QT) are adapting to the markets and finding better solutions going forward. This happened in the past but was less pronounced. The actual signals and performances are here https://logical-invest.com/app/historical.php
Also see a better explanation here: https://logical-invest.com/backtested-data-vs-historical-data/
I hope this makes sense.

[Frank Grossmann]

In November CA was bought by Broadcom and in December ESRX was bought by Cigna. Once these takovers have been done we had to update the N100 company list and recalculate the ranking. This ranking is then obviously different and somebody who starts the strategy mid month would have had to buy other companies than at the beginning of the month.
This is ok and does not mean that the 4 stocks at the beginning of the month are wrong. Anyway if you use QuantTrader you will see that intra-month there can be quite some changes in the ranking especially when we have such high volatile markets.
Normally when there is no takeover, you don’t see these changes on the website.

[Derrick]

Thank you for your responses. This makes sense. The link to the historical signals is very helpful.

[Author]

Posts