BOTER-BLOG

We still have lots of bugs on the first work-day of the new year.

* The n225 index has long been being failed to be acquired automatically as expected. So poor Arlene.
* It's very easy to make mistake on day_type within the input for the present day and the append of the historical data as well. We lost some points on this yesterday. I'm thinking to invent a new mechanism by which the near-settlement period is tackled in an automatic manner.
* The historical statistics has never been right since day one. It scares people away.
* The yahoo.com!

Kill bug.

First 3 opening days in 2007 shows that there will be another 365 amazing days in front of us.
Jan. 2, +304 points
Jan. 3, +115 points, touching High 20,623
Jan. 4, -409 points, touching Low 19,991

The happy news is: Day 1 win; Day 2 win; Day 3 win as well.

JK
1/4/07

On Friday, the Futures drops down by 236 points within less than 1hr after opening. And then, it goes up by 455 points and closes at the highest. I guess all the tactics about stop-loss would be unfunctioning in this situation. However, under the Boter's design, the risk of this kinda loss has been digested by the possibility calculation. Up-to-day, the stop-loss-related gain/loss is break-even. So, sometimes I think, the function of stop-loss is to protect the fund from broken rather than earning extra profit..

JK
Shanghai Sunday night

Seems the new year makes everything out of track.

The system said 'uncertain somehow' this morning. I had her fixed in an hour -- that's a bug due to being unable to predict there're so many holidays (Xmas and new year and the past weekend! Not an employer would be happy to see that).

The reason that I needed an hour is the multi-platform (and it's also the reason of making many bugs). I've talked to Archer and he's willing to help.

Leslie

After a couple of days working as a core-duo processor, I finally kill all the bugs found and finish modeling LAPUTA on the freeBSD platform. So far it's simply a beta version and we need time to stablize everything around.

I really hate to transfer the code piece by piece from Windows to UNIX again. It just makes me have an aching neck and this obviously violates Sam's Code III. Again, Arch has his solution and we will try it tomorrow. Arch is always able to find the best thing (e.g., his gf), hope this time Magic C++ is not the exception.

Check the test report attached on the page. It's inspiring.

leslie

Laptua has kicked off pilot for 3 days. Overall it is alright in terms of operation and stability. Except day 1 that HSIF got a big drop unexpectedly at opening, day 2 and day 3 is quiet though the HSI was up and down violently. Most of all, the Laputa copies one of the strengths of MII which is to hold a contact on hand as long as 2 same singals are in a row. It is proven by MII that this rule could help reduce many loss. Eagerly expect the show of Laputa in next week.

JK
Shnaghai, 1/12/2007

- The Laputa's new version, which gets rid of a couple of bugs, is now online after another couple of days' working. Most importantly, it implements the handling of the right-wing of the settlement day.
- The new version appeared to be doing pretty well on the past few days. The right-wing :)
- We deploy the Magic C++ and have it successfully compile our code after many painful hours. Hope it does help in the future. By reading their documents, we found that this company might be just founded by two Cantoneses after peeling of their international-like shell -- yes, it still has some stupid bugs, but hope they can do better as we understand they're rather not easy. Seems even the France Telecom is one of their customer.
- I had some discussion with Arc about how to smooth the modeling process. Now I have a plan.

leslie

Seems these days the HSI is simply drawn by the model lap_v1.0_070114. This model has contributed more than 1800 index points since the first trade-day of the year. It almost copies the miracles happened in the debut of m_II and Arlene.

leslie