Selection

[Bubba]

Dear all,

It's very close to the time to select the upcoming TSP Calc plan for the following year. I'm curious to see what people have selected as their path forward.

When responding, please include which plan (number) you've selected and why.

Personally, I'm still on the fence. Some of it depends on market timers that I look at.

Best,

Bubster

[michigande]

I've always kept an eye on the dailies, never truly followed one, but here's my $.02

The Seasonal Blind Tests make the most sense to me until the builder is back up and running. My understanding is that it's the only way to truly see if a strategy actually worked against real historical data by using X years of available data, say 2004 to 2014 to build a strategy, then using the remaining years of "blind" data that the strategy was not informed of and seeing if it worked on those years, 2015 to present in this case. The problem with this approach is you're using the already few years of data we have available to test the strategy against so I'm curious to hear what people think the sweet spot would be in terms of how many years of blind data is optimal?

By this reasoning, I've kept an eye on one of MJ's blind tests, #136637 from his Case 10: Highest Mean, GCSI Only. This strategy excludes 2020, an anomalous year in my opinion and has 13 years of blind data as I understand it.

#138970 is another I like from Case 30, Highest Mean, 10% Max SD, No I fund. This should use 11 years of blind data, also excludes 2020. Unless this December is blockbuster, it will fall outside the Standard Deviation, but blind years up to 2019 are at 3.85% Standard Deviation with a mean of 24% and up to present is at 8% deviation with a 19% mean.

[Bubba]

I've always kept an eye on the dailies, never truly followed one, but here's my $.02

The Seasonal Blind Tests make the most sense to me until the builder is back up and running. My understanding is that it's the only way to truly see if a strategy actually worked against real historical data by using X years of available data, say 2004 to 2014 to build a strategy, then using the remaining years of "blind" data that the strategy was not informed of and seeing if it worked on those years, 2015 to present in this case. The problem with this approach is you're using the already few years of data we have available to test the strategy against so I'm curious to hear what people think the sweet spot would be in terms of how many years of blind data is optimal?

By this reasoning, I've kept an eye on one of MJ's blind tests, #136637 from his Case 10: Highest Mean, GCSI Only. This strategy excludes 2020, an anomalous year in my opinion and has 13 years of blind data as I understand it.

#138970 is another I like from Case 30, Highest Mean, 10% Max SD, No I fund. This should use 11 years of blind data, also excludes 2020. Unless this December is blockbuster, it will fall outside the Standard Deviation, but blind years up to 2019 are at 3.85% Standard Deviation with a mean of 24% and up to present is at 8% deviation with a 19% mean.

Nice ideas.

I'm also interested to see what MJ is suggesting for this next year. I always like to look at his blind tests.

From what many "experts" are saying, it would seem that over the next few years emerging markets and international stocks would fare better than US stocks...but then Vanguard has been saying that for years...from that perspective the I Fund might be a good tool in your skillset (note that I know that the I Fund follows the EFA...which is not including the emerging markets).

I'm really on the line of enjoying my default...which is #131784.

Best,

Bubba

[pilotric99]

In a rising interest rate environment, I am leaning towards shifting to a zero F fund seasonal. A quick search shows 160726 as having a good mean (31.5%) with a relatively low SD @ 7.7. its also performed well this year, clocking in at 22% YTD. Its in G 35% of the time and I only 12%, as well.

Just a thought.

[mjedlin66]

The Seasonal Blind Tests make the most sense to me until the builder is back up and running.

I have been working on that for the past week. It is running now and is up to date through November. It is a bit slow.

I haven't done any manual data checks yet and I am working on making it faster in the next few days.

[KTSPcalc]

I see minimum yearly return as equally or more important than StdDev (though I respect I am in the minority). To me, it should matter less that a strategy has a high StdDev if the spread is across a set of higher percentage returns.

For example, I am tracking 159932. The strategy has a relatively high stddev of 8.61, but the lowest return from 2004 to 2020 is 2018 (24.14%).

In the past, I've suggested TSPCALC add a filter for minimum yearly return in the same way that you can choose a maximum total StdDev. Filtering strategies by minimum yearly returns while trying to maximize CAGR would (hopefully) allow you to build a strategy with the highest historical minimum returns even as the stddevs go up. I've been trying to get there by trial and error, but that would be a more direct path. MJ has put this in the someday category.

In any case, I haven't decided on my F risk appetite for 2022, but here are the strategies I have my eye on (data listed is from 2004 to 2020):

0 time in F = 160726, CAGR = 31.8, stdev = 7.54, and min. = 22.68 (2004). 2021 to date = 22.15.

8% time in F = 161530, CAGR = 33.15, stdev = 8.68, and min. = 23.13 (2018). 2021 to date = 22.07.

14% time in F = 159932, CAGR = 33.74, Stdev = 8.61, and min = 24.14 (2018). 2021 to date = 21.35.

[12squared]

My current favorite is #159757, created: 2021-10-29
https://tspcalc.com/seasonal.php?ID=159 ... 637-138970>
It evolved from looking at strats which performed best in 2015-19 when the Fed was not buying bonds. I constructed it by manually splicing "better" months from these and other strats based on their ratio of Mean to SD during those years. May is its weakest month, due to losing more than 6% in 2019. However, none of the stats endorsed so far had an up May '19 either.

FYI, I've never figured out to get TSPcalc's seasonal builder to do what I want for just one month in any finite amount of time.

96778 has even better performance from 2015-19, and a higher low year. However, it was negative from 2007-09. 159757 isn't perfect. It returned in in the single digits form 2006-08.

12²

[Bubba823]

I'm probably sticking with 70% C and 30% S. Prior to November my PIP was at 44%. The only change I made was to CA which I just set at 100% S.

[bloobs]

I'll probably stick to one with a moderate mean with low stddev rule for my TSP--which seems to have done the opposite this year lol. Dang outlier.

But across my entire investment portfolio of IRAs, a 401k, and taxable accounts ascribing to seasonals, I pick a *variety* of seasonals that span the gamut of high and low risk and return profiles. This usually entails picking a strat that goes into F/G at around the same time another seasonal is diving into C/S/I-- reducing market timing risks. So if one strat blows, another one or two should counteract it.

[Bubba]

I'm also curious to the beer bet from 2022. It would seem like Aitrus is winning this year, although I haven't looked at the recent stats due to the selloff.

Personally I have switched my strategy's moves to F this year with G, which didn't always bode well. Other times it was the right call.

Thanks to all whom answered. It's interesting to see yours answers and reasoning. I'm curious to Aitrus and MJ and the upcoming beer bet, but maybe we're too early in December for that one.

Thanks,

Bubba

[Aitrus]

I'm also curious to the beer bet from 2022. It would seem like Aitrus is winning this year, although I haven't looked at the recent stats due to the selloff.

Sorry to ruin the fun for you Bubba, but I didn't partake in the beer bet this year. Nor do I plan to in the future. Let me explain why (a few have asked me why my Fantasy shows G nowadays, so I figure now's as good a time as any).

My role here is to provide forum moderation and advice on how Seasonal and TSP works - nothing more. A while back, I became aware via PMs that my standing as moderator gave some folks the impression that my moves were better, more researched, or more appropriate than others. Meaning, that because I was moderator I was therefore also an authority on investing. I was functioning as a sort of Jim Cramer or Warren Buffet for more than a few people. They were making IFTs according to what my own moves were, and remaining blind to why I was making them - or even if they were good moves at all.

I'm not an investing guru, nor do I even pretend to be. If I were, then I'd quit my day job and go into trading and market commentary full time a-la Ira Epstein or something. If that's how I'm coming across, then that's a mistake on my part. I just stumbled on an old idea from the forum's past and decided to carry it forward because 1) it was interesting to me, 2) it was interesting to others, and 3) it helped people understand how the market works better (and myself as well as I went through the process of learning). MJ then carried the idea further with TSPCalc. I'm no genius, I'm just a regular guy - just ask MJ. I'm well versed on a very specific method of investing and have a minor talent for writing, nothing more.

As a result, I decided to remove my personal IFTs from public view. I still follow a seasonal strategy, but I don't share which one. I don't move anything on my Fantasy anymore, so it just shows G all year long. As far as the beer bet goes, Jahby's Mix is doing decently well for anybody following it, but I'm sure there are other people who are doing better and thus should be the ones winning the bet.

[Bubba]

I'm also curious to the beer bet from 2022. It would seem like Aitrus is winning this year, although I haven't looked at the recent stats due to the selloff.

Sorry to ruin the fun for you Bubba, but I didn't partake in the beer bet this year. Nor do I plan to in the future. Let me explain why (a few have asked me why my Fantasy shows G nowadays, so I figure now's as good a time as any).

My role here is to provide forum moderation and advice on how Seasonal and TSP works - nothing more. A while back, I became aware via PMs that my standing as moderator gave some folks the impression that my moves were better, more researched, or more appropriate than others. Meaning, that because I was moderator I was therefore also an authority on investing. I was functioning as a sort of Jim Cramer or Warren Buffet for more than a few people. They were making IFTs according to what my own moves were, and remaining blind to why I was making them - or even if they were good moves at all.

I'm not an investing guru, nor do I even pretend to be. If I were, then I'd quit my day job and go into trading and market commentary full time a-la Ira Epstein or something. If that's how I'm coming across, then that's a mistake on my part. I just stumbled on an old idea from the forum's past and decided to carry it forward because 1) it was interesting to me, 2) it was interesting to others, and 3) it helped people understand how the market works better (and myself as well as I went through the process of learning). MJ then carried the idea further with TSPCalc. I'm no genius, I'm just a regular guy - just ask MJ. I'm well versed on a very specific method of investing and have a minor talent for writing, nothing more.

As a result, I decided to remove my personal IFTs from public view. I still follow a seasonal strategy, but I don't share which one. I don't move anything on my Fantasy anymore, so it just shows G all year long. As far as the beer bet goes, Jahby's Mix is doing decently well for anybody following it, but I'm sure there are other people who are doing better and thus should be the ones winning the bet.

Sorry to hear that. I can imagine that you felt somewhat bothered when people would PM you as well as when they would complain that some of the moves didn't work.

Good luck on your moves, whichever they may be.

Best,

Bubba

[Aitrus]

Sorry to hear that. I can imagine that you felt somewhat bothered when people would PM you as well as when they would complain that some of the moves didn't work.

Good luck on your moves, whichever they may be.

Best,

Bubba

Them contacting me for advice didn't bother me. What concerned me was that they were asking me - a total stranger - to make their financial decisions for them, and they wanted to blindly follow whatever I told them. I don't mind providing advice and my view on things as part of their research and investing consideration. We should each be researching and learning as much as we can so as to make wise decisions with our money.

However, I'm not a fiduciary. I don't have the training or credentials necessary to make monetary decisions for other people. It's an ethical issue for me, not an issue of convenience.

[mjedlin66]

This may be relevant to the discussion:

https://tspcalc.com/blog/?p=43

[mjedlin66]

Also this, a look back at the 2020 beer bet strategies, and how they would have done this year:

Note the 2020 winning strategy would have won 2021 too and it only uses C&S...

https://tspcalc.com/seasonal.php?ID=82& ... 104-111961