FWIW, I did quite a bit of my own analysis on the Presidential Cycle. What I did was somewhat like what Burk did, in that I broke down the cycle by month and year, and did some statistical analyses.
I have a lot more to do, but in a nutshell what I found so far was this:
First, I looked at the COMPX (the data I had at the time on the cycle was from the SPX). If one goes long the SPX during the 27 month long part of the 4 year cycle, the returns over 14 cycles were always profitable with the SPX, and averaged 40% per cycle. For the COMPX, returns were about 52% per cycle. Like the SPX, there were no losing cycles. The range of returns was 13% (during the cycle that included the 1987 crash) to 75%. If one excludes the 1987 cycle as an outlier, the returns averaged 50% per cycle. The cumulative returns since 1970 over 9 cycles would have been over 3600%, which works out to an annualized yield of a bit over 100% per year.
Next, I began to wonder if the long strategy could be improved by eliminating months that tended to give lower returns. I noticed that 7 of the 27 months in the long part of the cycle showed average negative returns, and the win percentage was close to or less than 50%. So, I asked what happens to returns if you eliminate these low-yield or losing months, and just went to cash at these times. Taking that strategy, the sum of the returns improved from 471% to 608%, and the average cycle return improved from 52% to 68%; time in the market decreased to 42%, and 58% of the time you are in cash. Not bad.
Next, I wondered whether one could take a short position during the 21 month portion of the cycle where normally one would stay in cash. It was immediately apparent that simply taking a short position at the end of the long portion of the cycle would not work consistently. But could specific months be identified that yielded consistently bad returns?
It turned out that the answer is yes. There were 4 good candidate months, where a short position produced an average return of 5% per month (range, 3.5% to 7.0%). The win rate for these losing months with a short position ranged historically from 68% to 78% (average, 74%).
So what does that do to the returns from the full cycle, when you add the short position to the long position?
IF you take a position in the COMPX using a 200% leveraged mutual fund, then the returns per month average 10.1% for each month you are invested (you are invested either long or short 24 out of 48 months, or 50% of the time).
The simple sum of the returns over one cycle would be 243% (60.8% annualized yield). The cumulative yield over the 4 year period would be 876%, so $10,000 would grow to over $97,000.
Now, a few caveats. First, there is no way I know of to invest in the $COMPX. Probably the closest equivalent is the RUT. Second, these return calculations did not try to include returns from being in cash, which would have boosted the overall calculated returns somewhat, since the strategy calls for a cash (or fixed income) position 50% of the time.
Among the analyses I still want to do are to look at specific indices such as the RUT and the NDX, and maybe a sector like the SOX. I also want to examine whether maintaining a position in multiple indices at once provides any advantage, in terms of possibly lowering volatility and drawdowns. It seems reasonable that this might be so, since we know that the cycle works historically with at least 3 indices (Dow, SPX, and COMPX), but these indices do not move in lockstep. If so, I suspect that reduced overall volatility might come at the expense of somewhat reduced overall returns, but that remains to be seen.
Anyhow, I am encouraged. It is potentially a very low maintenance approach---there are only 16 transactions to make in the entire 4 years, or an average of about 1 transaction every 3 months. And, it is "blind", and requires no watching or analysis of the market at all (unless one wanted to incorporate a stop-loss strategy).
First month in this form of the cycle would be this coming July, so I want to complete all the analyses I want to do by then. The traditional approach to the cycle calls for an initial long position at the beginning of October of this year.
WS
Market Data 
Markets 
AI
