Selected work

Portfolio analytics methodology

Relative-Strength Portfolio Research System

This started with a frustration: relative-strength research gets messy when every chart asks a different question.

What I didComparison logic, research screens, and allocation views
Built withRelative-strength analysis, Beta analysis, Market segmentation
What came out of itThe finished workbook turned several market checks into one sequence I could run the same way each time.
RSPSRelative-strength research
Snapshot
Oct 02, 2024
Horizon
2D
Historical pairwise model snapshot
PairInputsFrameAverageState
ETH / BTC92D−1.00Short
SOL / ETH82D1.00Long
SOL / BTC82D1.00Long
ETH/BTC workbook trace2017 - 2024
2017202020222024
3 pairwise models2D horizonHistorical snapshot

Inside the workbook

The order of questions matters.

I started with the market regime, moved through the pairwise readings, and looked at individual assets last.
9connected research views
5decision stages
3pairwise models
2Dpairwise horizon
RSPSRelative-strength workbook
Snapshot
Oct 02, 2024
Review
Pairwise, regime, selection
ETH/BTC relative-price structureSource-workbook reconstruction
20172019202120232024
AFR model alignmentSource-workbook reconstruction
20172019202120232024
Indicator time coherenceSource-workbook reconstruction
20172019202120232024
Positive model stateNegative model state

I recreated the workbook's visual structure here. The traces do not show prices or performance.

Research sequence

The five questions I worked through.

01Market regime

TOTAL trend gate

02Major allocation

ETH / BTC

03Speculative strength

SOL / ETH and SOL / BTC

04Small-cap budget

OTHERS.D

05Candidate selection

Market cap, trend, beta, score

Recorded readings

Pairwise trend models

These are the inputs, time horizon, averages, and 80/20 rules from the historical snapshot.
Historical RSPS pairwise model states
ModelInputsFrameAverageStateAllocation logic
ETH / BTC92D−1.00ShortETH 20% / BTC 80%
SOL / ETH82D1.00LongSOL 80% / ETH 20%
SOL / BTC82D1.00LongSOL 80% / BTC 20%

Swipe or use Shift + scroll to review all columns.

Candidate screen

Multi-factor selection table

This sample shows how I compared market cap, trend, beta, and the final score.
Selected historical candidate screening rows
TickerMarket capBull USDBull SOLBTC betaSOL betaAvg betaScore
XRP$179,543,392,162110.720.400.562
SUI$12,505,026,341N/AN/A1.040.720.881
DOGE$49,391,582,499N/AN/A1.570.781.182
SHIB$11,097,647,584N/AN/A1.430.731.082
LINK$15,906,687,5791N/A1.020.620.821

Swipe or use Shift + scroll to review all columns.

These values are historical. They are not current holdings, prices, or recommendations.

The question behind it

What I was trying to understand

Relative strength, trend, beta, and market size answer different questions. Without a fixed order, it was easy to jump from one chart to another and lose the reasoning behind a choice.

The standard I set

What a useful answer needed

I wanted the broad market view to come first, the pairwise comparisons second, and the candidate list last.

My part

Where I did the work.

  • I built relative-strength comparisons across BTC, ETH, SOL, and broader market segments.
  • I brought trend, beta, market size, and selection criteria into the same workbook.
  • I created the portfolio-balance and rebalance planning views.
  • I kept each research input separate until the final candidate screen.

Process

How I got to an answer.

Normalize the comparison

I used the same format for every pair so I could compare the readings without changing the rules.

Add decision context

I checked trend, beta, and market size before letting a relative-strength reading influence the candidate list.

Design the planning view

I carried the research into a balance view that showed what a rebalance would change.

Where judgment entered

  • I kept the research inputs separate from the proposed allocation.
  • I compared several market segments instead of relying on one benchmark.
  • I showed historical model states without exposing holdings, balances, or live strategy details.

Takeaway

What stayed with me

The breakthrough was not another indicator. It was asking the same questions in the same order, which made the final choice easier to explain and challenge.

More work

Read the next projectExecutive Deferred Compensation TRS Hedge Model