Off-season: Explore the unlocked site! Projections reflect re-calculations of 2022, based on end-of-season knowledge.

Survivor Pool 2021 Example

In my Survivor Strategy article, I outlined the main features of my approach. The goal of this article is to explain how I will present the advice from my algorithm, so you can understand the optimal paths recommended. There is a logic to my choice of this format, which I will also explain. Please note the actual visual output is in progress (it is a new feature).

Key Features of my 3 Survivor Pathways

I will show 3 picks each week, each pick corresponding to a Survivor path with optimized expected longevity.

  • The 3 picks will all be different; no overlaps.

  • The paths are tiered (e.g. the gold path gets 1st priority).

  • A “back-up” pick is given for each path, for consideration as an alternative.

    • The back-up is simply one of the 2 other picks, if possible. If both teams were already used by the path, then a 4th option is given.

  • Whenever a main path fails (expected 3-4 times in each path), that path will continue next week as if the “back-up option” had been chosen.

This strategy with back-ups is to ensure my algorithm always highlights 3 options that are tied to a successful pathway for the duration of the season. It turns out “3” is a good number for securing a 90% chance that I can continue the line of advice.

Example from 2021

I have simulated 2021, to create 3 pathways and show how they proceed weekly. (For the sake of illustration, I assume perfect knowledge of the season from the start, to form the optimal paths.)

It turns out the first path failure occurred in week 6. So the spreads chart below shows the 3 sequences until week 7.

Here’s what to notice:

  • Firstly, take a look at the left-most lists. They are the 3 paths as generated from week 1, but clearly they must get modified over time: the algorithm avoid overlapping teams for the current week and for the subsequent week. So paths 2 and 3 will get updated over time.

  • Look at the colored boxes in the main chart. From left to right, it shows which teams were chosen.

  • Notice the colored dots, which show back-up recommendations. They were not continued as long as the main paths were winning.

  • Finally in week 6 was the first failure seen: Bills lost to Titans. So the Silver path ended, and week 7 proceeds as though the Colts had been chosen.

For an overview of the all the weekly adjustments made 2021 simulations, zoom into the below table. Only weeks with a failure are shown, and the each failure+correction is highlighted in blue:

There were 9 total failures among the 3 pathways, giving a failure rate of 18% (82% success). Hopefully this article gave you some idea about how the Survivor Pool table will continually help with multiple options each week, adjusting for failed paths.

The Challenge in Presenting Survivor Advice

You might be wondering why all the complexity.

The advice cannot be so simple as to show a single path, or single pick each week:

  • The team might lose, ending your season. Then my series of weekly recommendations would also end.

  • A suggested team might not be available to you, if you chosen that team already while following a different path.

  • You might just want to choose a different pick from “everyone else”, as a strategy to better your odds.

  • If you are playing multiple survivor pools, it is better to diversify— therefore preferable to have multiple paths.

So it’s better to show multiple picks each week.

However, presenting multiple picks also brings problems:

  • The multiple optimized paths will naturally overlap often, bringing back all the problems above.

  • There are millions of possible paths, but showing too many would appear chaotic and would not be true advice.

  • Even showing multiple paths, there is a very high certainty that any single one will fail before the season end.

My method aims at a compromise between simplicity and versatility.


Week 8

Tagged under