Why Don’t Your Pick Recommendations Choose The Top Option For All My Entries?

We explain how our portfolio strategy works, and why sometimes we recommend for something other than the top option in a particular pool.

We have built logic that enables us to recommend exact picks for all your Survivor entries, across all your pools. It allows us to treat your selections as a portfolio of picks, and not just treat individual picks separately.

However, there’s one question we keep getting asked:

Why doesn’t your pick recommendation page just show the top option for each individual entry? Why does it show lower ranked teams for some entries?

And that’s a totally understandable question! It’s a little confusing at first. So here’s the explanation.

Don’t Put All Your Eggs In One Basket

The pick rankings for each of your individual entries are just that — individual. They only take into account that single entry, and not your picks on any other entries, or in any other pools.

On the other hand, the Pick Recommendations logic looks at all your entries as a single portfolio, and tries to strike a good balance between risk and reward.

Most of the time, when the Pick Recommendations suggest you pick a team that isn’t the top-ranked one for that entry, it’s because the rankings are trying to prevent picking the same team with too many entries.

Just like we never recommend betting your entire bankroll on a single game, we also suggest not picking the same team with every entry in a large portfolio. When you pick the same team with all your entries, you drastically increase the chance that all your entries will be eliminated.

The Math Behind The Logic

Let’s imagine you have 5 entries in your portfolio, and they all have the same top 3 picks (we’re ignoring public pick data and future value in this example, but the same basic concepts apply even if you take those into account):

  • New England: 80% win odds
  • Pittsburgh: 75% win odds
  • Kansas City: 70% win odds

If you pick New England with all 5 picks, your potential outcomes are:

  • 80% … 5 entries left
  • 20% … ELIMINATED
  • (Average of 4.0 entries remaining)

On the other hand, if you split your entries with 2 on New England, 2 on Pittsburgh, and 1 on Kansas City, your potential outcomes are:

  • 42% … 5 entries left
  • 18% … 4 entries left
  • 24.5% … 3 entries left
  • 10.5% … 2 entries left
  • 3.5% … 1 entry left
  • 1.5% … ELIMINATED
  • (Average of 3.8 entries remaining)

So, by splitting your picks, you’ve slightly decreased your average number of expected remaining entries (from 4.0 to 3.8), but you’ve massively decreased you chance of total elimination (from 20% to 1.5%).

That reduction is risk is most of the point in entering a large portfolio of picks. If you keep stacking all your entries on one team — even if that team is a large favorite — the odds will most likely eventually catch up to you, and all your entries will go out at once.

For example, if you put all 5 entries each week on a team with 80% win odds (extremely high for Survivor), after three weeks, there’s a 49% chance you’ll be totally eliminated.

On the other hand, if you spread your entries like above, after three weeks there’s a only roughly a 7% chance you’ll be totally eliminated.

You Also Don’t Want To Spread Your Picks Too Thin

This is less common, but another reason we sometimes don’t recommend the top-ranked pick for an entry is because we don’t want to pick too many different teams in your portfolio.

The math here is essentially the opposite of above — if you pick too many teams, you’re almost guaranteed to have at least one of them lose.

Let’s use the above example again, only this time let’s split the picks over 5 teams (we’re even going to give those teams good win odds, so the average teams surviving doesn’t change):

  • New England: 80% win odds
  • Seattle: 80% win odds
  • Pittsburgh: 75% win odds
  • Green Bay: 75% win odds
  • Kansas City: 70% win odds

Picking those teams each once, here are your potential outcomes:

  • 25% … 5 entries left
  • 40% … 4 entries left
  • 25% … 3 entries left
  • 8% … 2 entries left
  • 1% … 1 entry left
  • 0.1% … ELIMINATED
  • (Average of 3.8 entries remaining)

In this case, you’ve only slightly improved your total elimination rate (from 1.5% to 0.1%), but you’ve decreased your chance of having all 4 entries survive from 42% to 25%. If that decrease had come with a lot of additional safety at bottom of the distribution, it would be worth it. But moving from 1.5% to 0.1% is simply not a big enough change to justify the worst top-end results.

It’s a Balancing Act

Spreading your picks across too few teams, or across too many, can both have detrimental effects on your long term pool outlook.

Our Pick Recommendations logic tries to manage your portfolio risk levels in order to achieve the right balance.

Often, that can mean taking a slightly lower ranked team from your pick rankings, in order to avoid having an overall portfolio that’s too skewed towards one team (or spread over too many). And if your pick rankings are different for different entries or pools, that may even mean taking a team that’s ranked lower than the number of unique teams in your portfolio (i.e. taking a team ranked 4th on an entry in order to stick to using only 3 teams across your portfolio).