Does your selection position in your draft matter?

This week, I decided to look into this age-old question: Does your position in the draft selection affect your chances of winning?

And so far, the answer seems to be: YES.

This was surprising to me based on the fact that having the top scorer in your pool was only a slight advantage in recent years. (a previous blog linked here)

The first thing I did was record the draft position of the eventual winning poolie (we call the owner of a fantasy hockey team a poolie) in every pool we have tracked since the 2011/2012 season.

And before I go on, I should define what I mean by draft position, just in case. Your draft position for most pools, is exactly like the NHL's entry draft selection position; that is, in what order do you select your players? Most pools will select in Snake order. For example, if you had 5 poolies drafting, you'd go in draft order for the first round: 1, 2, 3, 4, 5, then you'd snake back selecting in reverse order for the next five: 5, 4, 3, 2, 1, then back to 1, 2, 3, 4, 5 - etc.

Snake order tries to compensate the person drafting last by giving them two selections from a still fairly talent-rich section of the players.

But as I was thinking about this idea, I started to wonder, does it work?

Okay, on to the data that led to the answer, YES.

So first, I only considered pools that were not labeled as box drafts. In our terms, box drafts allow multiple poolies to have the same player on their roster. In pools like that, draft order does not matter. So we tried to limit our sample to pools that had exclusive player selections. Now for full, full transparency, administrators don't have to tell us if their pool is a box draft or a straight draft, so we chose to believe that a pool was a box draft if the same player appeared on multiple teams. This of course would falsely eliminate some pools that allow player trades, but we need to simplify our sample selection somehow.

Figure 1. Number of winning teams that selected players based on draft order. Since pools have a variety of numbers of poolies, the draft position is normalized here with 0-10% meaning they drafted in the first 10% of players in their pools, and 90-100% indicating they drafted last (or near last for large pools).

So, I pulled the data, sorted, normalized and plotted the data and got Figure 1. Now some things made sense to me, for example:

• Drafting early in the regular season seemed to be more of an advantage than in the playoffs, where it seems that picking high scoring players on winning teams matters more than picking a high scoring player that will be bounced in 4 games (or so we think right now - a future blog will look at that!)
• For the most part, while where you select in the draft matters, it's not an overwhelming, merely a few percentage points.
• EXCEPT
• If you just so happen to draft not in last place, but ALMOST in last place. Like if you draft 2nd last, you might as well leave the pool it's such an outlier.

So that concerned me. As well, there are other patterns where the number of winners seemed consistently low, like in the 30-40% and 60-70% ranges.

That got me thinking that maybe the under-representation there is due to the normalization - for example in a 7 poolie pool, none of positions will ever fall in the 30-40% range; position 2 is in the 20-30% range and position 3 is in the 40-50% range. So maybe that's why second last is so unlikely?

Okay, so we need to really look at pools with specific numbers of poolies in isolation - that is only compare pools with 8 poolies to other pools with 8 poolies and see how draft position affects your chances of winning. And to keep the number of graphs low, we just looked at pools with the most common number of poolies (10 poolies is the overall average, so we focused on pools in the 8-12 poolies range).

When we do that we get Figure 2 (again this is using ALL pools since 2011/2012 from TheHockeyPool.com).

Figure 2. Pools with 8, 9, 10, 11, and 12 poolies plotted against the historical probablility of winning the pool based on a poolie's draft position.

Once we do that, thankfully the odd dip in the second last position vanishes. But we do still see a few percentage points advantage to drafting first.

Now we mentioned earlier in this post that we looked into if having the top scorer in the league on your team influenced your chances of winning in a previous blog (link is here) and we saw a noted decline in this influence starting in 2017. So I then wanted to eliminate the influence of the pre-2017 first pick getting top-scorer and thus winning, so I re-ran the graph again with just pools since 2017 and got Figure 3.

Figure 3. Only pools since 2017 with 8, 9, 10, 11, and 12 poolies plotted agains the nistorical probability of winning the pool based on a poolie's draft position.

And surprisingly, this had the reverse effect. Instead of reducing the advantage of drafting first, it seems to indicate an advantage of 5% or more if you are drafting first.

Now of course there are caveats to this, we've put so many constraints on the input data, that our sample size is too small to really lock down the advantage. We are also limited to sampling only those who use TheHockeyPool.com, and we know there are keeper leagues, rotisserrie, unique scoring, and other factors that may skew the data.

So yes, it matters, but it is only a small percentage, and not insurmountable if you use your skill to make better selections.

What are your thoughts? I'd love to hear from you about this topic or about others you'd like me to cover in future blogs! Please let us know using the BLOG FEEDBACK link on the left side of page.