How Quickly Do Multi-Time All-Stars Reveal Themselves?

A very cursory glance at if there is a trend when it comes multi-time all-stars.

As we head into draft week, I wanted to take a quick analytical look at when a team can reasonably expect a player to reveal that they have the potential to be a multi-time all-star. This isn’t an in-depth dive like my article on the best way to draft from a few years ago, but I thought it complemented the findings of that study rather well.

Methodology

There was not much science put into deciding the sample and methodology. I started in year 2005-06 and took the all-star rosters from every season since then. From there I catalogued when each all-star made their first appearance, the number of total appearances, and put them in the following positional buckets: Guard, Forward, Big. From there I defined multi-time all-star as making at least three all-star games. I know two technically fits the bill and is a positive outcome, but I wanted to subjectively narrow the field to a stronger crop of players. This left me with a sample of 70 players to analyze.

Results

The chart included below displays a box plot of the first season each qualifying player appeared in an all-star game.

The mean and median are both squarely 4. In other words, the most likely first appearance for a multi-time all-star will happen in season four of that player’s career. The spread is quite large with Kyle Lowry and Chauncey Billups each reaching their first appearance in season 9, along with serval year one all-stars in Tim Duncan, Shaquille O’Neal, and Yao Ming. The size of the interquartile range suggests that Lowry and Billups are indeed outliers when it comes to their first appearances. Nevertheless, the data says on balance, you can expect the first appearance of a multi-time all-star in season four.

Next, I evaluated positional groups to see if certain position types develop quicker than others. The positions were divided into guard, forward, and big. The chart below displays the spread of first appearances for each position.

Guards saw the largest range of outcomes, with the lowest being season 2, the highest being season 9. The mean (hashed line) was 4.29 and the median (solid line) was 4. Guards also had the largest group of players qualifying for multi-time all-star status at 28. The majority of guards fell between 2.5 and 5.5 - which are the 25th and 75th percentiles respectively. The gap between the top end and the 75th percentile was not quite big enough to suggest an outlier.

Forwards had the narrowest range, with the lowest being season 2, and the highest being season 7. The mean was 3.75 and the median was 4. The sample for forwards meeting the criteria was 16 players - smallest group of the three. The majority of the forwards fell between 3 and 4, leaving the narrowest interquartile range of the group at 1. Due to how narrow the interquartile range was, any seasons above 5.5 would suggest outlier activity.

Bigs had the same range as guards but with the floor being season 1 and ceiling being season 8. The mean was 3.85 and the median was 3.5 The majority of bigs fell between 3 and 5, leaving an interquartile range of 2. The gap between the 75th percentile and the top end does not suggest the presence of outliers.

Conclusions

I thought it was quite telling that the mean and median outcomes for first all-star appearance was season 4. This dovetails really nicely with my findings for a previous article that most first round picks do not outproduce veteran minimum contracts until year 4 if they ever do, and it was the group of players within pick range 1-10 that saw the most separation.

I thought that the positional analysis was equally telling, as each group congregated around season 4 as well. At first glance it appeared that bigs “revealed” themselves more quickly but there was enough standard deviation overlap between the positional groups to suggest no statistically significant difference between the means or medians. Season 4 seems to be the magic number throughout.

I did find it curious that guards led the pack when it comes to multiple appearances, followed closely by bigs. Forwards being the smallest group might suggest the value and rarity of finding a multi-time all-star in that position, but it might also suggest the lowest likelihood of finding a player of that caliber, period. Could it be that betting on guards and bigs is the best path to acquiring these kinds of players? At the very least, you may also find a multi-time all-star much later into the careers of guards and bigs relative to the forwards.

I also found it interesting that there was a mild correlation between total number of appearances and year of first appearance as shown in the graph below:

Nevertheless - there appears to be a great deal of variance that isn’t accounted for and I am sure there is selection bias at play.

Based on the results, I think a reasonable assumption is that if the first all-star appearance generally occurs in season 4, the team is likely enjoying above average production by year 3. Nevertheless, these results once again reiterate how rare it is for players to walk into the league and have all-star level impact. Of the 70 players qualifying for the multi-time all-star criteria in this study, only 16 players had their first appearance in years one and two. There likely is a level of survivorship bias occurring here, as the earlier a player is selected an all-star, the more chances they will have to repeat.

I do think the findings support my overall research about player development and production. Teams drafting for “win now” scenarios are likely to find themselves on the same developmental timeline as teams drafting for “best player available”.