Built a data project looking at VGK trades and wanted to share because some of the results surprised me.
It measures trade value using on-ice impact (xGAR, expected goals above replacement), expected value from picks/prospects, and cap efficiency.
Curious what you guys think; anything here that looks off or surprising?
Full interactive version if you want to dig in here
4 comments
A couple notes on how I handled some of the edge cases:
The Fleury trade is probably the biggest question mark on here. The model shows it as a win for VGK, but that’s mostly driven by the context provided, Fleury’s post-trade value comes out negative because he went to a really bad Chicago team, and there’s essentially no incoming value on the VGK side since the return never played. So statistically it looks like VGK “won,” but I wouldn’t really interpret it that way in practice.
Retained salary also isn’t included in the value calculation. Most retention is short-term and usually tied to expiring deals, so it doesn’t move the needle much in this model, though there are exceptions like Hertl, who I ignored the $1.4 million AAV retained for, for the sake of consistency.
Not every draft pick is included in expected value either. Some were traded again before the selection, some players were moved before ever playing in the NHL, and some picks are just hard to trace cleanly back to one outcome. The goal here wasn’t to model the entire trade tree, just the direct value from each move involving Vegas.
I plan to add the Rasmus Andersson trade as well sometime post playoffs once some of the conditions on VGK’s draft picks have settled, and once we get an answer to Andersson’s future.
Would Wild Bill be considered a trade since we were given picks to take him in the expansion draft? If so that might be the top one.
This is why I really wouldn’t want to trade gms for someone else. Do they make mistakes? Sure, they’ve done a few harsh things.
But just look at the trades that made the team as good as they are. I dont think there’s many options, if any, that could be as good as they have been.
Ok so this is really cool!
I have some problems with the end result. You didn’t explain your methodology or provide sources for your stats and there’s a lot of missing variables and context that I think are important for evaluating each trade. I have a lot of questions about what you did, how you got there, and why you did it. There’s a lot more work to do here, but you have some good bones to start with.