What first move give you the best chance of a good result?

As I had a bit of spare time I decided it might be interesting to use that to do a bit of practicing of data analysis, as nowadays I dont get much time anymore to "play with data". There are several databases with chess games available for download on the internet, and after a bit of googling I found a 200k games database from chessdigits (https://web.chessdigits.com/data). Kudos to them.

He/she/they/them already did some interesting stuff with these data, and for example looked at when one should resign (which can be found here: https://web.chessdigits.com/articles/when-should-you-resign). Really interesting stuff.

Anyways...

I thought I'd start at the beginning with a very simple question, already answered by many, many, many people throughout time. But why not. So, playing white, what is the best first move to make to get a good result? Instead of just looking at winning, I am more interested in the slightly more interesting question (and more relevant to myself) of how not to loose. And if at all possible to win as well.

I wont be going into modelling details, but DM me if interested. In short, after removing problematic games I ended up with a dataset of 150k games. I cleaned those up, and used the R statistical program to run an ordered (win, draw, loose) logistic regression model adjusted for ELO of the white player and the ELO difference between the white and black player. There were various other possibilities that didnt improve the model much (or at all). 

So for this first post, I am just going to give to top 3 opening moves for white that give you an increased probability to at least draw, but ideally give you a win:

1. Nf3

2. c4

3. d4

This may well generate a new interest in the Reti opening! The English, of course, was always a popular choice. As is d4. 

Of course, there is much more I could have done. For example split by player level, look separately at blitz (c4!), rapid (Nf3) and classical chess (Nf3), etc...but that may all be for a future post.

If this sparked a new interest in the Reti opening, there is actually a really good book available on the topic: https://thinkerspublishing.com/product/adrien-demuth-the-modernized-reti/

*

You may also be interested in the worst opening moves? I am sure it comes as little surprise that opening with a3 or f3 does not do you any favors .

I hope you enjoyed this first analytic post. There will be more when I can find some time.

Also, feel free to argue and disagree about this material...