Hell hath frozen over and I've decided to explore the possibility of producing a set of ratings for the 1.5 engine. This would have been unthinkable for me in the past, but a few things have made me consider going down this path. Having tested some of the free 1.5 ratings on the market I was left unimpressed. I also have some more time on my hands this month and with all the base data used to build my regular 2.1 ratings already compiled I figured I'd experiment a little with the 1.5. Make no mistake however, the 2.1 engine remains vastly superior and is the one everyone should be using.
However, back to the 1.5. Having not used this engine for a very long time I'm now in the process of re-familiarizing myself with it; running correlations and baseline tests to evaluate relationships and connections from which the ratings will grow. I imagine I could have a complete set of ratings available within a couple of weeks.
These ratings would meet the same standards for quality as my ratings for the 2.1 engine. They would naturally be based on the same wealth of unique data and go through the same rigorous testing process to ensure as realistic a simulation as the 1.5 engine will allow.
To illustrate the level of detail in my ratings we can take a look at Rebound Control (RC). I don't know exactly what data others use to quantify this stat but I build RC on actual rebound statistics. These aren't freely available to just quickly copy paste into Excel, but rather I have to download the full play-by-play data from every single game (1349 of them this past season including the playoffs) from nhl.com (this one: http://www.nhl.com/scores/htmlreports/2 ... 030415.HTM), isolate every single recorded event and derive rebound statistics from there. RC in my ratings is therefore based on stats such as: rebound frequency (also referred to as Frozen Puck percentage in the advanced stats community) and rebound save percentage. Thus, it's the best measure of actual rebound control we can produce. It's not just a number based on statistics measuring things other than rebounds, or worse yet, people guesstimating based on their opinions of players (it does happen!). This brief description of my RC-process should also give some idea of why my ratings for the 2.1 engine are a bit more expensive than most others (RC is just one of thirty categories so you can imagine that the work involved in making a full set of my ratings is quite an endeavor, but oh so much fun!).
As for cost, these 1.5 ratings wouldn't be near as expensive as the 2.1 ratings, simply because the process of data collection is already done so there's (a lot) less work involved when making these ratings. I was thus thinking of a price of about $20 dollars for these as this would appear to be right on average for one season of the three commercial sets I could find listed on the forum (BRHL @ $35, Schifty031 @ $15 and 2M4B @ $10). So for once, I wouldn't be the most expensive option!
And since this is more of an experimental exercise for myself in testing the 1.5 engine again after so many years, I would feel hesitant to offer up a subscription service for this one at this time. I also don't know what type of interest there would be in this product but if the interest is there and the feedback on how the ratings perform is anything like what I'm getting on the 2.1 ratings, I might in future make these 1.5 ratings a recurring thing. Should that happen, people who bought them this year would certainly be able to benefit from that next year in the form of a discount similar to a subscription.
If you want to encourage this venture into madness feel free to do so.
This forum should be use to discust players ratings. / Ce forum devrait être utilisé par discuté des côtes des joueurs.
1 post • Page 1 of 1