For the fifth post in this series we're going to tie together some of the previous posts to propose some concrete steps to determine who your best players are and who to play the most often.
Kind of long but lots of good information in here.
To review, in the first post
we discussed how getting the first elimination in a game changes the average team's odds of winning from 50% to 75%. In the third post
we discussed how the odds of winning dropped to 28% when you are down by one point and all the way down to 4% when you are down by two points. If you haven't come to the logical conclusion of these two numbers let me propose that together they mean, in general, little things matter a HUGE
amount in college paintball. Specifically, getting to 5v4's and winning points when your team gets the first point elimination is probably the most important part of college paintball.
"But Alex", you say, "what about 3v2s/1v1s/that time I was SUPER baller and shot out 3 people etc?"
Excellent question! Let me put it this way: the most common situation you will face as a college paintball player is a 5v5 followed by being in a 5v4 or 4v5. Since there is such a huge difference in the odds of winning a 5v4 (75%) versus a 4v5 (25%) coupled with the high odds of winning when you are up on points then you:
a. want to maximize your chances of getting to a 5v4
b. want to minimize your chances of losing a 5v4
"But Alex", you say, "that is all well and good but let's say I'm a college paintball club president like in post #3
, how do I apply the above to figure out who my best players are?"
Another good question! So for (a), the easiest answer is to track how good your players are at laning. You can do this pretty easily by simply writing down who plays each point and recording how many times you eliminate a player in the first 25 seconds of a point (how I've been defining laning). You can use a basic plus/minus for each player (e.g. if your line lanes someone out then everyone gets a +1, if not a -1). Those of you who having taken statistics can also use regression and dummy variables for each player and opposing team to get even more accurate numbers. The goal of this is to figure out who gets you into 5v4 scenarios.
A similar process can be used to track who gets laned OUT. If you team is great at laning but for everyone person you lane out you lose one player to being laned out then the two cancel each other out.
Which bring us to (b). Here we are going to add a new statistic: the average the chance of any given player being eliminated in a 5v4 if you are on the 5: 5%. In other words, an average player will be eliminated 5 times for every 100 5v4's they find themselves in. To be very clear, this is different from an average TEAM's
odds of having one of their players eliminated in a 5v4 which is about 30%. In other words, if you have 5 average players in a 5v4 and the odds PER PLAYER of being eliminated are 5% then the odds of AT LEAST one player being eliminated are around 30%. The math isn't exact but the numbers are close. (See here for an example.
We are interested in the case of "Either dice is a particular number").
To apply this to a real world example, you are the coach of a college paintball team. You are at practice and you notice that your team is pretty good at laning. You start paying attention to your team more when they are in 5v4s and you see one of your players (let's call him Adam) get shot out when your team is on a 5v4. You might think at this point "Wow! Alex said that the odds of being eliminated in a 5v4 are only 5%. Adam must be pretty bad." That isn't necessarily the case. Remember, the odds of at least one player being eliminated on the 5 side of 5v4 are 30%. Another point to make is that this is one game. Anything can happen in one game.
So how do you address this? Well, one way would be to pick a random player (let's call him Roger) and keep track of how many times he gets eliminated in 5v4s. THIS time if you see Roger get eliminated then you would be correct that he might not be that good as the odds of a randomly selected player being eliminated is about 5%.
An even better way would be to keep track of all the players on your team every time they are in a 5v4 and record each time they get eliminated. For each player, if you divide the number of times they were eliminated divided by the number of times they were in a 5v4 you will get their "5v4 elimination percentage". If that number is substantially over 5% you may want to think about putting that player in a little less often.
"Ok," you say "but what about the difference in positions? e.g. a snake player may get out more than say a back center"
Again good point, back centers on average get eliminated in 5v4's effectively 0% of the time. Fronts on the other hand get eliminated closer to 8-9% of the time. To simplify matters, anyone who is eliminated more than 5% of time in a 5v4 needs to be looked at pretty closely and anyone who is getting eliminated more than 10% should not be on your A line.
VERY curious to hear what people think about this post.