Hey everyone, my name is Anthony, also known as LinkDFS! I have received a lot of questions regarding the projections that I post — so I have decided to write an article explaining the proper way to view the projections!
- Underdog Fantasy Affiliate Link (Promo Code: ROTORADAR = 100% Deposit Match up to $100)
- Install Free Sleeper App
- Play daily fantasy and higher/lower contests on Sleeper Picks
- Sports Injury Central
- Props.Cash (Promo Code: ROTORADAR 25% off their Monthly All Sports Package!)
- Pikkit (Code: ROTORADAR)
- Vivid Picks ($100 Sign-up Bonus Code: ROTORADAR)
- RunGoodProps (Promo Code: RADAR 50% off your first month!)
HOW TO READ THE MODELS:
- Every projection model has the following four items…
- A line. Whether for kills, headshots (HS), FPTS, etc. This is what the sites have the player set at
- Odds. These are the odds for the TEAM to win. They are NOT for the specific player
- A projection. This can be for kills, headshots (HS), FPTS, etc. This is what the model is projecting the player for
- A diff. This is the difference between what the line the sites have and what the model is projecting
- The process of reading the model is simple…
- If a player has a positive diff, take their over
- For example, GruBy has a +3.53 diff, which means he’s projected for 3.53 more kills than the site has his line
- If a player has a negative diff, take them under
- For example, MUTiRiS has a -2.59 diff, which means he’s projected for 2.59 fewer kills than the site has his line
- If a player has a positive diff, take their over
Thresholds of Plays to Consider:
CSGO & VALORANT & CS2
- The model should be used as a strong starting point, not convictions. Personally, I use the model to find the largest differences and then research those specific plays.
- At this time due to data limitations, the projections do not account for expected maps.
- Players that are highlighted in yellow (or have an asterisk, if you’re using the website version) means that their sample size is volatile. Additional research should ALWAYS be done before playing them.
- In BO1s (Map 1) for CSGO, a +/- 1 difference or greater is a must, anything lower is a fade. Please be aware that regardless of the difference for BO1s, the variance is significantly higher.
- In BO3s (Map 1+2) for CSGO, a +/- 2 difference or greater is the preferred target range for plays. Plays in the +/- 1.00-1.99 range should be considered as leans and played at a lower volume. Anything lower than 1.00 should be considered a fade.
- In BO5s (Map 1+2+3) for CSGO, a +/- 3 difference or greater is the preferred target range for plays. Plays in the +/- 1.50-2.99 range should be considered as leans and played at a lower volume. Anything lower than 1.50 should be considered a fade.
- Headshots follow the same general principles for the map sets listed above with the caveat that they’re more volatile than kills, especially for AWPers.
- When playing CSGO players from the same game, it’s a good idea to AVOID playing an over and under. Why? With an over, you want as many rounds as possible. With an under, you want as few rounds as possible. Essentially, negative correlation.
- If plays get bumped, they are still fine to play as long as they fit the criteria stated above.
- Note: VALORANT & CS2 are currently in the beta-testing phase and the margins for that are subject to change based on further testing.
CS2 Stack Sheet
- The purpose of CS2 Stack Sheet is to highlight the top correlation options. In CS2, over or under stacks of a specific game correlate well together as more rounds = more kills and less rounds = less kills. The Stack Sheet highlights those correlation plays with the help of the projections.
- On the sheet, there’s an SS column. SS = Stack Score, the higher the SS, the stronger the correlation play is based on the projections.
- The SS is still a WIP and will be refined over time. The SS should be used as a rough guideline.
- Only 1 stack should be used per game, if there’s an over and under stack for one game, the preferred option between the two is the one with a higher SS.
- A “Yes” with a RED background means it’s a viable UNDER stack.
- A “Yes” with a GREEN background means it’s a viable OVER stack.
- The blue background in the SS column highlights the Top 3 stacks based on their score.
- For PrizePicks, the preferred method of use for the Stack Sheet is to take 2 or 3 2-man stacks from different games and pair them in an entry for a 10x or 25x chance.
- For Underdog, with their payout changes, there’s no one right way to make use of the Stack Sheet though 4-picks with two 2-man stacks looks the most appealing.
- As always, these tools should be used as a resource for research and not conviction. As previously mentioned, this tool is a WIP and will be tweaked over time!
CSGO & CS2 w/Map projections
- The difference guidelines for the CSGO & CS2 projections listed above still apply here, however, the maps listed are the projected maps. The map projections are based on team analytics. The map projections are as it says, are projections, they WILL NOT be 100% accurate. Yes, the differences will look much larger than normal but no they are not nukes/locks/etc.
- Expect varying results as with these projections, we’re dealing with unknown variables (maps) that are not guaranteed to be what’s projected.
- In the Map columns, you’ll notice + / green cells and – / red cells
- If a player has a map that has a + icon / a green color coating, that means that are in the Top ~10% percentile of Kills Per Round (KPR) on the Map. This means on one of the projected maps, the player has a higher ceiling.
- If a player has a + icon / green color coating and is favored to go higher than his line, this is an added positive bonus as the player has a higher ceiling on the map
- If a player has a + icon / green color coating and is favored to go lower than his line, this is an added negative bonus. Players that fit this criteria should be played at a lower volume or avoided altogether, depending on the risk you’re willing to take.
- If a player has a map that has a – icon / a red color coating, that means that are in the Bottom ~10% percentile of Kills Per Round (KPR) on the Map. This means on one of the projected maps, the player has a lower ceiling.
- If a player has a – icon / red color coating and is favored to go lower than his line, this is an added positive bonus as the player has a lower ceiling on the map
- If a player has a – icon / red color coating and is favored to go higher than his line, this is an added negative bonus. Players that fit this criteria should be played at a lower volume or avoided altogether, depending on the risk you’re willing to take.
- If a player happens to have both a – icon / red color coating & a + icon / green color coating, regardless, it should be treated as an overall negative bonus. Players that fit this criteria should be played at a lower volume or avoided altogether, depending on the risk you’re willing to take.
- If a player has a map that has a + icon / a green color coating, that means that are in the Top ~10% percentile of Kills Per Round (KPR) on the Map. This means on one of the projected maps, the player has a higher ceiling.
- For CS2 w/Map projections, there is a possibility for double asterisk players (**) instead of the typical single asterisk (*). Double asterisk players are different than single asterisk players. Double asterisk players indicate that while the player profiles well, there’s a large range of outcomes on the likely maps for the player. To put it in perspective, the risk is between a regular asterisk player (*) and a normal line. Unlike single asterisk players (*), you do not need to do additional research for double asterisk players (**), just know their range of outcomes is higher than normal.
League of Legends & DOTA2
- With constant meta shifts, the LoL projections should be used as a strong starting point to your research process and not conviction.
- Pace (also known as CKPM) should be considered when taking higher/lowers and that can be found here for every region.
- Players that are highlighted in yellow (or asterisk, if you’re using the website version) means that their sample size is volatile. Additional research should ALWAYS be done before playing them.
- In BO1s (Map 1) for LoL Kills, a +/- 1 difference or greater is a must, anything lower is a fade. Please be aware that regardless of the difference for BO1s, the variance is significantly higher.
- In BO3s (Maps 1+2) for LoL Kills, a +/- 2 difference or greater is the preferred target range for plays. Plays in the +/- 1.00-1.99 range should be considered as leans and played at a lower volume. Anything lower than 1.00 should be considered a fade.
- In BO5s (Maps 1+2+3) for LoL Kills, a +/- 3 difference or greater is the preferred target range for plays. Plays in the +/- 1.50-2.99 range should be considered as leans and played at a lower volume. Anything lower than 1.50 should be considered a fade.
- Death and Assist follow the exact same criteria as Kills for LoL (subject to change since they are currently in beta).
- We’ve recently (at the time of writing this article) gotten Fantasy Points (FPTS) on Underdog for LoL. The target differentials for FPTS are WAY DIFFERENT than they are for kills.
- Why? 1 Kill = 3 FPTS so if our minimum target for Kills is +/- 1, it’d make no sense for the minimum for FPTS to be +/- 1 when 1 FPT equates to 0.33 of a kill.
- In BO1s (Map 1) for LoL FPTS, a +/- 3 difference or greater is a must, anything lower is a fade. Please be aware that regardless of the difference for BO1s, the variance is significantly higher.
- In BO3s (Maps 1+2) for LoL FPTS, a +/- 6 difference or greater is the preferred target range for plays. Plays in the +/- 3.00-5.99 range should be considered as leans and played at a lower volume. Anything lower than 3.00 should be considered a fade.
- In BO5s (Maps 1+2+3) for LoL Kills, a +/- 9 difference or greater is the preferred target range for plays. Plays in the +/- 6-8.99 range should be considered as leans and played at a lower volume. Anything lower than 6.00 should be considered a fade.
- After testing, the target differentials for FPTS are different for DOTA2 as the lines are much larger for DOTA2 than LoL due to the nature of the game.
- In BO1s (Map 1) for DOTA2 FPTS, a +/- 6 difference or greater is a must, anything lower is a fade. Please be aware that regardless of the difference for BO1s, the variance is significantly higher.
- In BO3s (Maps 1+2) for DOTA2 FPTS, a +/- 9 difference or greater is the preferred target range for plays. Plays in the +/- 6.00-8.99 range should be considered as leans and played at a lower volume. Anything lower than 6.00 should be considered a fade.
- In BO5s (Maps 1+2+3) for DOTA2 FPTS, a +/- 12 difference or greater is the preferred target range for plays. Plays in the +/- 9-11.99 range should be considered as leans and played at a lower volume. Anything lower than 9.00 should be considered a fade.
- Combo offerings are two players with one line (Example: Hope + Deokdam at 13.5) which is a new offering and the exact metrics for them are still TBD. Obviously, since there are two players in the offering, a higher threshold makes sense but what that exactly is TBD, testing is currently ongoing! With testing still ongoing, I’d treat combo offerings as a product that is still currently in BETA.
- When playing LoL, it’s preferred to avoid playing overs from both teams in the same game. Why? With the way LoL is structured unless a game goes very long (which is mathematically a low probability), the winning team often gets far more kills than the losing team.
- Something to keep in mind, the FPTS scale is essentially 3x of what the kills/deaths/assists scales are. When selecting a player, be sure to keep this in mind.
- For example: Let’s say Tom has a 3 FPTS difference and a 1.5 Kills difference (or deaths/assists difference), the kills would be the better option because when considering the same scaling, a 3 FPTS difference would be like a 1 kill difference, hence why a 1.5 kills difference would be more favorable than a 3 FPTS difference.
- Note: DOTA2 is currently in the beta-testing phase and the margins for that are subject to change based on further testing.