How to Build a Sports Prop Betting Tool

[@BBJamesB]

Where do you teach what you're doing with your formulas before? Do you have a video for that somewhere?

[@wageredontilt1649]

Some of them will be here: Back to Basics of Excel for Models kzbin.info/www/bejne/aJq2hqynpdyYeZo

[@wageredontilt1649]

This series will be getting more walkthroughs on excel as well, hopefully in next couple weeks

[@MJMoneyline]

hey brotha thanks for the video, im having an issue, the numbers keep resetting every time I paste in the J Column, is there something im doing wrong?

[@wageredontilt1649]

So, the numbers will continually reset, which is okay. The resetting is due to interacting with a cell. When you have a formula in a spread sheet (Google or Excel), it will rerun every time you interact with another cell. Since the formula is using a random number generator, it will rerun every time. This should be fine because based upon your sample size, and how many iterations you have set, it shouldn't change anything too drastically in the finalized output.

[@MJMoneyline]

@@wageredontilt1649 Thank You Man! I truly learned a lot! Would swap out catch % for an hit rate % let’s say for a running back prop or NBA prop?

[@wageredontilt1649]

@TheMoneylineJae you can do that as well. One thing I’d do though is look at some of the other videos on props and modeling. Dealing with RB and NBA baskets gets a bit more complex as you need to account for volume/plays, as well as how the defense will respond.

[@MJMoneyline]

@@wageredontilt1649 Man this is Awesome! I ran the model on Brandon Cooks Receptions o3.5 and that cleared! I can’t thank you enough

[@chrissima7512]

Thank you for the walk-through! A couple questions for you: What other factors go into to the projections and how could this be implemented? Also, how do the projections work on a weekly basis?

[@wageredontilt1649]

Glad it was helpful! So, for other factors that can go into the model would be differing stats using a very similar pattern. Example: If I wanted to do receiving yards for a WR, I would check who the team is that they are playing against and get their stats for passing yards allowed and evaluate their rank against QBs or DVOA. I'd use a QB stats for depth of passes per game as well, and the receivers targets and catches, as well as yardage. Once you have those stats, you would use a similar process of finding the Mean and the Standard Deviation, and running the simulations. Combining the stats will give you a much more in depth look. As far as a week by week basis, this can be a bit trickier. It would depend upon what sport you are doing, and what the statistical swings are for the sport. In the NFL the sample size is very small, so each game can have high variance making it more complicated and risky to adjust on a weekly basis. Something like the NHL and MLB would have a better guiding factor since they have so many games in a week. I tend to adjust my stats about every 2 weeks for most sports. There will always be times when a team is "hot" and creates an outlier in the data. If you take that outlier in without other "padding" data to show it is an outlier, then your numbers will likely be off. So in a shorter version, to avoid high variance in the data and getting a bad snap shot, I like to collect new stats every 2 weeks or so. It allows a larger sample size of change in players. Hope that helps! If not, let me know and I can try to give further examples in a video and do a brief walk through.

[@chrissima7512]

@@wageredontilt1649 That was great! I completely understand the high variance for NFL games because there are fewer games than MLB, NHL, and NBA. I'm glad you brought up DVOA, aDot, etc. because I use those in addition to other stats! I do have a follow-up question: In the sample your video demonstrates with Tyreek Hill, if I choose to include other statistics, such as DVOA, aDot, etc., I understand finding those statistic's mean and standard deviation. However, I am unsure where to include these means and standard deviations in my simulations: Would I add these additional means and standard deviations to the current cell formula running my simulations? (I believe your video demonstrated that the cell formula was located in B11) I cannot thank you enough for helping me out with this!

[@wageredontilt1649]

Yep, so you'd add some fields to look up the stats deviation and mean. Then include them in the formula for output. Before tossing them into the model itself, I'd do some quick manual tests on the stats you want to use to be sure they are relevant in the way they are "shaped". So, I wouldn't use a mean of DVOA, as DVOA is already a ranking value. So, I'd do some tests to see what that value actually looks like for player stats first. Then I'd be sure what stats I am using. If they are a per game, per play, per minute, per attempt, etc. and be sure that they are at the same rate. Your numbers will look wild if you use a mean value from a stat that is on a play by play basis vs a game by game basis. I am going to be going through one of my prop tools this week, and create a new video on a more complex prop model. Hope that helps!

[@chrissima7512]

@@wageredontilt1649 Very helpful and thanks again!

[@chrissima7512]

@@wageredontilt1649 Sorry for another message and question! My work-in-progress model is based on a receptions prop for Tyreek Hill on a game-by-game basis. I have all data/statistics from the past four years, such as his targets and reception % that was shown in your video. In addition, I researched a random team's (CIN Bengals) defense against. I pulled CIN's targets allowed/game against #1 WR over the course of the past 4 years and reception % allowed over the past four years. Would you have any advice about how to write the formula I would use? I'm researching how each formula works but I want to make sure the correct formula is used to get the correct projection. I think this is where I am lost on haha. Thanks!

[@eduardosalas3509]

Is this only for receptions or can it be for like other stat categories like receiving yards and other stats like that

[@wageredontilt1649]

This can be used for other stats like receiving yards, and other stats that are non-discrete (non-count). This is more of an intro model which is why I used it for a discrete type of prop like receptions. For discrete, you would want to use more of binom.dist or poisson depending upon what specifically you are looking for.

[@morganlindsey6713]

any way to do a video on nba or nhl?? im trying to figure out how many games to go back to calculate means and standard devs for different stats thanks great stuff man

[@wageredontilt1649]

Yep, I can do one on basketball. The NHL I can look at a separate, but those sports will take a slightly differing approach. I will try and have up on KZbin in a week or so.

[@SportsFormm]

Thanks for this video and sorry for the late response but would you consider taking into account the defense of the opponent and how that could play a role in any given stat?

[@wageredontilt1649]

Yes, I would definitely take that into account, as the defense will always have an impact on the output of a players performance. It will range based upon what type of play you are building the model for however

[@ianharkins2172]

Another question, is there a particular reason to the process for taking max and min years and getting the mean and standard dev that way? Instead of gathering game by game data for the same range of years and just pulling the mean and standard dev from that data?

[@wageredontilt1649]

Good question, and no. You can pull specific seasons game by game data and use mean, standard deviation or median. You can also used different subsets of data like home or away, etc. This was one of the earlier videos I made on this, and didn’t want to get too deep in what data to pull, vs just using a basic high and low concept to find the mean.

[@ivanomatijevic8822]

You should definitely do this for NBA especially points since there inst many of those on yt and tbh it would help me too

[@wageredontilt1649]

I do have a version of this for NBA 3p in the updated prop tool video that may be helpful. In the updated version I use a similar method but also account for the opponent

[@ianharkins2172]

Does this process work for statistics or events that don't follow a normal distribution? Something like a negative binomial?

[@wageredontilt1649]

The process would be the same but the formula you’d use binomDist or negative binom Dist formulas in excel

[@christophercobler7405]

Hey, Love the video it helped a lot! With the newer Excel if i hit control down arrow key i got to like 100k. Do you know how to prevent or customize excel to have a max of 10000 cells? Second, Do i have to leave the 10000 up per player or is there a way to keep the over/under % without it ? I am trying to build one for an entire board vs 8 players that is why I am curious. Thanks!

[@wageredontilt1649]

To limit the scrollable range, you can right click on the excel sheet, and click view code. Then when the VBA editor opens, there is a left panel. In the panel is the scrollable field. You'd enter the range you want to scroll. Example: A1:Z100005. So, without using VBA the only way to keep the over under would be to copy the cell, then right click the same cell. Select Paste Special. There will be a modal that opens. From the modal, you'd select "Values" and then click okay. That will take the actual numbers you want and paste them as the value and not a formula. That is the only way without vba.

[@christophercobler7405]

@@wageredontilt1649 Thank you for the feedback, I will give this a go after work tonight. One other question I was thinking about as I was working today is if you wanted to give a projection number along with your over under percentages is it as simple as multiplying the percentages to the line? In your example of Tyreek Hill at 5 for 75% for the under would the projection calculation be .75 * 5? Just curious if you have any thoughts on how to get a "Projection number?" Your video by far the best I have found and I followed it to a T so thank you, just trying to expand a little on it and im struggling to find a solid formula that I can "test results" with!

[@wageredontilt1649]

Unfortunately that method would not work because that 75% is a probability of summed options of 0, 1, 2, 3, 4. You'd want to look at also doing Poisson Distribution, or a slightly differing method of modeling. I have a few other videos on these, and currently working on one.

[@christophercobler7405]

@@wageredontilt1649 I look forward to the completion! I think this video will suffice for now ill use it to test results but I feel like it needs to be more current season based vs long term based because teams change players get better or worse etc. Not sure though when I tried it every time it gave me 100% over or 100% under as im sure its due to lack of the amount of data. So, I am currently going to use your method here and test results! Thank you for the responses!

[@christophercobler7405]

@@wageredontilt1649 Also, for players like Alec Pierce where this is his first season, I cant figure out how to do this method in a sense where I could test. Im thinking its just using maximums no minimums and see how accurate that can be but likely isnt accurate and would require the Poisson method. I cant seem to find an easy video to set that up and try that method out as well. Do you have a Poisson Distribution for props video by chance?

[@johnborges1127]

Hey. Great video! How would you suggest doing this for calculating an nba player’s points? I saw your other video with 3pts which makes sense. Would you try and factor in how many points that team gives up to the position like with the 3pts? Thanks.

[@wageredontilt1649]

Correct. If doing overall pts, it will be a bit more though. You’d wanna adjust for playing time, possession, and be sure to include FT. So how often do they get fouled and how often does the opponent foul. Id start off doing basics, then include pace and possessions. Id do pace after because it makes stuff more complex, but you’d want to include it. If a team is not up tempo and slows in the second half, you’d wanna account for that. Great question though! Let me know if you need any other assistance.

[@ClipsNotMovies]

@@wageredontilt1649 do you have a template based on nba player points if so is it available for download?

[@wageredontilt1649]

I do have a basic template that takes in standard stats but does not utilize deep stats. Would you still want that template?

[@johnborges1127]

@@wageredontilt1649 absolutely! That would be great.

[@JiggyKey]

@@wageredontilt1649how can I get a template for nba stats

[@famoustiz4624]

You have a video like this but for MLB?

[@wageredontilt1649]

I have not fully done a walk through on mlb. I did do some for Unabated that you could check out.

[@343frankieboii]

Hello, can I bet on a player leading his team in rebounds?

[@wageredontilt1649]

You would be able to if the book you use offers that bet

[@ryanlally6393]

If this was made on Google docs you should have this document open or share with people!!!

[@wageredontilt1649]

I am not a fan of actually sharing the Google Doc itself, however I do download the file and drop it into MediaFire. From there you can upload into your Google Doc if you'd like. The link for the prop tool is www.mediafire.com/file/11lovhq5kpmymi4/Player_Prop_Sheet.xlsx/file Let me know if you need any assistance with the formulas or anything!

[@ryanlally6393]

@@wageredontilt1649 gotta try and find the time and make one 😅thanks

[@TheeDonnieRey]

Appreciate the video man. Tryna get serious about my bets and I feel like all the professionals model. Just tryna learn how to do it is difficult. Thanks.

[@yessirski7868]

Can you make a mlb strikeouts model?

[@wageredontilt1649]

Sure, let me take a look into it. I have a basic model, however I want to try and provide an updated/upgraded version. Let me see what I can think through.

[@yessirski7868]

@@wageredontilt1649 keep me updated! Thank you!